docs/master: discussion.roff annotate

docs/master

annotate discussion.roff @ 140:5cc81413ae62

bib: Added some line breaks.

author	markus schnalke <meillo@marmaro.de>
date	Thu, 05 Jul 2012 13:12:40 +0200
parents	cc35686f359e
children	3361e53dfcd6

rev	line source
meillo@58	1 .H0 "Discussion
meillo@0	2 .P
meillo@58	3 This main chapter discusses the practical work done in the mmh project.
meillo@104	4 It is structured along the goals to achieve.
meillo@104	5 The concrete work done
meillo@58	6 is described in the examples of how the general goals were achieved.
meillo@87	7 The discussion compares the current version of mmh with the state of
meillo@87	8 nmh just before the mmh project started, i.e. Fall 2011.
meillo@87	9 Current changes of nmh will be mentioned only as side notes.
meillo@87	10 .\" XXX where do I discuss the parallel development of nmh?
meillo@58	11
meillo@58	12
meillo@58	13
meillo@133	14 .\" --------------------------------------------------------------
meillo@125	15 .H1 "Streamlining
meillo@58	16
meillo@0	17 .P
meillo@58	18 MH had been considered an all-in-one system for mail handling.
meillo@58	19 The community around nmh has a similar understanding.
meillo@87	20 In fundamental difference, mmh shall be a MUA only.
meillo@87	21 I believe that the development of all-in-one mail systems is obsolete.
meillo@87	22 Today, email is too complex to be fully covered by single projects.
meillo@87	23 Such a project won't be able to excel in all aspects.
meillo@87	24 Instead, the aspects of email should be covered my multiple projects,
meillo@87	25 which then can be combined to form a complete system.
meillo@87	26 Excellent implementations for the various aspects of email exist already.
meillo@87	27 Just to name three examples: Postfix is a specialized MTA,
meillo@87	28 Procmail is a specialized MDA, and Fetchmail is a specialized MRA.
meillo@89	29 I believe that it is best to use such specialized tools instead of
meillo@87	30 providing the same function again as a side-component in the project.
meillo@58	31 .P
meillo@87	32 Doing something well, requires to focus on a small set of specific aspects.
meillo@87	33 Under the assumption that focused development produces better results
meillo@100	34 in the particular area, specialized projects will be superior
meillo@87	35 in their field of focus.
meillo@87	36 Hence, all-in-one mail system projects \(en no matter if monolithic
meillo@87	37 or modular \(en will never be the best choice in any of the fields.
meillo@87	38 Even in providing the best consistent all-in-one system they are likely
meillo@87	39 to be beaten by projects that focus only on integrating existing mail
meillo@89	40 components to a homogeneous system.
meillo@87	41 .P
meillo@87	42 The limiting resource in Free Software community development
meillo@87	43 is usually man power.
meillo@87	44 If the development power is spread over a large development area,
meillo@87	45 it becomes even more difficult to compete with the specialists in the
meillo@87	46 various fields.
meillo@87	47 The concrete situation for MH-based mail systems is even tougher,
meillo@87	48 given the small and aged community, including both developers and users,
meillo@87	49 it has.
meillo@87	50 .P
meillo@87	51 In consequence, I believe that the available development resources
meillo@100	52 should focus on the point where MH is most unique.
meillo@87	53 This is clearly the user interface \(en the MUA.
meillo@125	54 Peripheral parts should be removed to streamline mmh for the MUA task.
meillo@60	55
meillo@60	56
meillo@100	57 .H2 "Mail Transfer Facilities
meillo@60	58 .P
meillo@60	59 In contrast to nmh, which also provides mail submission and mail retrieval
meillo@87	60 agents, mmh is a MUA only.
meillo@100	61 This general difference initiated the development of mmh.
meillo@66	62 Removing the mail transfer facilities had been the first work task
meillo@76	63 in the mmh project.
meillo@60	64 .P
meillo@105	65 Focusing on one mail agent role only is motivated by Eric Allman's
meillo@105	66 experience with Sendmail.
meillo@105	67 He identified limiting Sendmail the MTA task had be one reason for
meillo@105	68 its success:
meillo@105	69 .[ [
meillo@105	70 costales sendmail
meillo@105	71 .], p. xviii]
meillo@105	72 .QS
meillo@105	73 Second, I limited myself to the routing function \(en
meillo@110	74 I wouldn't write user agents or delivery back-ends.
meillo@105	75 This was a departure of the dominant through of the time,
meillo@105	76 in which routing logic, local delivery, and often the network code
meillo@105	77 were incorporated directly into the user agents.
meillo@105	78 .QE
meillo@105	79 .P
meillo@105	80 In mmh, the Mail Submission Agent (MSA) is called
meillo@105	81 \fIMessage Transfer Service\fP (MTS).
meillo@105	82 This facility, implemented by the
meillo@105	83 .Pn post
meillo@105	84 command, established network connections and spoke SMTP to submit
meillo@60	85 messages for relay to the outside world.
meillo@105	86 The changes in email demanded changes in this part of nmh too.
meillo@89	87 Encryption and authentication for network connections
meillo@87	88 needed to be supported, hence TLS and SASL were introduced into nmh.
meillo@87	89 This added complexity to nmh without improving it in its core functions.
meillo@87	90 Also, keeping up with recent developments in the field of
meillo@87	91 mail transfer requires development power and specialists.
meillo@87	92 In mmh this whole facility was simply cut off.
meillo@76	93 .Ci f6aa95b724fd8c791164abe7ee5468bf5c34f226
meillo@76	94 .Ci fecd5d34f65597a4dfa16aeabea7d74b191532c3
meillo@76	95 .Ci 156d35f6425bea4c1ed3c4c79783dc613379c65b
meillo@87	96 Instead, mmh depends on an external MSA.
meillo@60	97 The only outgoing interface available to mmh is the
meillo@60	98 .Pn sendmail
meillo@87	99 command, which almost any MSA provides.
meillo@87	100 If not, a wrapper program can be written.
meillo@87	101 It must read the message from the standard input, extract the
meillo@87	102 recipient addresses from the message header, and hand the message
meillo@87	103 over to the MSA.
meillo@87	104 For example, a wrapper script for qmail would be:
meillo@87	105 .VS
meillo@87	106 #!/bin/sh
meillo@138	107 exec qmail-inject # ignore command line arguments
meillo@87	108 VE
meillo@87	109 The requirement to parse the recipient addresses out of the message header
meillo@87	110 is likely to be removed in the future.
meillo@87	111 Then mmh would give the recipient addresses as command line arguments.
meillo@100	112 This appears to be the better interface.
meillo@87	113 .\" XXX implement it
meillo@60	114 .P
meillo@60	115 To retrieve mail, the
meillo@60	116 .Pn inc
meillo@100	117 command acted as Mail Retrieval Agent (MRA).
meillo@100	118 It established network connections
meillo@76	119 and spoke POP3 to retrieve mail from remote servers.
meillo@76	120 As with mail submission, the network connections required encryption and
meillo@87	121 authentication, thus TLS and SASL were added.
meillo@87	122 Support for message retrieval through IMAP will become necessary
meillo@100	123 to be added soon, too, and likewise for any other changes in mail transfer.
meillo@100	124 Not so for mmh because it has dropped the support for retrieving mail
meillo@100	125 from remote locations.
meillo@76	126 .Ci ab7b48411962d26439f92f35ed084d3d6275459c
meillo@76	127 Instead, it depends on an external tool to cover this task.
meillo@100	128 In mmh exist two paths for messages to enter mmh's mail storage:
meillo@100	129 (1) Mail can be incorporated with
meillo@60	130 .Pn inc
meillo@87	131 from the system maildrop, or (2) with
meillo@60	132 .Pn rcvstore
meillo@87	133 by reading them, one at a time, from the standard input.
meillo@60	134 .P
meillo@60	135 With the removal of the MSA and MRA, mmh converted from an all-in-one
meillo@87	136 mail system to being a MUA only.
meillo@60	137 Now, of course, mmh depends on third-party software.
meillo@87	138 An external MSA is required to transfer mail to the outside world;
meillo@60	139 an external MRA is required to retrieve mail from remote machines.
meillo@60	140 There exist excellent implementations of such software,
meillo@76	141 which do this specific task likely better than the internal
meillo@87	142 versions had done it.
meillo@87	143 Also, the best suiting programs can be freely chosen.
meillo@60	144 .P
meillo@60	145 As it had already been possible to use an external MSA or MRA,
meillo@60	146 why not keep the internal version for convenience?
meillo@76	147 The question whether there is sense in having a fall-back pager in all
meillo@76	148 the command line tools, for the cases when
meillo@60	149 .Pn more
meillo@60	150 or
meillo@60	151 .Pn less
meillo@76	152 aren't available, appears to be ridiculous.
meillo@100	153 Of course, MSAs and MRAs are more complex than text pagers
meillo@87	154 and not necessarily available but still the concept of orthogonal
meillo@87	155 design holds: ``Write programs that do one thing and do it well.''
meillo@87	156 .[
meillo@87	157 mcilroy unix phil
meillo@87	158 p. 53
meillo@87	159 .]
meillo@87	160 .[
meillo@87	161 mcilroy bstj foreword
meillo@87	162 .]
meillo@87	163 Here, this part of the Unix philosophy was applied not only
meillo@87	164 to the programs but to the project itself.
meillo@87	165 In other words:
meillo@87	166 ``Develop projects that focus on one thing and do it well.''
meillo@87	167 Projects grown complex should be split for the same reasons programs grown
meillo@87	168 complex should be split.
meillo@100	169 If it is conceptionally more elegant to have the MSA and MRA as
meillo@87	170 separate projects then they should be separated.
meillo@87	171 This is the case here, in my opinion.
meillo@87	172 The RFCs propose this separation by clearly distinguishing the different
meillo@87	173 mail handling tasks.
meillo@87	174 .[
meillo@87	175 rfc 821
meillo@87	176 .]
meillo@87	177 The small interfaces between the mail agents support the separation.
meillo@76	178 .P
meillo@87	179 In the beginning, email had been small and simple.
meillo@100	180 At that time,
meillo@60	181 .Pn /bin/mail
meillo@100	182 had covered anything there was to email and still had been small
meillo@100	183 and simple.
meillo@100	184 Later, the essential complexity of email increased.
meillo@87	185 (Essential complexity is the complexity defined by the problem itself.\0
meillo@87	186 .[[
meillo@87	187 brooks no silver bullet
meillo@87	188 .]])
meillo@87	189 Email systems reacted to this change: They grew.
meillo@100	190 RFCs started to introduce the concept of mail agents to separate the
meillo@100	191 various tasks because they became more extensive and new tasks appeared.
meillo@100	192 As the mail systems grew even more, parts were split off.
meillo@100	193 In nmh, for instance, the POP server, which was included in the original
meillo@100	194 MH, was removed.
meillo@100	195 Now is the time to go one step further and split the MSA and MRA off, too.
meillo@87	196 Not only does this decrease the code size of the project,
meillo@87	197 but, more important, it unburdens mmh of the whole field of
meillo@87	198 message transfer with all its implications for the project.
meillo@100	199 There is no more need to concern with changes in network transfer.
meillo@76	200 This independence is received by depending on an external program
meillo@76	201 that covers the field.
meillo@76	202 Today, this is a reasonable exchange.
meillo@60	203 .P
meillo@100	204 Functionality can be added in three different ways:
meillo@87	205 .BU
meillo@87	206 Implementing the function originally in the project.
meillo@87	207 .BU
meillo@87	208 Depending on a library that provides the function.
meillo@87	209 .BU
meillo@87	210 Depending on a program that provides the function.
meillo@87	211 .P
meillo@87	212 Whereas adding the function originally to the project increases the
meillo@76	213 code size most and requires most maintenance and development work,
meillo@87	214 it makes the project most independent of other software.
meillo@87	215 Using libraries or external programs require less maintenance work
meillo@87	216 but introduces dependencies on external software.
meillo@87	217 Programs have the smallest interfaces and provide the best separation
meillo@87	218 but possibly limit the information exchange.
meillo@87	219 External libraries are stronger connected than external programs,
meillo@87	220 thus information can be exchanged more flexible.
meillo@87	221 Adding code to a project increases maintenance work.
meillo@87	222 .\" XXX ref
meillo@100	223 Implementing complex functions originally in the project adds
meillo@87	224 a lot of code.
meillo@87	225 This should be avoided if possible.
meillo@66	226 Hence, the dependencies only change in kind, not in their existence.
meillo@66	227 In mmh, library dependencies on
meillo@66	228 .Pn libsasl2
meillo@66	229 and
meillo@66	230 .Pn libcrypto /\c
meillo@66	231 .Pn libssl
meillo@66	232 were treated against program dependencies on an MSA and an MRA.
meillo@87	233 This also meant treating build-time dependencies against run-time
meillo@87	234 dependencies.
meillo@66	235 Besides program dependencies providing the stronger separation
meillo@66	236 and being more flexible, they also allowed
meillo@66	237 over 6\\|000 lines of code to be removed from mmh.
meillo@66	238 This made mmh's code base about 12\\|% smaller.
meillo@87	239 Reducing the project's code size by such an amount without actually
meillo@87	240 losing functionality is a convincing argument.
meillo@87	241 Actually, as external MSAs and MRAs are likely superior to the
meillo@87	242 project's internal versions, the common user even gains functionality.
meillo@66	243 .P
meillo@76	244 Users of MH should not have problems to set up an external MSA and MRA.
meillo@60	245 Also, the popular MSAs and MRAs have large communities and a lot
meillo@60	246 of documentation available.
meillo@87	247 Choices for MSAs range from full-featured MTAs like
meillo@60	248 .I Postfix
meillo@87	249 over mid-size MTAs like
meillo@60	250 .I masqmail
meillo@60	251 and
meillo@60	252 .I dma
meillo@60	253 to small forwarders like
meillo@60	254 .I ssmtp
meillo@60	255 and
meillo@60	256 .I nullmailer .
meillo@60	257 Choices for MRAs include
meillo@60	258 .I fetchmail ,
meillo@60	259 .I getmail ,
meillo@60	260 .I mpop
meillo@60	261 and
meillo@60	262 .I fdm .
meillo@60	263
meillo@60	264
meillo@100	265 .H2 "Non-MUA Tools
meillo@60	266 .P
meillo@87	267 One goal of mmh is to remove the tools that are not part of the MUA's task.
meillo@89	268 Further more, any tools that don't improve the MUA's job significantly
meillo@87	269 should be removed.
meillo@87	270 Loosely related and rarely used tools distract from the lean appearance.
meillo@87	271 They require maintenance work without adding much to the core task.
meillo@125	272 By removing these tools, the project shall become more streamlined
meillo@87	273 and focused.
meillo@76	274 In mmh the following tools are not available anymore:
meillo@62	275 .BU
meillo@58	276 .Pn conflict
meillo@87	277 was removed
meillo@76	278 .Ci 8b235097cbd11d728c07b966cf131aa7133ce5a9
meillo@87	279 because it is a mail system maintenance tool that is not MUA-related.
meillo@87	280 It even checked
meillo@58	281 .Fn /etc/passwd
meillo@58	282 and
meillo@58	283 .Fn /etc/group
meillo@87	284 for consistency, which is completely unrelated to email.
meillo@87	285 A tool like
meillo@87	286 .Pn conflict
meillo@87	287 is surely useful, but it should not be shipped with mmh.
meillo@76	288 .\" XXX historic reasons?
meillo@62	289 .BU
meillo@58	290 .Pn rcvtty
meillo@87	291 was removed
meillo@87	292 .Ci 14767c94b3827be7c867196467ed7aea5f6f49b0
meillo@89	293 because its use case of writing to the user's terminal
meillo@76	294 on receiving of mail is obsolete.
meillo@87	295 If users like to be informed of new mail, the shell's
meillo@58	296 .Ev MAILPATH
meillo@87	297 variable or graphical notifications are technically more appealing.
meillo@100	298 Writing directly to terminals is hardly ever wanted today.
meillo@62	299 If though one wants to have it this way, the standard tool
meillo@58	300 .Pn write
meillo@58	301 can be used in a way similar to:
meillo@82	302 .VS
meillo@58	303 scan -file - \| write `id -un`
meillo@82	304 VE
meillo@62	305 .BU
meillo@58	306 .Pn viamail
meillo@87	307 was removed
meillo@87	308 .Ci eda72d6a7a7c20ff123043fb7f19c509ea01f932
meillo@87	309 when the new attachment system was activated, because
meillo@58	310 .Pn forw
meillo@76	311 could then cover the task itself.
meillo@62	312 The program
meillo@58	313 .Pn sendfiles
meillo@62	314 was rewritten as a shell script wrapper around
meillo@58	315 .Pn forw .
meillo@76	316 .Ci 0e82199cf3c991a173e0ac8aa776efdb3ded61e6
meillo@62	317 .BU
meillo@58	318 .Pn msgchk
meillo@87	319 was removed
meillo@87	320 .Ci bb9360ead7eb7a3fedcce2eeedfc660014e41dbe ,
meillo@87	321 because it lost its use case when POP support was removed.
meillo@76	322 A call to
meillo@58	323 .Pn msgchk
meillo@87	324 provided hardly more information than:
meillo@82	325 .VS
meillo@58	326 ls -l /var/mail/meillo
meillo@82	327 VE
meillo@100	328 It did distinguish between old and new mail, but
meillo@100	329 this detail information can be retrieved with
meillo@76	330 .Pn stat (1),
meillo@62	331 too.
meillo@100	332 A small shell script could be written to print the information
meillo@76	333 in a similar way, if truly necessary.
meillo@76	334 As mmh's
meillo@76	335 .Pn inc
meillo@87	336 only incorporates mail from the user's local maildrop,
meillo@62	337 and thus no data transfers over slow networks are involved,
meillo@76	338 there's hardly any need to check for new mail before incorporating it.
meillo@62	339 .BU
meillo@58	340 .Pn msh
meillo@87	341 was removed
meillo@76	342 .Ci 916690191222433a6923a4be54b0d8f6ac01bd02
meillo@87	343 because the tool was in conflict with the philosophy of MH.
meillo@76	344 It provided an interactive shell to access the features of MH,
meillo@76	345 but it wasn't just a shell, tailored to the needs of mail handling.
meillo@76	346 Instead it was one large program that had several MH tools built in.
meillo@76	347 This conflicts with the major feature of MH of being a tool chest.
meillo@76	348 .Pn msh 's
meillo@76	349 main use case had been accessing Bulletin Boards, which have seized to
meillo@62	350 be popular.
meillo@62	351 .P
meillo@62	352 Removing
meillo@58	353 .Pn msh ,
meillo@76	354 together with the truly archaic code relicts
meillo@58	355 .Pn vmh
meillo@58	356 and
meillo@58	357 .Pn wmh ,
meillo@62	358 saved more than 7\\|000 lines of C code \(en
meillo@66	359 about 15\\|% of the project's original source code amount.
meillo@100	360 Having less code \(en with equal readability, of course \(en
meillo@76	361 for the same functionality is an advantage.
meillo@63	362 Less code means less bugs and less maintenance work.
meillo@76	363 As
meillo@63	364 .Pn rcvtty
meillo@63	365 and
meillo@63	366 .Pn msgchk
meillo@87	367 are assumed to be rarely used and can be implemented in different ways,
meillo@87	368 why should one keep them?
meillo@125	369 Removing them streamlines mmh.
meillo@63	370 .Pn viamail 's
meillo@63	371 use case is now partly obsolete and partly covered by
meillo@63	372 .Pn forw ,
meillo@76	373 hence there's no reason to still maintain it.
meillo@63	374 .Pn conflict
meillo@76	375 is not related to the mail client, and
meillo@63	376 .Pn msh
meillo@63	377 conflicts with the basic concept of MH.
meillo@76	378 Theses two tools might still be useful, but they should not be part of mmh.
meillo@63	379 .P
meillo@76	380 Finally, there's
meillo@76	381 .Pn slocal .
meillo@76	382 .Pn slocal
meillo@76	383 is an MDA and thus not directly MUA-related.
meillo@100	384 It should be removed from mmh, because including it conflicts with
meillo@100	385 the idea that mmh is a MUA only.
meillo@100	386 .Pn slocal
meillo@100	387 should rather become a separate project.
meillo@87	388 However,
meillo@76	389 .Pn slocal
meillo@76	390 provides rule-based processing of messages, like filing them into
meillo@76	391 different folders, which is otherwise not available in mmh.
meillo@87	392 Although
meillo@76	393 .Pn slocal
meillo@87	394 does neither pull in dependencies nor does it include a separate
meillo@100	395 technical area (cf. Sec. XXX), still,
meillo@100	396 it accounts for about 1\\|000 lines of code that need to be maintained.
meillo@76	397 As
meillo@76	398 .Pn slocal
meillo@76	399 is almost self-standing, it should be split off into a separate project.
meillo@76	400 This would cut the strong connection between the MUA mmh and the MDA
meillo@76	401 .Pn slocal .
meillo@87	402 For anyone not using MH,
meillo@87	403 .Pn slocal
meillo@87	404 would become yet another independent MDA, like
meillo@87	405 .I procmail .
meillo@100	406 Then
meillo@87	407 .Pn slocal
meillo@100	408 could be installed without the complete MH system.
meillo@76	409 Likewise, mmh users could decide to use
meillo@76	410 .I procmail
meillo@87	411 without having a second, unused MDA,
meillo@87	412 .Pn slocal ,
meillo@76	413 installed.
meillo@100	414 That appears to be conceptionally the best solution.
meillo@76	415 Yet,
meillo@76	416 .Pn slocal
meillo@87	417 is not split off.
meillo@100	418 I defer the decision over
meillo@78	419 .Pn slocal
meillo@100	420 in need for deeper investigation.
meillo@100	421 In the meanwhile, it remains part of mmh.
meillo@100	422 That does not hurt because
meillo@100	423 .Pn slocal
meillo@100	424 is unrelated to the rest of the project.
meillo@0	425
meillo@58	426
meillo@133	427
meillo@134	428 .H2 "Displaying Messages
meillo@131	429 .P
meillo@133	430 Since the very beginning, already in the first concept paper,
meillo@58	431 .Pn show
meillo@62	432 had been MH's message display program.
meillo@58	433 .Pn show
meillo@76	434 mapped message numbers and sequences to files and invoked
meillo@58	435 .Pn mhl
meillo@89	436 to have the files formatted.
meillo@88	437 With MIME, this approach wasn't sufficient anymore.
meillo@100	438 MIME messages can consist of multiple parts. Some parts are not
meillo@100	439 directly displayable and text content might be encoded in
meillo@58	440 foreign charsets.
meillo@58	441 .Pn show 's
meillo@76	442 understanding of messages and
meillo@58	443 .Pn mhl 's
meillo@88	444 display capabilities couldn't cope with the task any longer.
meillo@62	445 .P
meillo@88	446 Instead of extending these tools, additional tools were written from
meillo@88	447 scratch and added to the MH tool chest.
meillo@88	448 Doing so is encouraged by the tool chest approach.
meillo@88	449 Modular design is a great advantage for extending a system,
meillo@88	450 as new tools can be added without interfering with existing ones.
meillo@62	451 First, the new MIME features were added in form of the single program
meillo@58	452 .Pn mhn .
meillo@58	453 The command
meillo@82	454 .Cl "mhn -show 42
meillo@58	455 would show the MIME message numbered 42.
meillo@58	456 With the 1.0 release of nmh in February 1999, Richard Coleman finished
meillo@58	457 the split of
meillo@58	458 .Pn mhn
meillo@88	459 into a set of specialized tools, which together covered the
meillo@88	460 multiple aspects of MIME.
meillo@88	461 One of them was
meillo@69	462 .Pn mhshow ,
meillo@88	463 which replaced
meillo@88	464 .Cl "mhn -show" .
meillo@88	465 It was capable of displaying MIME messages appropriately.
meillo@62	466 .P
meillo@88	467 From then on, two message display tools were part of nmh,
meillo@76	468 .Pn show
meillo@76	469 and
meillo@76	470 .Pn mhshow .
meillo@88	471 To ease the life of users,
meillo@69	472 .Pn show
meillo@69	473 was extended to automatically hand the job over to
meillo@69	474 .Pn mhshow
meillo@69	475 if displaying the message would be beyond
meillo@69	476 .Pn show 's
meillo@69	477 abilities.
meillo@88	478 In consequence, the user would simply invoke
meillo@69	479 .Pn show
meillo@69	480 (possibly through
meillo@69	481 .Pn next
meillo@69	482 or
meillo@69	483 .Pn prev )
meillo@69	484 and get the message printed with either
meillo@69	485 .Pn show
meillo@69	486 or
meillo@69	487 .Pn mhshow ,
meillo@69	488 whatever was more appropriate.
meillo@69	489 .P
meillo@69	490 Having two similar tools for essentially the same task is redundant.
meillo@88	491 Usually,
meillo@88	492 users wouldn't distinguish between
meillo@88	493 .Pn show
meillo@88	494 and
meillo@88	495 .Pn mhshow
meillo@88	496 in their daily mail reading.
meillo@88	497 Having two separate display programs was therefore mainly unnecessary
meillo@88	498 from a user's point of view.
meillo@88	499 Besides, the development of both programs needed to be in sync,
meillo@76	500 to ensure that the programs behaved in a similar way,
meillo@76	501 because they were used like a single tool.
meillo@76	502 Different behavior would have surprised the user.
meillo@69	503 .P
meillo@69	504 Today, non-MIME messages are rather seen to be a special case of
meillo@100	505 MIME messages, although it is the other way round.
meillo@69	506 As
meillo@69	507 .Pn mhshow
meillo@88	508 had already be able to display non-MIME messages, it appeared natural
meillo@69	509 to drop
meillo@69	510 .Pn show
meillo@69	511 in favor of using
meillo@69	512 .Pn mhshow
meillo@69	513 exclusively.
meillo@88	514 .Ci 4c1efddfd499300c7e74263e57d8aa137e84c853
meillo@88	515 Removing
meillo@88	516 .Pn show
meillo@88	517 is no loss in function, because functionally
meillo@88	518 .Pn mhshow
meillo@88	519 covers it completely.
meillo@88	520 The old behavior of
meillo@88	521 .Pn show
meillo@88	522 can still be emulated with the simple command line:
meillo@88	523 .VS
meillo@88	524 mhl `mhpath c`
meillo@88	525 VE
meillo@88	526 .P
meillo@76	527 For convenience,
meillo@76	528 .Pn mhshow
meillo@88	529 was renamed to
meillo@88	530 .Pn show
meillo@88	531 after
meillo@88	532 .Pn show
meillo@88	533 was gone.
meillo@88	534 It is clear that such a rename may confuse future developers when
meillo@88	535 trying to understand the history.
meillo@88	536 Nevertheless, I consider the convenience on the user's side,
meillo@88	537 to call
meillo@88	538 .Pn show
meillo@88	539 when they want a message to be displayed, to outweigh the inconvenience
meillo@88	540 on the developer's side when understanding the project history.
meillo@69	541 .P
meillo@88	542 To prepare for the transition,
meillo@69	543 .Pn mhshow
meillo@69	544 was reworked to behave more like
meillo@69	545 .Pn show
meillo@69	546 first.
meillo@88	547 (cf. Sec. XXX)
meillo@88	548 Once the tools behaved more alike, the replacing appeared to be
meillo@88	549 even more natural.
meillo@88	550 Today, mmh's new
meillo@69	551 .Pn show
meillo@88	552 became the one single message display program again, with the difference
meillo@88	553 that today it handles MIME messages as well as non-MIME messages.
meillo@88	554 The outcome of the transition is one program less to maintain,
meillo@88	555 no second display program for users to deal with,
meillo@88	556 and less system complexity.
meillo@69	557 .P
meillo@88	558 Still, removing the old
meillo@69	559 .Pn show
meillo@88	560 hurts in one regard: It had been such a simple program.
meillo@88	561 Its lean elegance is missing to the new
meillo@69	562 .Pn show .
meillo@88	563 But there is no chance;
meillo@88	564 supporting MIME demands for higher essential complexity.
meillo@58	565
meillo@134	566 .ig
meillo@134	567 XXX
meillo@134	568 Consider including text on scan listings here
meillo@58	569
meillo@134	570 Scan listings shall not contain body content. Hence, removed this feature.
meillo@134	571 Scan listings shall operator on message headers and non-message information
meillo@134	572 only. Displaying the beginning of the body complicates everything too much.
meillo@134	573 That's no surprise, because it's something completely different. If you
meillo@134	574 want to examine the body, then use show(1)/mhshow(1).
meillo@134	575 Changed the default scan formats accordingly.
meillo@134	576 .Ci 70b2643e0da8485174480c644ad9785c84f5bff4
meillo@134	577 ..
meillo@131	578
meillo@131	579
meillo@131	580
meillo@133	581
meillo@100	582 .H2 "Configure Options
meillo@58	583 .P
meillo@76	584 Customization is a double-edged sword.
meillo@76	585 It allows better suiting setups, but not for free.
meillo@76	586 There is the cost of code complexity to be able to customize.
meillo@76	587 There is the cost of less tested setups, because there are
meillo@72	588 more possible setups and especially corner-cases.
meillo@76	589 And, there is the cost of choice itself.
meillo@76	590 The code complexity directly affects the developers.
meillo@72	591 Less tested code affects both, users and developers.
meillo@76	592 The problem of choice affects the users, for once by having to
meillo@100	593 choose, but also by more complex interfaces that require more documentation.
meillo@72	594 Whenever options add little advantages, they should be considered for
meillo@72	595 removal.
meillo@72	596 I have reduced the number of project-specific configure options from
meillo@72	597 fifteen to three.
meillo@74	598
meillo@76	599 .U3 "Mail Transfer Facilities
meillo@74	600 .P
meillo@85	601 With the removal of the mail transfer facilities five configure
meillo@85	602 options vanished:
meillo@85	603 .P
meillo@85	604 The switches
meillo@85	605 .Sw --with-tls
meillo@85	606 and
meillo@85	607 .Sw --with-cyrus-sasl
meillo@89	608 had activated the support for transfer encryption and authentication.
meillo@85	609 This is not needed anymore.
meillo@85	610 .Ci fecd5d34f65597a4dfa16aeabea7d74b191532c3
meillo@85	611 .Ci 156d35f6425bea4c1ed3c4c79783dc613379c65b
meillo@85	612 .P
meillo@85	613 The configure switch
meillo@85	614 .Sw --enable-pop
meillo@85	615 activated the message retrieval facility.
meillo@85	616 The code area that would be conditionally compiled in for TLS and SASL
meillo@85	617 support had been small.
meillo@85	618 The conditionally compiled code area for POP support had been much larger.
meillo@85	619 Whereas the code base changes would only slightly change on toggling
meillo@85	620 TLS or SASL support, it changed much on toggling POP support.
meillo@85	621 The changes in the code base could hardly be overviewed.
meillo@85	622 By having POP support togglable a second code base had been created,
meillo@85	623 one that needed to be tested.
meillo@85	624 This situation is basically similar for the conditional TLS and SASL
meillo@85	625 code, but there the changes are minor and can yet be overviewed.
meillo@85	626 Still, conditional compilation of a code base creates variations
meillo@85	627 of the original program.
meillo@85	628 More variations require more testing and maintenance work.
meillo@85	629 .P
meillo@85	630 Two other options only specified default configuration values:
meillo@100	631 .Sw --with-mts
meillo@100	632 defined the default transport service, either
meillo@100	633 .Ar smtp
meillo@100	634 or
meillo@100	635 .Ar sendmail .
meillo@85	636 In mmh this fixed to
meillo@85	637 .Ar sendmail .
meillo@85	638 .Ci f6aa95b724fd8c791164abe7ee5468bf5c34f226
meillo@85	639 With
meillo@100	640 .Sw --with-smtpservers
meillo@85	641 default SMTP servers for the
meillo@85	642 .Ar smtp
meillo@85	643 transport service could be specified.
meillo@72	644 .Ci 128545e06224233b7e91fc4c83f8830252fe16c9
meillo@85	645 Both of them became irrelevant.
meillo@72	646
meillo@74	647 .U3 "Backup Prefix
meillo@74	648 .P
meillo@76	649 The backup prefix is the string that was prepended to message
meillo@76	650 filenames to tag them as deleted.
meillo@76	651 By default it had been the comma character `\f(CW,\fP'.
meillo@78	652 In July 2000, Kimmo Suominen introduced
meillo@78	653 the configure option
meillo@78	654 .Sw --with-hash-backup
meillo@78	655 to change the default to the hash symbol `\f(CW#\fP'.
meillo@78	656 The choice was probably personal preference, because first, the
meillo@78	657 option was named
meillo@78	658 .Sw --with-backup-prefix.
meillo@78	659 and had the prefix symbol as argument.
meillo@100	660 But giving the hash symbol as argument caused too many problems
meillo@100	661 for Autoconf,
meillo@100	662 thus the option was limited to use the hash symbol as the default prefix.
meillo@100	663 This supports the assumption, that the choice for the hash was
meillo@100	664 personal preference only.
meillo@100	665 Being related or not, words that start with the hash symbol
meillo@78	666 introduce a comment in the Unix shell.
meillo@72	667 Thus, the command line
meillo@72	668 .Cl "rm #13 #15
meillo@72	669 calls
meillo@72	670 .Pn rm
meillo@72	671 without arguments because the first hash symbol starts the comment
meillo@72	672 that reaches until the end of the line.
meillo@72	673 To delete the backup files,
meillo@72	674 .Cl "rm ./#13 ./#15"
meillo@72	675 needs to be used.
meillo@100	676 Using the hash as backup prefix can be seen as a precaution against
meillo@78	677 data loss.
meillo@78	678 .P
meillo@72	679 I removed the configure option but added the profile entry
meillo@72	680 .Pe backup-prefix ,
meillo@72	681 which allows to specify an arbitrary string as backup prefix.
meillo@72	682 .Ci 6c40d481d661d532dd527eaf34cebb6d3f8ed086
meillo@76	683 Profile entries are the common method to change mmh's behavior.
meillo@76	684 This change did not remove the choice but moved it to a location where
meillo@72	685 it suited better.
meillo@76	686 .P
meillo@78	687 Eventually, however, the new trash folder concept
meillo@78	688 .Cf "Sec. XXX
meillo@78	689 obsoleted the concept of the backup prefix completely.
meillo@78	690 .Ci 8edc5aaf86f9f77124664f6801bc6c6cdf258173
meillo@133	691 .Ci ca0b3e830b86700d9e5e31b1784de2bdcaf58fc5
meillo@133	692
meillo@76	693
meillo@76	694 .U3 "Editor and Pager
meillo@74	695 .P
meillo@74	696 The two configure options
meillo@74	697 .CW --with-editor=EDITOR
meillo@74	698 .CW --with-pager=PAGER
meillo@74	699 were used to specify the default editor and pager at configure time.
meillo@109	700 Doing so at configure time made sense in the eighties,
meillo@76	701 when the set of available editors and pagers varied much across
meillo@76	702 different systems.
meillo@89	703 Today, the situation is more homogeneous.
meillo@74	704 The programs
meillo@74	705 .Pn vi
meillo@74	706 and
meillo@74	707 .Pn more
meillo@76	708 can be expected to be available on every Unix system,
meillo@74	709 as they are specified by POSIX since two decades.
meillo@74	710 (The specifications for
meillo@74	711 .Pn vi
meillo@74	712 and
meillo@74	713 .Pn more
meillo@74	714 appeared in
meillo@74	715 .[
meillo@74	716 posix 1987
meillo@74	717 .]
meillo@74	718 and,
meillo@74	719 .[
meillo@74	720 posix 1992
meillo@74	721 .]
meillo@74	722 respectively.)
meillo@74	723 As a first step, these two tools were hard-coded as defaults.
meillo@74	724 .Ci 5d43a99db70c12a673028c7758c20cbe3e13ef5f
meillo@74	725 Not changed were the
meillo@74	726 .Pe editor
meillo@74	727 and
meillo@74	728 .Pe moreproc
meillo@76	729 profile entries, which allowed the user to override the system defaults.
meillo@74	730 Later, the concept was reworked to respect the standard environment
meillo@74	731 variables
meillo@74	732 .Ev VISUAL
meillo@74	733 and
meillo@74	734 .Ev PAGER
meillo@76	735 if they are set.
meillo@74	736 Today, mmh determines the editor to use in the following order,
meillo@74	737 taking the first available and non-empty item:
meillo@74	738 .IP (1)
meillo@74	739 Environment variable
meillo@74	740 .Ev MMHEDITOR
meillo@74	741 .IP (2)
meillo@74	742 Profile entry
meillo@74	743 .Pe Editor
meillo@74	744 .IP (3)
meillo@74	745 Environment variable
meillo@74	746 .Ev VISUAL
meillo@74	747 .IP (4)
meillo@74	748 Environment variable
meillo@74	749 .Ev EDITOR
meillo@74	750 .IP (5)
meillo@74	751 Command
meillo@74	752 .Pn vi .
meillo@74	753 .P
meillo@76	754 .Ci f85f4b7ae62e3d05a945dcd46ead51f0a2a89a9b
meillo@76	755 .P
meillo@89	756 The pager to use is determined in a similar order,
meillo@74	757 also taking the first available and non-empty item:
meillo@74	758 .IP (1)
meillo@74	759 Environment variable
meillo@74	760 .Ev MMHPAGER
meillo@74	761 .IP (2)
meillo@74	762 Profile entry
meillo@74	763 .Pe Pager
meillo@74	764 (replaces
meillo@74	765 .Pe moreproc )
meillo@74	766 .IP (3)
meillo@74	767 Environment variable
meillo@74	768 .Ev PAGER
meillo@74	769 .IP (4)
meillo@74	770 Command
meillo@74	771 .Pn more .
meillo@74	772 .P
meillo@74	773 .Ci 0c4214ea2aec6497d0d67b436bbee9bc1d225f1e
meillo@74	774 .P
meillo@76	775 By respecting the
meillo@74	776 .Ev VISUAL /\c
meillo@74	777 .Ev EDITOR
meillo@74	778 and
meillo@74	779 .Ev PAGER
meillo@76	780 environment variables,
meillo@76	781 the new behavior confirms better to the common style on Unix systems.
meillo@76	782 Additionally, the new approach is more uniform and clearer to users.
meillo@72	783
meillo@72	784
meillo@76	785 .U3 "ndbm
meillo@72	786 .P
meillo@74	787 .Pn slocal
meillo@78	788 used to depend on
meillo@74	789 .I ndbm ,
meillo@74	790 a database library.
meillo@76	791 The database is used to store the `\fLMessage-ID\fP's of all
meillo@76	792 messages delivered.
meillo@74	793 This enables
meillo@74	794 .Pn slocal
meillo@74	795 to suppress delivering the same message to the same user twice.
meillo@74	796 (This features was enabled by the
meillo@74	797 .Sw -suppressdup
meillo@74	798 switch.)
meillo@74	799 .P
meillo@100	800 A variety of versions of the database library exist.
meillo@78	801 .[
meillo@78	802 wolter unix incompat notes dbm
meillo@78	803 .]
meillo@74	804 Complicated autoconf code was needed to detect them correctly.
meillo@74	805 Further more, the configure switches
meillo@74	806 .Sw --with-ndbm=ARG
meillo@74	807 and
meillo@74	808 .Sw --with-ndbmheader=ARG
meillo@74	809 were added to help with difficult setups that would
meillo@78	810 not be detected automatically or correctly.
meillo@74	811 .P
meillo@74	812 By removing the suppress duplicates feature of
meillo@74	813 .Pn slocal ,
meillo@74	814 the dependency on
meillo@74	815 .I ndbm
meillo@78	816 vanished and 120 lines of complex autoconf code could be saved.
meillo@74	817 .Ci ecd6d6a20cb7a1507e3a20d6c4cb3a1cf14c6bbf
meillo@89	818 The change removed functionality too, but that is minor to the
meillo@78	819 improvement by dropping the dependency and the complex autoconf code.
meillo@72	820
meillo@74	821 .U3 "mh-e Support
meillo@72	822 .P
meillo@74	823 The configure option
meillo@74	824 .Sw --disable-mhe
meillo@74	825 was removed when the mh-e support was reworked.
meillo@74	826 Mh-e is the Emacs front-end to MH.
meillo@76	827 It requires MH to provide minor additional functions.
meillo@76	828 The
meillo@76	829 .Sw --disable-mhe
meillo@76	830 configure option could switch these extensions off.
meillo@76	831 After removing the support for old versions of mh-e,
meillo@74	832 only the
meillo@74	833 .Sw -build
meillo@76	834 switches of
meillo@74	835 .Pn forw
meillo@74	836 and
meillo@74	837 .Pn repl
meillo@76	838 are left to be mh-e extensions.
meillo@76	839 They are now always built in because they add little code and complexity.
meillo@76	840 In consequence, the
meillo@74	841 .Sw --disable-mhe
meillo@76	842 configure option was removed
meillo@72	843 .Ci a7ce7b4a580d77b6c2c4d980812beb589aa4c643
meillo@74	844 Removing the option removed a second code setup that would have
meillo@74	845 needed to be tested.
meillo@76	846 This change was first done in nmh and thereafter merged into mmh.
meillo@76	847 .P
meillo@76	848 The interface changes in mmh require mh-e to be adjusted in order
meillo@76	849 to be able to use mmh as back-end.
meillo@76	850 This will require minor changes to mh-e, but removing the
meillo@76	851 .Sw -build
meillo@76	852 switches would require more rework.
meillo@72	853
meillo@74	854 .U3 "Masquerading
meillo@72	855 .P
meillo@74	856 The configure option
meillo@74	857 .Sw --enable-masquerade
meillo@76	858 could take up to three arguments:
meillo@76	859 `draft_from', `mmailid', and `username_extension'.
meillo@74	860 They activated different types of address masquerading.
meillo@74	861 All of them were implemented in the SMTP-speaking
meillo@74	862 .Pn post
meillo@76	863 command, which provided an MSA.
meillo@76	864 Address masquerading is an MTA's task and mmh does not cover
meillo@76	865 this field anymore.
meillo@76	866 Hence, true masquerading needs to be implemented in the external MTA.
meillo@74	867 .P
meillo@74	868 The
meillo@74	869 .I mmailid
meillo@74	870 masquerading type is the oldest one of the three and the only one
meillo@74	871 available in the original MH.
meillo@74	872 It provided a
meillo@74	873 .I username
meillo@74	874 to
meillo@74	875 .I fakeusername
meillo@76	876 mapping, based on the password file's GECOS field.
meillo@74	877 The man page
meillo@74	878 .Mp mh-tailor(5)
meillo@74	879 described the use case as being the following:
meillo@98	880 .QS
meillo@74	881 This is useful if you want the messages you send to always
meillo@74	882 appear to come from the name of an MTA alias rather than your
meillo@74	883 actual account name. For instance, many organizations set up
meillo@74	884 `First.Last' sendmail aliases for all users. If this is
meillo@74	885 the case, the GECOS field for each user should look like:
meillo@74	886 ``First [Middle] Last <First.Last>''
meillo@98	887 .QE
meillo@74	888 .P
meillo@74	889 As mmh sends outgoing mail via the local MTA only,
meillo@76	890 the best location to do such global rewrites is there.
meillo@74	891 Besides, the MTA is conceptionally the right location because it
meillo@74	892 does the reverse mapping for incoming mail (aliasing), too.
meillo@76	893 Further more, masquerading set up there is readily available for all
meillo@74	894 mail software on the system.
meillo@76	895 Hence, mmailid masquerading was removed.
meillo@74	896 .Ci 0836c8000ccb34b59410ef1c15b1b7feac70ce5f
meillo@74	897 .P
meillo@74	898 The
meillo@74	899 .I username_extension
meillo@76	900 masquerading type did not replace the username but would append a suffix,
meillo@76	901 specified by the
meillo@74	902 .Ev USERNAME_EXTENSION
meillo@76	903 environment variable, to it.
meillo@76	904 This provided support for the
meillo@74	905 .I user-extension
meillo@74	906 feature of qmail and the similar
meillo@74	907 .I "plussed user
meillo@74	908 processing of sendmail.
meillo@74	909 The decision to remove this username_extension masquerading was
meillo@74	910 motivated by the fact that
meillo@74	911 .Pn spost
meillo@76	912 hadn't supported it already.
meillo@76	913 .Ci 2abae0bfd0ad5bf898461e50aa4b466d641f23d9
meillo@76	914 Username extensions are possible in mmh, but less convenient to use.
meillo@76	915 .\" XXX format file %(getenv USERNAME_EXTENSION)
meillo@74	916 .P
meillo@74	917 The
meillo@74	918 .I draft_from
meillo@74	919 masquerading type instructed
meillo@74	920 .Pn post
meillo@84	921 to use the value of the
meillo@84	922 .Hd From
meillo@84	923 header field as SMTP envelope sender.
meillo@76	924 Sender addresses could be replaced completely.
meillo@74	925 .Ci b14ea6073f77b4359aaf3fddd0e105989db9
meillo@76	926 Mmh offers a kind of masquerading similar in effect, but
meillo@74	927 with technical differences.
meillo@76	928 As mmh does not transfer messages itself, the local MTA has final control
meillo@76	929 over the sender's address. Any masquerading mmh introduces may be reverted
meillo@76	930 by the MTA.
meillo@76	931 In times of pedantic spam checking, an MTA will take care to use
meillo@76	932 sensible envelope sender addresses to keep its own reputation up.
meillo@84	933 Nonetheless, the MUA can set the
meillo@84	934 .Hd From
meillo@84	935 header field and thereby propose
meillo@76	936 a sender address to the MTA.
meillo@74	937 The MTA may then decide to take that one or generate the canonical sender
meillo@74	938 address for use as envelope sender address.
meillo@74	939 .P
meillo@74	940 In mmh, the MTA will always extract the recipient and sender from the
meillo@84	941 message header (\c
meillo@74	942 .Pn sendmail 's
meillo@74	943 .Sw -t
meillo@74	944 switch).
meillo@84	945 The
meillo@84	946 .Hd From
meillo@84	947 header field of the draft may be set arbitrary by the user.
meillo@74	948 If it is missing, the canonical sender address will be generated by the MTA.
meillo@74	949
meillo@74	950 .U3 "Remaining Options
meillo@74	951 .P
meillo@74	952 Two configure options remain in mmh.
meillo@74	953 One is the locking method to use:
meillo@74	954 .Sw --with-locking=[dot\|fcntl\|flock\|lockf] .
meillo@76	955 The idea of removing all methods except the portable dot locking
meillo@76	956 and having that one as the default is appealing, but this change
meillo@76	957 requires deeper technical investigation into the topic.
meillo@76	958 The other option,
meillo@74	959 .Sw --enable-debug ,
meillo@74	960 compiles the programs with debugging symbols and does not strip them.
meillo@74	961 This option is likely to stay.
meillo@72	962
meillo@72	963
meillo@58	964
meillo@63	965
meillo@100	966 .H2 "Command Line Switches
meillo@58	967 .P
meillo@93	968 The command line switches of MH tools follow the X Window style.
meillo@93	969 They are words, introduced by a single dash.
meillo@93	970 For example:
meillo@93	971 .Cl "-truncate" .
meillo@93	972 Every program in mmh has two generic switches:
meillo@93	973 .Sw -help ,
meillo@93	974 to print a short message on how to use the program, and
meillo@93	975 .Sw -Version ,
meillo@93	976 to tell what version of mmh the program belongs to.
meillo@93	977 .P
meillo@93	978 Switches change the behavior of programs.
meillo@93	979 Programs that do one thing in one way require no switches.
meillo@93	980 In most cases, doing something in exactly one way is too limiting.
meillo@97	981 If there is basically one task to accomplish, but it should be done
meillo@93	982 in various ways, switches are a good approach to alter the behavior
meillo@93	983 of a program.
meillo@93	984 Changing the behavior of programs provides flexibility and customization
meillo@97	985 to users, but at the same time it complicates the code, documentation and
meillo@93	986 usage of the program.
meillo@97	987 .\" XXX: Ref
meillo@93	988 Therefore, the number of switches should be kept small.
meillo@93	989 A small set of well-chosen switches does no harm.
meillo@93	990 But usually, the number of switches increases over time.
meillo@93	991 Already in 1985, Rose and Romine have identified this as a major
meillo@93	992 problem of MH:
meillo@93	993 .[ [
meillo@93	994 rose romine real work
meillo@93	995 .], p. 12]
meillo@98	996 .QS
meillo@93	997 A complaint often heard about systems which undergo substantial development
meillo@93	998 by many people over a number of years, is that more and more options are
meillo@93	999 introduced which add little to the functionality but greatly increase the
meillo@93	1000 amount of information a user needs to know in order to get useful work done.
meillo@93	1001 This is usually referred to as creeping featurism.
meillo@93	1002 .QP
meillo@93	1003 Unfortunately MH, having undergone six years of off-and-on development by
meillo@93	1004 ten or so well-meaning programmers (the present authors included),
meillo@93	1005 suffers mightily from this.
meillo@98	1006 .QE
meillo@93	1007 .P
meillo@97	1008 Being reluctant to adding new switches \(en or `options',
meillo@97	1009 as Rose and Romine call them \(en is one part of a counter-action,
meillo@97	1010 the other part is removing hardly used switches.
meillo@97	1011 Nmh's tools had lots of switches already implemented,
meillo@97	1012 hence, cleaning up by removing some of them was the more important part
meillo@97	1013 of the counter-action.
meillo@93	1014 Removing existing functionality is always difficult because it
meillo@93	1015 breaks programs that use these functions.
meillo@93	1016 Also, for every obsolete feature, there'll always be someone who still
meillo@93	1017 uses it and thus opposes its removal.
meillo@93	1018 This puts the developer into the position,
meillo@93	1019 where sensible improvements to style are regarded as destructive acts.
meillo@97	1020 Yet, living with the featurism is far worse, in my eyes, because
meillo@97	1021 future needs will demand adding further features,
meillo@93	1022 worsening the situation more and more.
meillo@93	1023 Rose and Romine added in a footnote,
meillo@93	1024 ``[...]
meillo@93	1025 .Pn send
meillo@93	1026 will no doubt acquire an endless number of switches in the years to come.''
meillo@97	1027 Although clearly humorous, the comment points to the nature of the problem.
meillo@97	1028 Refusing to add any new switches would encounter the problem at its root,
meillo@97	1029 but this is not practical.
meillo@97	1030 New needs will require new switches and it would be unwise to block
meillo@97	1031 them strictly.
meillo@97	1032 Nevertheless, removing obsolete switches still is an effective approach
meillo@97	1033 to deal with the problem.
meillo@97	1034 Working on an experimental branch without an established user base,
meillo@97	1035 eased my work because I did not offend users when I removed existing
meillo@110	1036 functions.
meillo@93	1037 .P
meillo@93	1038 Rose and Romine counted 24 visible and 9 more hidden switches for
meillo@93	1039 .Pn send .
meillo@97	1040 In nmh, they increased up to 32 visible and 12 hidden ones.
meillo@97	1041 At the time of writing, no more than 7 visible switches and 1 hidden switch
meillo@97	1042 have remained in mmh's
meillo@97	1043 .Pn send .
meillo@97	1044 (These numbers include two generic switches, help and version.)
meillo@93	1045 .P
meillo@97	1046 Fig. XXX
meillo@93	1047 .\" XXX Ref
meillo@97	1048 displays the number of switches for each of the tools that is available
meillo@97	1049 in both, nmh and mmh.
meillo@100	1050 The tools are sorted by the number of switches they had in nmh.
meillo@100	1051 Visible and hidden switches were counted,
meillo@97	1052 but not the generic help and version switches.
meillo@93	1053 Whereas in the beginning of the project, the average tool had 11 switches,
meillo@93	1054 now it has no more than 5 \(en only half as many.
meillo@93	1055 If the `no' switches and similar inverse variant are folded onto
meillo@100	1056 their counter-parts, the average tool had 8 switches in pre-mmh times and
meillo@100	1057 has 4 now.
meillo@93	1058 The total number of functional switches in mmh dropped from 465
meillo@93	1059 to 234.
meillo@58	1060
meillo@93	1061 .KS
meillo@93	1062 .in 1c
meillo@93	1063 .so input/switches.grap
meillo@93	1064 .KE
meillo@58	1065
meillo@93	1066 .P
meillo@93	1067 A part of the switches vanished after functions were removed.
meillo@93	1068 This was the case for network mail transfer, for instance.
meillo@97	1069 Sometimes, however, the work flow was the other way:
meillo@97	1070 I looked through the
meillo@97	1071 .Mp mh-chart (7)
meillo@97	1072 man page to identify the tools with apparently too many switches.
meillo@97	1073 Then considering the value of each of the switches by examining
meillo@97	1074 the tool's man page and source code, aided by recherche and testing.
meillo@97	1075 This way, the removal of functions was suggested by the aim to reduce
meillo@97	1076 the number of switches per command.
meillo@97	1077
meillo@58	1078
meillo@93	1079 .U3 "Draft Folder Facility
meillo@93	1080 .P
meillo@100	1081 A change early in the project was the complete transition from
meillo@93	1082 the single draft message to the draft folder facility.
meillo@97	1083 .Ci 337338b404931f06f0db2119c9e145e8ca5a9860
meillo@109	1084 The draft folder facility was introduced in the mid-eighties, when
meillo@100	1085 Rose and Romine called it a ``relatively new feature''.
meillo@93	1086 .[
meillo@93	1087 rose romine real work
meillo@93	1088 .]
meillo@110	1089 Since then, the facility had existed but was inactive by default.
meillo@93	1090 The default activation and the related rework of the tools made it
meillo@93	1091 possible to remove the
meillo@93	1092 .Sw -[no]draftfolder ,
meillo@93	1093 and
meillo@93	1094 .Sw -draftmessage
meillo@93	1095 switches from
meillo@93	1096 .Pn comp ,
meillo@93	1097 .Pn repl ,
meillo@93	1098 .Pn forw ,
meillo@93	1099 .Pn dist ,
meillo@93	1100 .Pn whatnow ,
meillo@93	1101 and
meillo@93	1102 .Pn send .
meillo@97	1103 .Ci 337338b404931f06f0db2119c9e145e8ca5a9860
meillo@97	1104 The only flexibility removed with this change is having multiple
meillo@97	1105 draft folders within one profile.
meillo@97	1106 I consider this a theoretical problem only.
meillo@93	1107 In the same go, the
meillo@93	1108 .Sw -draft
meillo@93	1109 switch of
meillo@93	1110 .Pn anno ,
meillo@93	1111 .Pn refile ,
meillo@93	1112 and
meillo@93	1113 .Pn send
meillo@93	1114 was removed.
meillo@93	1115 The special-casing of `the' draft message became irrelevant after
meillo@93	1116 the rework of the draft system.
meillo@93	1117 (See Sec. XXX.)
meillo@95	1118 Equally,
meillo@95	1119 .Pn comp
meillo@95	1120 lost its
meillo@95	1121 .Sw -file
meillo@95	1122 switch.
meillo@95	1123 The draft folder facility, together with the
meillo@95	1124 .Sw -form
meillo@95	1125 switch, are sufficient.
meillo@93	1126
meillo@95	1127
meillo@102	1128 .U3 "In Place Editing
meillo@93	1129 .P
meillo@93	1130 .Pn anno
meillo@93	1131 had the switches
meillo@93	1132 .Sw -[no]inplace
meillo@100	1133 to either annotate the message in place and thus preserve hard links,
meillo@93	1134 or annotate a copy to replace the original message, breaking hard links.
meillo@97	1135 Following the assumption that linked messages should truly be the
meillo@97	1136 same message, and annotating it should not break the link, the
meillo@93	1137 .Sw -[no]inplace
meillo@93	1138 switches were removed and the previous default
meillo@93	1139 .Sw -inplace
meillo@93	1140 was made the only behavior.
meillo@97	1141 .Ci c8195849d2e366c569271abb0f5f60f4ebf0b4d0
meillo@93	1142 The
meillo@93	1143 .Sw -[no]inplace
meillo@93	1144 switches of
meillo@93	1145 .Pn repl ,
meillo@93	1146 .Pn forw ,
meillo@93	1147 and
meillo@93	1148 .Pn dist
meillo@93	1149 could be removed, too, as they were simply passed through to
meillo@93	1150 .Pn anno .
meillo@93	1151 .P
meillo@93	1152 .Pn burst
meillo@93	1153 also had
meillo@93	1154 .Sw -[no]inplace
meillo@95	1155 switches, but with different meaning.
meillo@95	1156 With
meillo@95	1157 .Sw -inplace ,
meillo@95	1158 the digest had been replaced by the table of contents (i.e. the
meillo@110	1159 introduction text) and the burst messages were placed right
meillo@95	1160 after this message, renumbering all following messages.
meillo@95	1161 Also, any trailing text of the digest was lost, though,
meillo@95	1162 in practice, it usually consists of an end-of-digest marker only.
meillo@95	1163 Nontheless, this behavior appeared less elegant than the
meillo@95	1164 .Sw -noinplace
meillo@95	1165 behavior, which already had been the default.
meillo@95	1166 Nmh's
meillo@95	1167 .Mp burst (1)
meillo@95	1168 man page reads:
meillo@95	1169 .sp \n(PDu
meillo@98	1170 .QS
meillo@93	1171 If -noinplace is given, each digest is preserved, no table
meillo@93	1172 of contents is produced, and the messages contained within
meillo@93	1173 the digest are placed at the end of the folder. Other messages
meillo@93	1174 are not tampered with in any way.
meillo@98	1175 .QE
meillo@95	1176 .LP
meillo@93	1177 The decision to drop the
meillo@93	1178 .Sw -inplace
meillo@95	1179 behavior was supported by the code complexity and the possible data loss
meillo@95	1180 it caused.
meillo@93	1181 .Sw -noinplace
meillo@95	1182 was chosen to be the definitive behavior.
meillo@97	1183 .Ci 68a686adeb39223a5e1ad35e4a24890ec053679d
meillo@93	1184
meillo@95	1185
meillo@95	1186 .U3 "Forms and Format Strings
meillo@93	1187 .P
meillo@95	1188 Historically, the tools that had
meillo@95	1189 .Sw -form
meillo@95	1190 switches to supply a form file had
meillo@95	1191 .Sw -format
meillo@95	1192 switches as well to supply the contents of a form file as a string
meillo@95	1193 on the command line directly.
meillo@95	1194 In consequence, the following two lines equaled:
meillo@95	1195 .VS
meillo@95	1196 scan -form scan.mailx
meillo@95	1197 scan -format "`cat .../scan.mailx`"
meillo@95	1198 VE
meillo@95	1199 The
meillo@95	1200 .Sw -format
meillo@95	1201 switches were dropped in favor for extending the
meillo@95	1202 .Sw -form
meillo@95	1203 switches.
meillo@97	1204 .Ci f51956be123db66b00138f80464d06f030dbb88d
meillo@95	1205 If their argument starts with an equal sign (`='),
meillo@95	1206 then the rest of the argument is taken as a format string,
meillo@95	1207 otherwise the arguments is treated as the name of a format file.
meillo@95	1208 Thus, now the following two lines equal:
meillo@95	1209 .VS
meillo@95	1210 scan -form scan.mailx
meillo@95	1211 scan -form "=`cat .../scan.mailx`"
meillo@95	1212 VE
meillo@95	1213 This rework removed the prefix collision between
meillo@95	1214 .Sw -form
meillo@95	1215 and
meillo@95	1216 .Sw -format .
meillo@95	1217 Now, typing
meillo@95	1218 .Sw -fo
meillo@95	1219 suffices to specify form or format string.
meillo@95	1220 .P
meillo@95	1221 The different meaning of
meillo@95	1222 .Sw -format
meillo@95	1223 for
meillo@95	1224 .Pn repl
meillo@95	1225 and
meillo@95	1226 .Pn forw
meillo@95	1227 was removed in mmh.
meillo@95	1228 .Pn forw
meillo@95	1229 was completely switched to MIME-type forwarding, thus removing the
meillo@95	1230 .Sw -[no]format .
meillo@97	1231 .Ci 6e271608b7b9c23771523f88d23a4d3593010cf1
meillo@95	1232 For
meillo@95	1233 .Pn repl ,
meillo@95	1234 the
meillo@95	1235 .Sw -[no]format
meillo@95	1236 switches were reworked to
meillo@95	1237 .Sw -[no]filter
meillo@95	1238 switches.
meillo@97	1239 .Ci 67411b1f95d6ec987b4c732459e1ba8a8ac192c6
meillo@95	1240 The
meillo@95	1241 .Sw -format
meillo@95	1242 switches of
meillo@95	1243 .Pn send
meillo@95	1244 and
meillo@95	1245 .Pn post ,
meillo@95	1246 which had a third meaning,
meillo@95	1247 were removed likewise.
meillo@97	1248 .Ci f3cb7cde0e6f10451b6848678d95860d512224b9
meillo@95	1249 Eventually, the ambiguity of the
meillo@95	1250 .Sw -format
meillo@95	1251 switches was resolved by not anymore having any such switch in mmh.
meillo@95	1252
meillo@95	1253
meillo@95	1254 .U3 "MIME Tools
meillo@95	1255 .P
meillo@95	1256 The MIME tools, which were once part of
meillo@100	1257 .Pn mhn
meillo@100	1258 [sic!],
meillo@95	1259 had several switches that added little practical value to the programs.
meillo@95	1260 The
meillo@95	1261 .Sw -[no]realsize
meillo@95	1262 switches of
meillo@95	1263 .Pn mhbuild
meillo@95	1264 and
meillo@95	1265 .Pn mhlist
meillo@97	1266 were removed, doing real size calculations always now
meillo@97	1267 .Ci 8d8f1c3abc586c005c904e52c4adbfe694d2201c ,
meillo@97	1268 as
meillo@95	1269 ``This provides an accurate count at the expense of a small delay.''
meillo@95	1270 This small delay is not noticable on modern systems.
meillo@95	1271 .P
meillo@95	1272 The
meillo@95	1273 .Sw -[no]check
meillo@95	1274 switches were removed together with the support for
meillo@95	1275 .Hd Content-MD5
meillo@95	1276 header fields.
meillo@95	1277 .[
meillo@95	1278 rfc 1864
meillo@95	1279 .]
meillo@97	1280 .Ci 31dc797eb5178970d68962ca8939da3fd9a8efda
meillo@95	1281 (See Sec. XXX)
meillo@95	1282 .P
meillo@95	1283 The
meillo@95	1284 .Sw -[no]ebcdicsafe
meillo@95	1285 and
meillo@95	1286 .Sw -[no]rfc934mode
meillo@95	1287 switches of
meillo@95	1288 .Pn mhbuild
meillo@95	1289 were removed because they are considered obsolete.
meillo@97	1290 .Ci 01a3480928da485b4d6109d36d751dfa71799d58
meillo@97	1291 .Ci 3363e2624dce0eb8164cf8b3f1ab385c8ff72e88
meillo@95	1292 .P
meillo@95	1293 Content caching of external MIME parts, activated with the
meillo@95	1294 .Sw -rcache
meillo@95	1295 and
meillo@95	1296 .Sw -wcache
meillo@95	1297 switches was completely removed.
meillo@97	1298 .Ci d1fefd9f614e4dc3cda16da6c69133c1b2005269
meillo@97	1299 External MIME parts are rare today, having a caching facility
meillo@96	1300 for them is appears to be unnecessary.
meillo@95	1301 .P
meillo@95	1302 In pre-MIME times,
meillo@95	1303 .Pn mhl
meillo@95	1304 had covered many tasks that are part of MIME handling today.
meillo@95	1305 Therefore,
meillo@95	1306 .Pn mhl
meillo@95	1307 could be simplified to a large extend, reducing the number of its
meillo@95	1308 switches from 21 to 6.
meillo@97	1309 .Ci 350ad6d3542a07639213cf2a4fe524e829c1e7b6
meillo@97	1310 .Ci 0e46503be3c855bddaeae3843e1b659279c35d70
meillo@95	1311
meillo@95	1312
meillo@95	1313
meillo@95	1314
meillo@95	1315 .U3 "Header Printing
meillo@95	1316 .P
meillo@95	1317 .Pn folder 's
meillo@95	1318 data output is self-explaining enough that
meillo@95	1319 displaying the header line makes few sense.
meillo@95	1320 Hence, the
meillo@95	1321 .Sw -[no]header
meillo@95	1322 switch was removed and headers are never printed.
meillo@97	1323 .Ci 601cc73d1fa05ce96faa728f036d6c51b91701c7
meillo@95	1324 .P
meillo@95	1325 In
meillo@95	1326 .Pn mhlist ,
meillo@95	1327 the
meillo@95	1328 .Sw -[no]header
meillo@95	1329 switches were removed, too.
meillo@97	1330 .Ci b24f96523aaf60e44e04a3ffb1d22e69a13a602f
meillo@95	1331 But in this case headers are always printed,
meillo@95	1332 because the output is not self-explaining.
meillo@95	1333 .P
meillo@95	1334 .Pn scan
meillo@95	1335 also had
meillo@95	1336 .Sw -[no]header
meillo@95	1337 switches.
meillo@95	1338 Printing the header had been sensible until the introduction of
meillo@95	1339 format strings made it impossible to display the column headings.
meillo@95	1340 Only the folder name and the current date remained to be printed.
meillo@95	1341 As this information can be perfectly retrieved by
meillo@95	1342 .Pn folder
meillo@95	1343 and
meillo@95	1344 .Pn date ,
meillo@95	1345 consequently, the switches were removed.
meillo@97	1346 .Ci c477dc5d1d03fa6d9a8ab3dd3508c63cbddc044e
meillo@95	1347 .P
meillo@95	1348 By removing all
meillo@95	1349 .Sw -header
meillo@95	1350 switches, the collision with
meillo@95	1351 .Sw -help
meillo@95	1352 on the first two letters was resolved.
meillo@95	1353 Currently,
meillo@95	1354 .Sw -h
meillo@95	1355 evaluates to
meillo@95	1356 .Sw -help
meillo@95	1357 for all tools of mmh.
meillo@95	1358
meillo@95	1359
meillo@139	1360 .U3 "Suppressing Edits or the Invocation of the WhatNow Shell
meillo@95	1361 .P
meillo@95	1362 The
meillo@95	1363 .Sw -noedit
meillo@100	1364 switch of
meillo@95	1365 .Pn comp ,
meillo@95	1366 .Pn repl ,
meillo@95	1367 .Pn forw ,
meillo@95	1368 .Pn dist ,
meillo@95	1369 and
meillo@95	1370 .Pn whatnow
meillo@95	1371 was removed, but it can now be replaced by specifying
meillo@95	1372 .Sw -editor
meillo@95	1373 with an empty argument.
meillo@97	1374 .Ci 75fca31a5b9d5c1a99c74ab14c94438d8852fba9
meillo@95	1375 (Specifying
meillo@95	1376 .Cl "-editor true
meillo@95	1377 is nearly the same, only differing by the previous editor being set.)
meillo@95	1378 .P
meillo@95	1379 The more important change is the removal of the
meillo@95	1380 .Sw -nowhatnowproc
meillo@95	1381 switch.
meillo@97	1382 .Ci ee4f43cf2ef0084ec698e4e87159a94c01940622
meillo@95	1383 This switch had introduced an awkward behavior, as explained in nmh's
meillo@95	1384 man page for
meillo@95	1385 .Mp comp (1):
meillo@98	1386 .QS
meillo@98	1387 The \-editor editor switch indicates the editor to use for
meillo@98	1388 the initial edit. Upon exiting from the editor, comp will
meillo@98	1389 invoke the whatnow program. See whatnow(1) for a discussion
meillo@98	1390 of available options. The invocation of this program can be
meillo@98	1391 inhibited by using the \-nowhatnowproc switch. (In truth of
meillo@98	1392 fact, it is the whatnow program which starts the initial
meillo@98	1393 edit. Hence, \-nowhatnowproc will prevent any edit from
meillo@95	1394 occurring.)
meillo@98	1395 .QE
meillo@95	1396 .P
meillo@95	1397 Effectively, the
meillo@95	1398 .Sw -nowhatnowproc
meillo@100	1399 switch creates only a draft message.
meillo@95	1400 As
meillo@95	1401 .Cl "-whatnowproc true
meillo@95	1402 causes the same behavior, the
meillo@95	1403 .Sw -nowhatnowproc
meillo@95	1404 switch was removed for being redundant.
meillo@100	1405 Likely, the
meillo@95	1406 .Sw -nowhatnowproc
meillo@100	1407 switch was intended to be used by front-ends.
meillo@95	1408
meillo@95	1409
meillo@95	1410
meillo@95	1411 .U3 "Various
meillo@95	1412 .BU
meillo@139	1413 With the removal of MMDF maildrop format support,
meillo@139	1414 .Pn packf
meillo@139	1415 and
meillo@139	1416 .Pn rcvpack
meillo@139	1417 no longer needed their
meillo@139	1418 .Sw -mbox
meillo@139	1419 and
meillo@139	1420 .Sw -mmdf
meillo@139	1421 switches.
meillo@139	1422 .Sw -mbox
meillo@139	1423 is the sole behavior now.
meillo@139	1424 .Ci 3916ab66ad5d183705ac12357621ea8661afd3c0
meillo@139	1425 In the same go,
meillo@139	1426 .Pn packf
meillo@139	1427 and
meillo@139	1428 .Pn rcvpack
meillo@139	1429 were reworked (see Sec. XXX) and their
meillo@139	1430 .Sw -file
meillo@139	1431 switch became unnecessary.
meillo@139	1432 .Ci ca1023716d4c2ab890696f3e41fa0d94267a940e
meillo@139	1433
meillo@139	1434 .BU
meillo@139	1435 Mmh's tools will no longer clear the screen (\c
meillo@139	1436 .Pn scan 's
meillo@139	1437 and
meillo@139	1438 .Pn mhl 's
meillo@139	1439 .Sw -[no]clear
meillo@139	1440 switches
meillo@139	1441 .Ci e57b17343dcb3ff373ef4dd089fbe778f0c7c270
meillo@139	1442 .Ci 943765e7ac5693ae177fd8d2b5a2440e53ce816e ).
meillo@139	1443 Neither will
meillo@139	1444 .Pn mhl
meillo@139	1445 ring the bell (\c
meillo@139	1446 .Sw -[no]bell
meillo@139	1447 .Ci e11983f44e59d8de236affa5b0d0d3067c192e24 )
meillo@139	1448 nor page the output itself (\c
meillo@139	1449 .Sw -length
meillo@139	1450 .Ci 5b9d883db0318ed2b84bb82dee880d7381f99188 ).
meillo@139	1451 Generally, the pager to use is no longer specified with the
meillo@139	1452 .Sw -[no]moreproc
meillo@139	1453 command line switches for
meillo@139	1454 .Pn mhl
meillo@139	1455 and
meillo@139	1456 .Pn show /\c
meillo@139	1457 .Pn mhshow .
meillo@139	1458 .Ci 39e87a75b5c2d3572ec72e717720b44af291e88a
meillo@139	1459
meillo@139	1460 .BU
meillo@96	1461 In order to avoid prefix collisions among switch names, the
meillo@95	1462 .Sw -version
meillo@95	1463 switch was renamed to
meillo@95	1464 .Sw -Version
meillo@95	1465 (with capital `V').
meillo@97	1466 .Ci 32b2354dbaf4bf934936eb5b102a4a3d2fdd209a
meillo@95	1467 Every program has the
meillo@95	1468 .Sw -version
meillo@95	1469 switch but its first three letters collided with the
meillo@95	1470 .Sw -verbose
meillo@95	1471 switch, present in many programs.
meillo@95	1472 The rename solved this problem once for all.
meillo@95	1473 Although this rename breaks a basic interface, having the
meillo@95	1474 .Sw -V
meillo@95	1475 abbreviation to display the version information, isn't all too bad.
meillo@139	1476
meillo@95	1477 .BU
meillo@95	1478 .Sw -[no]preserve
meillo@95	1479 of
meillo@95	1480 .Pn refile
meillo@95	1481 was removed because what use was it anyway?
meillo@98	1482 .QS
meillo@95	1483 Normally when a message is refiled, for each destination
meillo@95	1484 folder it is assigned the number which is one above the current
meillo@95	1485 highest message number in that folder. Use of the
meillo@95	1486 \-preserv [sic!] switch will override this message renaming, and try
meillo@95	1487 to preserve the number of the message. If a conflict for a
meillo@95	1488 particular folder occurs when using the \-preserve switch,
meillo@95	1489 then refile will use the next available message number which
meillo@95	1490 is above the message number you wish to preserve.
meillo@98	1491 .QE
meillo@139	1492
meillo@95	1493 .BU
meillo@95	1494 The removal of the
meillo@95	1495 .Sw -[no]reverse
meillo@95	1496 switches of
meillo@95	1497 .Pn scan
meillo@97	1498 .Ci 8edc5aaf86f9f77124664f6801bc6c6cdf258173
meillo@95	1499 is a bug fix, supported by the comments
meillo@95	1500 ``\-[no]reverse under #ifdef BERK (I really HATE this)''
meillo@95	1501 by Rose and
meillo@95	1502 ``Lists messages in reverse order with the `\-reverse' switch.
meillo@95	1503 This should be considered a bug.'' by Romine in the documentation.
meillo@97	1504 The question remains why neither Rose and Romine had fixed this
meillo@109	1505 bug in the eighties when they wrote these comments nor has anyone
meillo@95	1506 thereafter.
meillo@93	1507
meillo@93	1508
meillo@93	1509 .ig
meillo@93	1510
meillo@95	1511 forw: [no]dashstuffing(mhl)
meillo@93	1512
meillo@95	1513 mhshow: [no]pause [no]serialonly
meillo@93	1514
meillo@93	1515 mhmail: resent queued
meillo@93	1516 inc: snoop, (pop)
meillo@93	1517
meillo@95	1518 mhl: [no]faceproc folder sleep
meillo@95	1519 [no]dashstuffing(forw) digest list volume number issue number
meillo@93	1520
meillo@95	1521 prompter: [no]doteof
meillo@93	1522
meillo@93	1523 refile: [no]preserve [no]unlink [no]rmmproc
meillo@93	1524
meillo@95	1525 send: [no]forward [no]mime [no]msgid
meillo@93	1526 [no]push split [no]unique (sasl) width snoop [no]dashstuffing
meillo@93	1527 attach attachformat
meillo@93	1528 whatnow: (noedit) attach
meillo@93	1529
meillo@93	1530 slocal: [no]suppressdups
meillo@93	1531
meillo@95	1532 spost: [no]filter [no]backup width [no]push idanno
meillo@93	1533 [no]check(whom) whom(whom)
meillo@93	1534
meillo@93	1535 whom: ???
meillo@93	1536
meillo@95	1537 ..
meillo@93	1538
meillo@93	1539
meillo@93	1540 .ig
meillo@93	1541
meillo@93	1542 .P
meillo@93	1543 In the best case, all switches are unambiguous on the first character,
meillo@93	1544 or on the three-letter prefix for the `no' variants.
meillo@96	1545 Reducing switch prefix collisions, shortens the necessary prefix length
meillo@93	1546 the user must type.
meillo@93	1547 Having less switches helps best.
meillo@93	1548
meillo@93	1549 ..
meillo@58	1550
meillo@95	1551
meillo@102	1552 .\" XXX: whatnow prompt commands
meillo@102	1553
meillo@102	1554
meillo@95	1555
meillo@95	1556
meillo@133	1557 .\" --------------------------------------------------------------
meillo@74	1558 .H1 "Modernizing
meillo@102	1559 .P
meillo@118	1560 In the over thirty years of MH's existence, its code base was
meillo@118	1561 extended more and more.
meillo@118	1562 New features entered the project and became alternatives to the
meillo@118	1563 existing behavior.
meillo@118	1564 Relicts from several decades have gathered in the code base,
meillo@118	1565 but seldom obsolete features were dropped.
meillo@118	1566 This section describes the removing of old code
meillo@118	1567 and the modernizing of the default setup.
meillo@118	1568 It focuses on the functional aspect only;
meillo@118	1569 the non-functional aspects of code style are discussed in
meillo@118	1570 .\" FIXME REF
meillo@118	1571 Sec. XXX.
meillo@58	1572
meillo@58	1573
meillo@100	1574 .H2 "Code Relicts
meillo@0	1575 .P
meillo@109	1576 My position to drop obsolete functions of mmh, in order to remove old code,
meillo@104	1577 is much more revolutional than the nmh community likes to have it.
meillo@104	1578 Working on an experimental version, I was able to quickly drop
meillo@104	1579 functionality I considered ancient.
meillo@104	1580 The need for consensus with peers would have slowed this process down.
meillo@104	1581 Without the need to justify my decisions, I was able to rush forward.
meillo@110	1582 In December 2011, Paul Vixie motivated the nmh developers to just
meillo@104	1583 do the work:
meillo@104	1584 .[
meillo@104	1585 paul vixie edginess nmh-workers
meillo@104	1586 .]
meillo@104	1587 .QS
meillo@104	1588 let's stop walking on egg shells with this code base. there's no need to
meillo@104	1589 discuss whether to keep using vfork, just note in [sic!] passing, [...]
meillo@104	1590 we don't need a separate branch for removing vmh
meillo@104	1591 or ridding ourselves of #ifdef's or removing posix replacement functions
meillo@104	1592 or depending on pure ansi/posix "libc".
meillo@104	1593 .QP
meillo@104	1594 these things should each be a day or two of work and the "main branch"
meillo@104	1595 should just be modern. [...]
meillo@104	1596 let's push forward, aggressively.
meillo@104	1597 .QE
meillo@104	1598 .LP
meillo@104	1599 I did so already in the months before.
meillo@104	1600 I pushed forward.
meillo@104	1601 I simply dropped the cruft.
meillo@104	1602 .P
meillo@104	1603 The decision to drop a feature was based on literature research and
meillo@104	1604 careful thinking, but whether having had contact to this particular
meillo@104	1605 feature within my own computer life served as a rule of thumb.
meillo@109	1606 Always, I explained my reasons in the commit messages
meillo@109	1607 in the version control system.
meillo@104	1608 Hence, others can comprehend my view and argue for undoing the change
meillo@104	1609 if I have missed an important aspect.
meillo@109	1610 I was quick in dropping parts.
meillo@109	1611 I rather re-included falsely dropped parts than going a slower pace.
meillo@109	1612 Mmh is experimental work; it required tough decisions.
meillo@12	1613
meillo@102	1614
meillo@104	1615 .U3 "Forking
meillo@12	1616 .P
meillo@109	1617 Being a tool chest, MH creates many processes.
meillo@104	1618 In earlier times
meillo@104	1619 .Fu fork()
meillo@104	1620 had been an expensive system call, because the process's image needed
meillo@104	1621 to be duplicated completely at once.
meillo@109	1622 This was especially painful in the common case when the image gets
meillo@104	1623 replaced by a call to
meillo@104	1624 .Fu exec()
meillo@104	1625 right after having forked the child process.
meillo@104	1626 The
meillo@104	1627 .Fu vfork()
meillo@104	1628 system call was invented to speed up this particular case.
meillo@104	1629 It completely omits the duplication of the image.
meillo@104	1630 On old systems this resulted in significant speed ups.
meillo@104	1631 Therefore MH used
meillo@104	1632 .Fu vfork()
meillo@104	1633 whenever possible.
meillo@12	1634 .P
meillo@104	1635 Modern memory management units support copy-on-write semantics, which make
meillo@104	1636 .Fu fork()
meillo@104	1637 almost as fast as
meillo@104	1638 .Fu vfork() .
meillo@104	1639 The man page of
meillo@104	1640 .Mp vfork (2)
meillo@104	1641 in FreeBSD 8.0 states:
meillo@104	1642 .QS
meillo@104	1643 This system call will be eliminated when proper system sharing mechanisms
meillo@104	1644 are implemented. Users should not depend on the memory sharing semantics
meillo@104	1645 of vfork() as it will, in that case, be made synonymous to fork(2).
meillo@104	1646 .QE
meillo@104	1647 .LP
meillo@104	1648 Vixie supports the removal with the note that ``the last
meillo@104	1649 system on which fork was so slow that an mh user would notice it, was
meillo@104	1650 Eunice. that was 1987''.
meillo@104	1651 .[
meillo@104	1652 nmh-workers vixie edginess
meillo@104	1653 .]
meillo@104	1654 I replaced all calls to
meillo@104	1655 .Fu vfork()
meillo@104	1656 with calls to
meillo@104	1657 .Fu fork() .
meillo@109	1658 .Ci 40821f5c1316e9205a08375e7075909cc9968e7d
meillo@104	1659 .P
meillo@104	1660 Related to the costs of
meillo@104	1661 .Fu fork()
meillo@104	1662 is the probability of its success.
meillo@109	1663 In the eighties, on heavy loaded systems, calls to
meillo@104	1664 .Fu fork()
meillo@104	1665 were prone to failure.
meillo@104	1666 Hence, many of the
meillo@104	1667 .Fu fork()
meillo@104	1668 calls in the code were wrapped into loops to retry the
meillo@104	1669 .Fu fork()
meillo@109	1670 several times, to increase the changes to succeed, eventually.
meillo@109	1671 On modern systems, a failing
meillo@104	1672 .Fu fork()
meillo@109	1673 call is unusual.
meillo@104	1674 Hence, in the rare case when
meillo@104	1675 .Fu fork()
meillo@104	1676 fails, mmh programs simply abort.
meillo@109	1677 .Ci 5fbf37ee68e018998ada61eeab73e035b26834b6
meillo@12	1678
meillo@12	1679
meillo@109	1680 .U3 "Header Fields
meillo@104	1681 .BU
meillo@84	1682 The
meillo@84	1683 .Hd Encrypted
meillo@104	1684 header field was introduced by RFC\\|822,
meillo@109	1685 but already marked as legacy in RFC\\|2822.
meillo@109	1686 Today, OpenPGP provides the basis for standardized exchange of encrypted
meillo@104	1687 messages [RFC\\|4880, RFC\\|3156].
meillo@109	1688 Hence, the support for
meillo@104	1689 .Hd Encrypted
meillo@104	1690 header fields is removed in mmh.
meillo@109	1691 .Ci 064527f7b57ab050e5af13e15ad99aeeab125857
meillo@104	1692 .BU
meillo@84	1693 Native support for
meillo@84	1694 .Hd Face
meillo@104	1695 header fields has been removed, as well.
meillo@109	1696 .Ci 8e5be81f784682822f5e868c1bf3c8624682bd23
meillo@104	1697 This feature is similar to the
meillo@84	1698 .Hd X-Face
meillo@84	1699 header field in its intent,
meillo@21	1700 but takes a different approach to store the image.
meillo@84	1701 Instead of encoding the image data directly into the header field,
meillo@109	1702 it contains the hostname and UDP port where the image
meillo@109	1703 date can be retrieved.
meillo@109	1704 There exists even a third Face system,
meillo@109	1705 which is the successor of
meillo@109	1706 .Hd X-Face ,
meillo@109	1707 although it re-uses the
meillo@104	1708 .Hd Face
meillo@109	1709 header field.
meillo@109	1710 It was invented in 2005 and supports colored PNG images.
meillo@104	1711 None of the Face systems described here is popular today.
meillo@104	1712 Hence, mmh has no direct support for them.
meillo@104	1713 .BU
meillo@104	1714 The
meillo@104	1715 .Hd Content-MD5
meillo@104	1716 header field was introduced by RFC\\|1864.
meillo@104	1717 It provides detection of data corruption during the transfer.
meillo@104	1718 But it can not ensure verbatim end-to-end delivery of the contents
meillo@104	1719 [RFC\\|1864].
meillo@104	1720 The proper approach to verify content integrity in an
meillo@104	1721 end-to-end relationship is the use of digital cryptography.
meillo@104	1722 .\" XXX (RFCs FIXME).
meillo@104	1723 On the other hand, transfer protocols should detect corruption during
meillo@109	1724 the transmission.
meillo@109	1725 The TCP includes a checksum field therefore.
meillo@104	1726 These two approaches in combinations render the
meillo@104	1727 .Hd Content-MD5
meillo@104	1728 header field superfluous.
meillo@109	1729 Not a single one out of 4\\|200 messages from two decades
meillo@109	1730 in an nmh-workers mailing list archive contains a
meillo@104	1731 .Hd Content-MD5
meillo@104	1732 header field.
meillo@104	1733 Neither did any of the 60\\|000 messages in my personal mail storage.
meillo@104	1734 Removing the support for this header field,
meillo@104	1735 removed the last place where MD5 computation was needed.
meillo@109	1736 .Ci 31dc797eb5178970d68962ca8939da3fd9a8efda
meillo@104	1737 Hence, the MD5 code could be removed as well.
meillo@104	1738 Over 500 lines of code vanished by this one change.
meillo@104	1739
meillo@104	1740
meillo@104	1741 .U3 "MMDF maildrop support
meillo@21	1742 .P
meillo@104	1743 This type of format is conceptionally similar to the mbox format,
meillo@139	1744 but uses a different message delimiter (`\fL\\1\\1\\1\\1\fP',
meillo@139	1745 commonly written as `\fL^A^A^A^A\fP', instead of `\fLFrom\0\fP').
meillo@104	1746 Mbox is the de-facto standard maildrop format on Unix,
meillo@109	1747 whereas the MMDF maildrop format became forgotten.
meillo@104	1748 I did drop MMDF maildrop format support.
meillo@109	1749 Mbox is the only packed mailbox format supported in mmh.
meillo@104	1750 .P
meillo@109	1751 The simplifications within the code were moderate.
meillo@109	1752 Mainly, the reading and writing of MMDF mailbox files was removed.
meillo@109	1753 But also, switches of
meillo@109	1754 .Pn packf
meillo@104	1755 and
meillo@109	1756 .Pn rcvpack
meillo@109	1757 could be removed.
meillo@109	1758 .Ci 3916ab66ad5d183705ac12357621ea8661afd3c0
meillo@109	1759 In the message parsing function
meillo@109	1760 .Fn sbr/m_getfld.c ,
meillo@109	1761 knowledge of MMDF packed mail boxes was removed.
meillo@109	1762 .Ci 684ec30d81e1223a282764452f4902ed4ad1c754
meillo@109	1763 Further code structure simplifications may be possible there,
meillo@109	1764 because only one single packed mailbox format is left to be supported.
meillo@104	1765 I have not worked on them yet because
meillo@104	1766 .Fu m_getfld()
meillo@104	1767 is heavily optimized and thus dangerous to touch.
meillo@104	1768 The risk of damaging the intricate workings of the optimized code is
meillo@104	1769 too high.
meillo@104	1770 .\" XXX: move somewhere else
meillo@104	1771 This problem is know to the developers of nmh, too.
meillo@109	1772 They also avoid touching this minefield.
meillo@104	1773
meillo@12	1774
meillo@101	1775 .U3 "Prompter's Control Keys
meillo@20	1776 .P
meillo@20	1777 The program
meillo@20	1778 .Pn prompter
meillo@104	1779 queries the user to fill in a message form.
meillo@104	1780 When used by
meillo@20	1781 .Pn comp
meillo@104	1782 as
meillo@104	1783 .Cl "comp -editor prompter" ,
meillo@20	1784 the resulting behavior is similar to
meillo@20	1785 .Pn mailx .
meillo@51	1786 Apparently,
meillo@20	1787 .Pn prompter
meillo@104	1788 hadn't been touched lately.
meillo@104	1789 Otherwise it's hardly explainable why it
meillo@20	1790 still offered the switches
meillo@84	1791 .Sw -erase
meillo@84	1792 .Ar chr
meillo@20	1793 and
meillo@84	1794 .Sw -kill
meillo@84	1795 .Ar chr
meillo@20	1796 to name the characters for command line editing.
meillo@21	1797 The times when this had been necessary are long time gone.
meillo@20	1798 Today these things work out-of-the-box, and if not, are configured
meillo@20	1799 with the standard tool
meillo@20	1800 .Pn stty .
meillo@104	1801 The switches are removed now
meillo@104	1802 .Ci 0bd9750710cdbab80cfb4036dd87af20afe1552f .
meillo@20	1803
meillo@104	1804
meillo@109	1805 .U3 "Hardcopy Terminal Support
meillo@21	1806 .P
meillo@109	1807 More of a funny anecdote is a check for being connected to a
meillo@109	1808 hardcopy terminal.
meillo@109	1809 It remained in the code until Spring 2012, when I finally removed it
meillo@104	1810 .Ci b7764c4a6b71d37918a97594d866258f154017ca .
meillo@109	1811 I would be truly happy to see such a terminal in action today,
meillo@109	1812 maybe even being able to work on it.
meillo@109	1813 But I fear my chances are null.
meillo@21	1814 .P
meillo@109	1815 The check only prevented a pager to be placed between the printing
meillo@104	1816 program (\c
meillo@104	1817 .Pn mhl )
meillo@104	1818 and the terminal.
meillo@109	1819 In nmh, this could have been ensured statically with the
meillo@104	1820 .Sw -nomoreproc
meillo@109	1821 at the command line, too.
meillo@121	1822 In mmh, setting the profile entry
meillo@104	1823 .Pe Pager
meillo@104	1824 or the environment variable
meillo@104	1825 .Ev PAGER
meillo@104	1826 to
meillo@109	1827 .Pn cat
meillo@109	1828 does the job.
meillo@104	1829
meillo@104	1830
meillo@21	1831
meillo@12	1832
meillo@58	1833 .H2 "Attachments
meillo@22	1834 .P
meillo@101	1835 The mind model of email attachments is unrelated to MIME.
meillo@101	1836 Although the MIME RFCs (2045 through 2049) define the technical
meillo@109	1837 requirements for having attachments, they do not mention the word
meillo@101	1838 ``attachment''.
meillo@101	1839 Instead of attachments, MIME talks about ``multi-part message bodies''
meillo@101	1840 [RFC\\|2045], a more general concept.
meillo@101	1841 Multi-part messages are messages
meillo@101	1842 ``in which one or more different
meillo@101	1843 sets of data are combined in a single body''
meillo@101	1844 [RFC\\|2046].
meillo@101	1845 MIME keeps its descriptions generic;
meillo@101	1846 it does not imply specific usage models.
meillo@109	1847 One usage model became prevalent: attachments.
meillo@101	1848 The idea is having a main text document with files of arbitrary kind
meillo@101	1849 attached to it.
meillo@101	1850 In MIME terms, this is a multi-part message having a text part first
meillo@110	1851 and parts of arbitrary type following.
meillo@101	1852 .P
meillo@101	1853 MH's MIME support is a direct implementation of the RFCs.
meillo@101	1854 The perception of the topic described in the RFCs is clearly visible
meillo@101	1855 in MH's implementation.
meillo@109	1856 In result, MH had all the MIME features but no idea of attachments.
meillo@109	1857 But users don't need all the MIME features,
meillo@109	1858 they want convenient attachment handling.
meillo@109	1859
meillo@102	1860
meillo@102	1861 .U3 "Composing MIME Messages
meillo@102	1862 .P
meillo@102	1863 In order to improve the situation on the message composing side,
meillo@102	1864 Jon Steinhart had added an attachment system to nmh in 2002.
meillo@101	1865 .Ci 7480dbc14bc90f2d872d434205c0784704213252
meillo@102	1866 In the file
meillo@102	1867 .Fn docs/README-ATTACHMENTS ,
meillo@102	1868 he described his motivation to do so as such:
meillo@101	1869 .QS
meillo@101	1870 Although nmh contains the necessary functionality for MIME message handing,
meillo@101	1871 the interface to this functionality is pretty obtuse.
meillo@101	1872 There's no way that I'm ever going to convince my partner to write
meillo@101	1873 .Pn mhbuild
meillo@101	1874 composition files!
meillo@101	1875 .QE
meillo@102	1876 .LP
meillo@102	1877 With this change, the mind model of attachments entered nmh.
meillo@102	1878 In the same document:
meillo@101	1879 .QS
meillo@101	1880 These changes simplify the task of managing attachments on draft files.
meillo@101	1881 They allow attachments to be added, listed, and deleted.
meillo@101	1882 MIME messages are automatically created when drafts with attachments
meillo@101	1883 are sent.
meillo@101	1884 .QE
meillo@102	1885 .LP
meillo@102	1886 Unfortunately, the attachment system,
meillo@102	1887 like any new facilities in nmh,
meillo@110	1888 was inactive by default.
meillo@101	1889 .P
meillo@101	1890 During my work in Argentina, I tried to improve the attachment system.
meillo@102	1891 But, because of great opposition in the nmh community,
meillo@102	1892 my patch died as a proposal on the mailing list, after long discussions.
meillo@101	1893 .[
meillo@101	1894 nmh-workers attachment proposal
meillo@101	1895 .]
meillo@110	1896 In January 2012, I extended the patch and applied it to mmh.
meillo@101	1897 .Ci 8ff284ff9167eff8f5349481529332d59ed913b1
meillo@102	1898 In mmh, the attachment system is active by default.
meillo@102	1899 Instead of command line switches, the
meillo@102	1900 .Pe Attachment-Header
meillo@102	1901 profile entry is used to specify
meillo@102	1902 the name of the attachment header field.
meillo@102	1903 It is pre-defined to
meillo@102	1904 .Hd Attach .
meillo@101	1905 .P
meillo@101	1906 To add an attachment to a draft, simply add an attachment header:
meillo@101	1907 .VS
meillo@101	1908 To: bob
meillo@101	1909 Subject: The file you wanted
meillo@101	1910 Attach: /path/to/the/file-bob-wanted
meillo@101	1911 --------
meillo@101	1912 Here it is.
meillo@101	1913 VE
meillo@101	1914 The header field can be added to the draft manually in the editor,
meillo@102	1915 or by using the `attach' command at the WhatNow prompt, or
meillo@102	1916 non-interactively with
meillo@101	1917 .Pn anno :
meillo@101	1918 .VS
meillo@102	1919 anno -append -nodate -component Attach -text /path/to/attachment
meillo@101	1920 VE
meillo@102	1921 Drafts with attachment headers are converted to MIME automatically by
meillo@102	1922 .Pn send .
meillo@102	1923 The conversion to MIME is invisible to the user.
meillo@102	1924 The draft stored in the draft folder is always in source form, with
meillo@101	1925 attachment headers.
meillo@101	1926 If the MIMEification fails, for instance because the file to attach
meillo@101	1927 is not accessible, the original draft is not changed.
meillo@101	1928 .P
meillo@102	1929 The attachment system handles the forwarding of messages, too.
meillo@101	1930 If the attachment header value starts with a plus character (`+'),
meillo@101	1931 like in
meillo@101	1932 .Cl "Attach: +bob 30 42" ,
meillo@101	1933 The given messages in the specified folder will be attached.
meillo@101	1934 This allowed to simplify
meillo@101	1935 .Pn forw .
meillo@101	1936 .Ci f41f04cf4ceca7355232cf7413e59afafccc9550
meillo@101	1937 .P
meillo@101	1938 Closely related to attachments is non-ASCII text content,
meillo@101	1939 because it requires MIME too.
meillo@102	1940 In nmh, the user needed to call `mime' at the WhatNow prompt
meillo@101	1941 to have the draft converted to MIME.
meillo@102	1942 This was necessary whenever the draft contained non-ASCII characters.
meillo@101	1943 If the user did not call `mime', a broken message would be sent.
meillo@101	1944 Therefore, the
meillo@101	1945 .Pe automimeproc
meillo@101	1946 profile entry could be specified to have the `mime' command invoked
meillo@102	1947 automatically each time.
meillo@101	1948 Unfortunately, this approach conflicted with with attachment system
meillo@101	1949 because the draft would already be in MIME format at the time
meillo@101	1950 when the attachment system wanted to MIMEify it.
meillo@102	1951 To use nmh's attachment system, `mime' must not be called at the
meillo@102	1952 WhatNow prompt and
meillo@101	1953 .Pe automimeproc
meillo@102	1954 must not be set in the profile.
meillo@101	1955 But then the case of non-ASCII text without attachment headers was
meillo@101	1956 not caught.
meillo@102	1957 All in all, the solution was complex and irritating.
meillo@102	1958 My patch from December 2010 would have simplified the situation.
meillo@102	1959 .P
meillo@101	1960 Mmh's current solution is even more elaborate.
meillo@101	1961 Any necessary MIMEification is done automatically.
meillo@101	1962 There is no `mime' command at the WhatNow prompt anymore.
meillo@102	1963 The draft will be converted automatically to MIME when either an
meillo@102	1964 attachment header or non-ASCII text is present.
meillo@101	1965 Further more, the special meaning of the hash character (`#')
meillo@102	1966 at line beginnings in the draft message is removed.
meillo@102	1967 Users need not at all deal with the whole topic.
meillo@101	1968 .P
meillo@102	1969 Although the new approach does not anymore support arbitrary MIME
meillo@102	1970 compositions directly, the full power of
meillo@101	1971 .Pn mhbuild
meillo@101	1972 can still be accessed.
meillo@102	1973 Given no attachment headers are included, the user can create
meillo@101	1974 .Pn mhbuild
meillo@102	1975 composition drafts like in nmh.
meillo@101	1976 Then, at the WhatNow prompt, he needs to invoke
meillo@101	1977 .Cl "edit mhbuild
meillo@101	1978 to convert it to MIME.
meillo@110	1979 Because the resulting draft does neither contain non-ASCII characters
meillo@102	1980 nor has it attachment headers, the attachment system will not touch it.
meillo@101	1981 .P
meillo@121	1982 The approach taken in mmh is tailored towards todays most common case:
meillo@101	1983 a text part with possibly attachments.
meillo@102	1984 This case is simplified a lot for users.
meillo@102	1985
meillo@112	1986
meillo@102	1987 .U3 "MIME Type Guessing
meillo@102	1988 .P
meillo@102	1989 The use of
meillo@101	1990 .Pn mhbuild
meillo@102	1991 composition drafts had one notable advantage over attachment headers
meillo@102	1992 from the programmer's point of view: The user provides the appropriate
meillo@102	1993 MIME types for files to include.
meillo@102	1994 The attachment system needs to find out the correct MIME type itself.
meillo@102	1995 This is a difficult task, yet it spares the user irritating work.
meillo@102	1996 Determining the correct MIME type of content is partly mechanical,
meillo@102	1997 partly intelligent work.
meillo@102	1998 Forcing the user to find out the correct MIME type,
meillo@102	1999 forces him to do partly mechanical work.
meillo@102	2000 Letting the computer do the work, can lead to bad choices for difficult
meillo@102	2001 content.
meillo@102	2002 For mmh, the latter option was chosen.
meillo@102	2003 .P
meillo@102	2004 Determining the MIME type by the suffix of the file name is a dumb
meillo@102	2005 approach, yet it is simple to implement and provides good results
meillo@102	2006 for the common cases.
meillo@102	2007 Mmh implements this approach in the
meillo@102	2008 .Pn print-mimetype
meillo@102	2009 script.
meillo@112	2010 .Ci 4b5944268ea0da7bb30598a27857304758ea9b44
meillo@102	2011 Using it is the default choice.
meillo@102	2012 .P
meillo@112	2013 A far better, though less portable, approach is the use of
meillo@102	2014 .Pn file .
meillo@102	2015 This standard tool tries to determine the type of files.
meillo@102	2016 Unfortunately, its capabilities and accuracy varies from system to system.
meillo@102	2017 Additionally, its output was only intended for human beings,
meillo@102	2018 but not to be used by programs.
meillo@102	2019 It varies much.
meillo@102	2020 Nevertheless, modern versions of GNU
meillo@102	2021 .Pn file ,
meillo@102	2022 which is prevalent on the popular GNU/Linux systems,
meillo@102	2023 provides MIME type output in machine-readable form.
meillo@102	2024 Although this solution is highly system-dependent,
meillo@102	2025 it solves the difficult problem well.
meillo@102	2026 On systems where GNU
meillo@102	2027 .Pn file ,
meillo@102	2028 version 5.04 or higher, is available it should be used.
meillo@102	2029 One needs to specify the following profile entry to do so:
meillo@112	2030 .Ci 3baec236a39c5c89a9bda8dbd988d643a21decc6
meillo@102	2031 .VS
meillo@102	2032 Mime-Type-Query: file -b --mime
meillo@102	2033 VE
meillo@102	2034 .LP
meillo@102	2035 Other versions of
meillo@102	2036 .Pn file
meillo@102	2037 might possibly be usable with wrapper scripts to reformat the output.
meillo@102	2038 The diversity among
meillo@102	2039 .Pn file
meillo@102	2040 implementations is great; one needs to check the local variant.
meillo@102	2041 .P
meillo@102	2042 If no MIME type can be determined, text content gets sent as
meillo@102	2043 `text/plain' and anything else under the generic fall-back type
meillo@102	2044 `application/octet-stream'.
meillo@102	2045 It is not possible in mmh to override the automatic MIME type guessing
meillo@102	2046 for a specific file.
meillo@102	2047 To do so, the user would need to know in advance for which file
meillo@102	2048 the automatic guessing does fail, or the system would require interaction.
meillo@102	2049 I consider both cases impractical.
meillo@102	2050 The existing solution should be sufficient.
meillo@102	2051 If not, the user may always fall back to
meillo@102	2052 .Pn mhbuild
meillo@102	2053 composition drafts and ignore the attachment system.
meillo@101	2054
meillo@102	2055
meillo@102	2056 .U3 "Storing Attachments
meillo@102	2057 .P
meillo@108	2058 Extracting MIME parts of a message and storing them to disk is done by
meillo@108	2059 .Pn mhstore .
meillo@108	2060 The program has two operation modes,
meillo@108	2061 .Sw -auto
meillo@108	2062 and
meillo@108	2063 .Sw -noauto .
meillo@108	2064 With the former one, each part is stored under the filename given in the
meillo@108	2065 MIME part's meta information, if available.
meillo@108	2066 This naming information is usually available for modern attachments.
meillo@108	2067 If no filename is available, this MIME part is stored as if
meillo@108	2068 .Sw -noauto
meillo@108	2069 would have been specified.
meillo@108	2070 In the
meillo@108	2071 .Sw -noauto
meillo@108	2072 mode, the parts are processed according to rules, defined by
meillo@108	2073 .Pe mhstore-store-*
meillo@108	2074 profile entries.
meillo@108	2075 These rules define generic filename templates for storing
meillo@108	2076 or commands to post-process the contents in arbitrary ways.
meillo@108	2077 If no matching rule is available the part is stored under a generic
meillo@108	2078 filename, built from message number, MIME part number, and MIME type.
meillo@108	2079 .P
meillo@108	2080 The
meillo@108	2081 .Sw -noauto
meillo@108	2082 mode had been the default in nmh because it was considered safe,
meillo@108	2083 in contrast to the
meillo@108	2084 .Sw -auto
meillo@108	2085 mode.
meillo@108	2086 In mmh,
meillo@108	2087 .Sw -auto
meillo@108	2088 is not dangerous anymore.
meillo@108	2089 Two changes were necessary:
meillo@108	2090 .BU
meillo@108	2091 Any directory path is removed from the proposed filename.
meillo@108	2092 Thus, the files are always stored in the expected directory.
meillo@108	2093 .Ci 41b6eadbcecf63c9a66aa5e582011987494abefb
meillo@108	2094 .BU
meillo@108	2095 Tar files are not extracted automatically any more.
meillo@108	2096 Thus, the rest of the file system will not be touched.
meillo@108	2097 .Ci 94c80042eae3383c812d9552089953f9846b1bb6
meillo@108	2098 .LP
meillo@108	2099 Now, the outcome of mmh's
meillo@108	2100 .Cl "mhstore -auto
meillo@110	2101 can be foreseen from the output of
meillo@108	2102 .Cl "mhlist -verbose" .
meillo@108	2103 .P
meillo@108	2104 The
meillo@108	2105 .Sw -noauto
meillo@108	2106 mode is seen to be more powerful but less convenient.
meillo@108	2107 On the other hand,
meillo@108	2108 .Sw -auto
meillo@108	2109 is safe now and
meillo@108	2110 storing attachments under their original name is intuitive.
meillo@108	2111 Hence,
meillo@108	2112 .Sw -auto
meillo@108	2113 serves better as the default option.
meillo@108	2114 .Ci 3410b680416c49a7617491af38bc1929855a331d
meillo@108	2115 .P
meillo@108	2116 Files are stored into the directory given by the
meillo@108	2117 .Pe Nmh-Storage
meillo@108	2118 profile entry, if set, or
meillo@108	2119 into the current working directory, otherwise.
meillo@108	2120 Storing to different directories is only possible with
meillo@108	2121 .Pe mhstore-store-*
meillo@108	2122 profile entries.
meillo@108	2123 .P
meillo@108	2124 Still, in both modes, existing files get overwritten silently.
meillo@108	2125 This can be considered a bug.
meillo@108	2126 Yet, each other behavior has its draw-backs, too.
meillo@108	2127 Refusing to replace files requires adding a
meillo@108	2128 .Sw -force
meillo@108	2129 option.
meillo@108	2130 Users will likely need to invoke
meillo@108	2131 .Pn mhstore
meillo@108	2132 a second time with
meillo@108	2133 .Sw -force
meillo@108	2134 then.
meillo@108	2135 Eventually, only the user can decide in the concrete case.
meillo@108	2136 This requires interaction, which I like to avoid if possible.
meillo@108	2137 Appending a unique suffix to the filename is another bad option.
meillo@108	2138 For now, the behavior remains as it is.
meillo@108	2139 .P
meillo@108	2140 In mmh, only MIME parts of type message are special in
meillo@108	2141 .Pn mhstore 's
meillo@108	2142 .Sw -auto
meillo@108	2143 mode.
meillo@108	2144 Instead of storing message/rfc822 parts as files to disk,
meillo@108	2145 they are stored as messages into the current mail folder.
meillo@108	2146 The same applies to message/partial, only, the parts are reassembled
meillo@108	2147 automatically before.
meillo@108	2148 Parts of type message/external-body are not automatically retrieved
meillo@108	2149 anymore. Instead, Information on how to retrieve them is output.
meillo@108	2150 Not supporting this rare case saved nearly one thousand lines of code.
meillo@108	2151 .Ci 55e1d8c654ee0f7c45b9361ce34617983b454c32
meillo@108	2152 .\" XXX mention somewhere else too: (The profile entry `nmh-access-ftp'
meillo@108	2153 .\" and sbr/ruserpass.c for reading ~/.netrc are gone now.)
meillo@108	2154 Not special anymore is `application/octet-stream; type=tar'.
meillo@108	2155 Automatically extracting such MIME parts had been the dangerous part
meillo@108	2156 of the
meillo@108	2157 .Sw -auto
meillo@108	2158 mode.
meillo@108	2159 .Ci 94c80042eae3383c812d9552089953f9846b1bb6
meillo@108	2160
meillo@102	2161
meillo@102	2162
meillo@102	2163 .U3 "Showing MIME Messages
meillo@102	2164 .P
meillo@114	2165 The program
meillo@114	2166 .Pn mhshow
meillo@114	2167 had been written to display MIME messages.
meillo@114	2168 It implemented the conceptional view of the MIME RFCs.
meillo@114	2169 Nmh's
meillo@114	2170 .Pn mhshow
meillo@114	2171 handled each MIME part independently, presenting them separately
meillo@114	2172 to the user.
meillo@114	2173 This does not match today's understanding of email attachments,
meillo@114	2174 where displaying a message is seen to be a single, integrated operation.
meillo@114	2175 Today, email messages are expected to consist of a main text part
meillo@114	2176 plus possibly attachments.
meillo@114	2177 They are not any more seen to be arbitrary MIME hierarchies with
meillo@114	2178 information on how to display the individual parts.
meillo@114	2179 I adjusted
meillo@114	2180 .Pn mhshow 's
meillo@114	2181 behavior to the modern view on the topic.
meillo@114	2182 .P
meillo@114	2183 Note that this section completely ignores the original
meillo@114	2184 .Pn show
meillo@114	2185 program, because it was not capable to display MIME messages
meillo@114	2186 and is no longer part of mmh.
meillo@114	2187 Although
meillo@114	2188 .Pn mhshow
meillo@114	2189 was renamed to
meillo@114	2190 .Pn show
meillo@114	2191 in mmh, this section uses the name
meillo@114	2192 .Pn mhshow ,
meillo@114	2193 in order to avoid confusion.
meillo@114	2194 .P
meillo@114	2195 In mmh, the basic idea is that
meillo@114	2196 .Pn mhshow
meillo@114	2197 should display a message in one single pager session.
meillo@114	2198 Therefore,
meillo@114	2199 .Pn mhshow
meillo@114	2200 invokes a pager session for all its output,
meillo@114	2201 whenever it prints to a terminal.
meillo@114	2202 .Ci a4197ea6ffc5c1550e8b52d5a654bcaaaee04a4e
meillo@114	2203 In consequence,
meillo@114	2204 .Pn mhl
meillo@114	2205 does no more invoke a pager.
meillo@114	2206 .Ci 0e46503be3c855bddaeae3843e1b659279c35d70
meillo@114	2207 With
meillo@114	2208 .Pn mhshow
meillo@114	2209 replacing the original
meillo@114	2210 .Pn show ,
meillo@114	2211 output from
meillo@114	2212 .Pn mhl
meillo@114	2213 does not go to the terminal directly, but through
meillo@114	2214 .Pn mhshow .
meillo@114	2215 Hence,
meillo@114	2216 .Pn mhl
meillo@114	2217 does not need to invoke a pager.
meillo@114	2218 The one and only job of
meillo@114	2219 .Pn mhl
meillo@114	2220 is to format messages or parts of them.
meillo@114	2221 The only place in mmh, where a pager is invoked is
meillo@114	2222 .Pn mhshow .
meillo@114	2223 .P
meillo@114	2224 .Pe mhshow-show-*
meillo@114	2225 profile entries can be used to display MIME parts in a specific way.
meillo@114	2226 For instance, PDF and Postscript files could be converted to plain text
meillo@114	2227 to display them in the terminal.
meillo@114	2228 In mmh, the displaying of MIME parts will always be done serially.
meillo@114	2229 The request to display the MIME type `multipart/parallel' in parallel
meillo@114	2230 is ignored.
meillo@114	2231 It is simply treated as `multipart/mixed'.
meillo@114	2232 .Ci d0581ba306a7299113a346f9b4c46ce97bc4cef6
meillo@114	2233 This could already be requested with the, now removed,
meillo@114	2234 .Sw -serialonly
meillo@114	2235 switch of
meillo@114	2236 .Pn mhshow .
meillo@114	2237 As MIME parts are always processed exclusively , i.e. serially,
meillo@114	2238 the `%e' escape in
meillo@114	2239 .Pe mhshow-show-*
meillo@114	2240 profile entries became useless and was thus removed.
meillo@114	2241 .Ci a20d405db09b7ccca74d3e8c57550883da49e1ae
meillo@114	2242 .P
meillo@114	2243 In the intended setup, only text content would be displayed.
meillo@114	2244 Non-text content would be converted to text by appropriate
meillo@114	2245 .Pe mhshow-show-*
meillo@114	2246 profile entries before, if possible and wanted.
meillo@114	2247 All output would be displayed in a single pager session.
meillo@114	2248 Other kinds of attachments are ignored.
meillo@114	2249 With
meillo@114	2250 .Pe mhshow-show-*
meillo@114	2251 profile entries for them, they can be displayed serially along
meillo@114	2252 the message.
meillo@114	2253 For parallel display, the attachments need to be stored to disk first.
meillo@114	2254 .P
meillo@114	2255 To display text content in foreign charsets, they need to be converted
meillo@114	2256 to the native charset.
meillo@114	2257 Therefore,
meillo@114	2258 .Pe mhshow-charset-*
meillo@114	2259 profile entries used to be needed.
meillo@121	2260 In mmh, the conversion is done automatically by piping the text through
meillo@114	2261 the
meillo@114	2262 .Pn iconv
meillo@114	2263 command, if necessary.
meillo@114	2264 .Ci 2433122c20baccb10b70b49c04c6b0497b5b3b60
meillo@114	2265 Custom
meillo@114	2266 .Pe mhshow-show-*
meillo@114	2267 rules for textual content might need a
meillo@114	2268 .Cl "iconv -f %c %f \|
meillo@114	2269 prefix to have the text converted to the native charset.
meillo@114	2270 .P
meillo@121	2271 Although the conversion of foreign charsets to the native one
meillo@114	2272 has improved, it is not consistent enough.
meillo@114	2273 Further work needs to be done and
meillo@114	2274 the basic concepts in this field need to be re-thought.
meillo@114	2275 Though, the default setup of mmh displays message in foreign charsets
meillo@114	2276 correctly without the need to configure anything.
meillo@114	2277
meillo@114	2278
meillo@114	2279 .ig
meillo@114	2280
meillo@114	2281 .P
meillo@114	2282 mhshow/mhstore: Removed support for retrieving message/external-body parts.
meillo@114	2283 These tools won't download the contents automatically anymore. Instead,
meillo@114	2284 they print the information needed to get the contents. If someone should
meillo@114	2285 really receive one of those rare message/external-body messages, he can
meillo@114	2286 do the job manually. We save nearly a thousand lines of code. That's worth
meillo@114	2287 it!
meillo@114	2288 (The profile entry `nmh-access-ftp' and sbr/ruserpass.c for reading
meillo@114	2289 ~/.netrc are gone now.)
meillo@114	2290 .Ci 55e1d8c654ee0f7c45b9361ce34617983b454c32
meillo@114	2291
meillo@114	2292 ..
meillo@102	2293
meillo@58	2294
meillo@58	2295
meillo@58	2296 .H2 "Digital Cryptography
meillo@22	2297 .P
meillo@58	2298 Signing and encryption.
meillo@112	2299 .P
meillo@112	2300 FIXME
meillo@58	2301
meillo@58	2302
meillo@102	2303
meillo@133	2304 .H2 "Draft and Trash Folder
meillo@131	2305 .P
meillo@58	2306
meillo@131	2307 .U3 "Draft Folder
meillo@131	2308 .P
meillo@131	2309 In the beginning, MH had the concept of a draft message.
meillo@131	2310 This is the file
meillo@131	2311 .Fn draft
meillo@131	2312 in the MH directory, which is treated special.
meillo@131	2313 On composing a message, this draft file was used.
meillo@131	2314 As the draft file was one particular file, only one draft could be
meillo@131	2315 managed at any time.
meillo@131	2316 When starting to compose another message before the former one was sent,
meillo@131	2317 the user had to decide among:
meillo@131	2318 .BU
meillo@131	2319 Use the old draft to finish and send it before starting with a new one.
meillo@131	2320 .BU
meillo@131	2321 Discard the old draft, replacing it with the new one.
meillo@131	2322 .BU
meillo@131	2323 Preserve the old draft by refiling it to a folder.
meillo@131	2324 .P
meillo@131	2325 This was, it was only possible to work in alternation on multiple drafts.
meillo@131	2326 Therefore, the current draft needed to be refiled to a folder and
meillo@131	2327 another one re-using for editing.
meillo@131	2328 Working on multiple drafts at the same time was impossible.
meillo@131	2329 The usual approach of switching to a different MH context did not
meillo@131	2330 change anything.
meillo@131	2331 .P
meillo@131	2332 The draft folder facility exists to
meillo@131	2333 allow true parallel editing of drafts, in a straight forward way.
meillo@131	2334 It was introduced by Marshall T. Rose, already in 1984.
meillo@131	2335 Similar to other new features, the draft folder was inactive by default.
meillo@131	2336 Even in nmh, the highly useful draft folder was not available
meillo@131	2337 out-of-the-box.
meillo@131	2338 At least, Richard Coleman added the man page
meillo@131	2339 .Mp mh-draft (5)
meillo@131	2340 to better document the feature.
meillo@131	2341 .P
meillo@131	2342 Not using the draft folder facility has the single advantage of having
meillo@131	2343 the draft file at a static location.
meillo@131	2344 This is simple in simple cases but the concept does not scale for more
meillo@131	2345 complex cases.
meillo@131	2346 The concept of the draft message is too limited for the problem.
meillo@131	2347 Therefore the draft folder was introduced.
meillo@131	2348 It is the more powerful and more natural concept.
meillo@131	2349 The draft folder is a folder like any other folder in MH.
meillo@131	2350 Its messages can be listed like any other messages.
meillo@131	2351 A draft message is no longer a special case.
meillo@131	2352 Tools do not need special switches to work on the draft message.
meillo@131	2353 Hence corner-cases were removed.
meillo@131	2354 .P
meillo@131	2355 The trivial part of the work was activating the draft folder with a
meillo@131	2356 default name.
meillo@131	2357 I chose the name
meillo@131	2358 .Fn +drafts
meillo@131	2359 for obvious reasons.
meillo@131	2360 In consequence, the command line switches
meillo@131	2361 .Sw -draftfolder
meillo@131	2362 and
meillo@131	2363 .Sw -draftmessage
meillo@131	2364 could be removed.
meillo@131	2365 More difficult but also more improving was updating the tools to the
meillo@131	2366 new concept.
meillo@131	2367 For nearly three decades, the tools needed to support two draft handling
meillo@131	2368 approaches.
meillo@131	2369 By fully switching to the draft folder, the tools could be simplified
meillo@131	2370 by dropping the awkward draft message handling code.
meillo@131	2371 .Sw -draft
meillo@131	2372 switches were removed because operating on a draft message is no longer
meillo@131	2373 special.
meillo@131	2374 It became indistinguishable to operating on any other message.
meillo@131	2375 There is no more need to query the user for draft handling.
meillo@131	2376 It is always possible to add another new draft.
meillo@131	2377 Refiling drafts is without difference to refiling other messages.
meillo@131	2378 All these special cases are gone.
meillo@131	2379 Yet, one draft-related switch remained.
meillo@131	2380 .Pn comp
meillo@131	2381 still has
meillo@131	2382 .Sw -[no]use
meillo@131	2383 for switching between two modes:
meillo@131	2384 .BU
meillo@131	2385 .Sw -use :
meillo@131	2386 Modify an existing draft.
meillo@131	2387 .BU
meillo@131	2388 .Sw -nouse :
meillo@131	2389 Compose a new draft, possibly taking some existing message as a form.
meillo@131	2390 .P
meillo@131	2391 In either case, the behavior of
meillo@131	2392 .Pn comp
meillo@131	2393 is deterministic.
meillo@131	2394 .P
meillo@131	2395 .Pn send
meillo@131	2396 now operates on the current message in the draft folder by default.
meillo@131	2397 As message and folder can both be overridden by specifying them on
meillo@131	2398 the command line, it is possible to send any message in the mail storage
meillo@131	2399 by simply specifying its number and folder.
meillo@131	2400 In contrast to the other tools,
meillo@131	2401 .Pn send
meillo@131	2402 takes the draft folder as its default folder.
meillo@131	2403 .P
meillo@131	2404 Dropping the draft message concept in favor for the draft folder concept,
meillo@131	2405 removed special cases with regular cases.
meillo@131	2406 This simplified the source code of the tools, as well as the concepts.
meillo@131	2407 In mmh, draft management does not break with the MH concepts
meillo@131	2408 but applies them.
meillo@133	2409 .Cl "scan +drafts" ,
meillo@133	2410 for instance, is a truly natural request.
meillo@131	2411 Most of the work was already done by Rose in the eighties.
meillo@133	2412 The original improvement of mmh is dropping the old draft message approach
meillo@133	2413 and thus simplifying the tools, the documentation and the system as a whole.
meillo@131	2414 Although my part in the draft handling improvement was small,
meillo@133	2415 it was an important one.
meillo@131	2416
meillo@131	2417
meillo@131	2418 .U3 "Trash Folder
meillo@131	2419 .P
meillo@131	2420 Similar to the situation for drafts is the situation for removed messages.
meillo@131	2421 Historically, a message was ``deleted'' by prepending a specific
meillo@131	2422 \fIbackup prefix\fP, usually the comma character,
meillo@131	2423 to the file name.
meillo@131	2424 The specific message would vanish from MH because only files with
meillo@131	2425 non-digit characters in their name are not treated as messages.
meillo@131	2426 Although files remained in the file system,
meillo@131	2427 the messages were no more visible in MH.
meillo@131	2428 To truly delete them, a maintenance job is needed.
meillo@131	2429 Usually a cron job is installed to delete them after a grace time.
meillo@131	2430 For instance:
meillo@131	2431 .VS
meillo@131	2432 find $HOME/Mail -type f -name ',*' -ctime +7 -delete
meillo@131	2433 VE
meillo@131	2434 In such a setup, the original message can be restored
meillo@131	2435 within the grace time interval by stripping the
meillo@131	2436 the backup prefix from the file name.
meillo@131	2437 But one can not rely on this statement.
meillo@131	2438 If the last message of a folder with six messages (1-6) is removed,
meillo@131	2439 message
meillo@131	2440 .Fn 6 ,
meillo@131	2441 becomes file
meillo@131	2442 .Fn ,6 .
meillo@131	2443 If then a new message enters the same folder, it will be given
meillo@131	2444 the number one higher than the highest existing message.
meillo@131	2445 In this case the message is named
meillo@131	2446 .Fn 6
meillo@131	2447 then.
meillo@131	2448 If this message is removed as well,
meillo@131	2449 then the backup of the former message gets overwritten.
meillo@131	2450 Hence, the ability to restore removed messages does not only depend on
meillo@131	2451 the ``sweeping cron job'' but also on the removing of further messages.
meillo@131	2452 It is undesirable to have such obscure and complex mechanisms.
meillo@131	2453 The user should be given a small set of clear assertions.
meillo@131	2454 ``Removed files are restorable within a seven-day grace time.''
meillo@131	2455 is such a clear assertion.
meillo@131	2456 With the addition ``... unless a message with the same name in the
meillo@131	2457 same folder is removed before.'' the statement becomes complex.
meillo@131	2458 A user will hardly be able to keep track of any removal to know
meillo@131	2459 if the assertion still holds true for a specific file.
meillo@131	2460 The the real mechanism is practically obscure to the user.
meillo@131	2461 The consequences of further removals are not obvious.
meillo@131	2462 .P
meillo@131	2463 Further more, the backup files are scattered within the whole mail storage.
meillo@131	2464 This complicates managing them.
meillo@131	2465 It is possible, with help of
meillo@131	2466 .Pn find ,
meillo@131	2467 but everything would be more convenient
meillo@131	2468 if the deleted messages would be collected in one place.
meillo@131	2469 .P
meillo@131	2470 The profile entry
meillo@131	2471 .Pe rmmproc
meillo@131	2472 (previously named
meillo@131	2473 .Pe Delete-Prog )
meillo@131	2474 was introduced very early to improve the situation.
meillo@131	2475 It could be set to any command, which would be executed to removed
meillo@131	2476 the specified messages.
meillo@131	2477 This would override the default action, described above.
meillo@131	2478 Refiling the to-be-removed files to a garbage folder is the usual example.
meillo@131	2479 Nmh's man page
meillo@131	2480 .Mp rmm (1)
meillo@131	2481 proposes to set the
meillo@131	2482 .Pe rmmproc
meillo@131	2483 to
meillo@131	2484 .Cl "refile +d
meillo@131	2485 to move messages to the garbage folder,
meillo@131	2486 .Fn +d ,
meillo@131	2487 instead of renaming them with the backup prefix.
meillo@131	2488 The man page proposes additionally the expunge command
meillo@131	2489 .Cl "rm `mhpath +d all`
meillo@131	2490 to empty the garbage folder.
meillo@131	2491 .P
meillo@131	2492 Removing messages in such a way has advantages.
meillo@131	2493 The mail storage is prevented from being cluttered with removed messages
meillo@131	2494 because they are all collected in one place.
meillo@131	2495 Existing and removed messages are thus separated more strictly.
meillo@131	2496 No backup files are silently overwritten.
meillo@131	2497 Most important is the ability to keep removed messages in the MH domain.
meillo@131	2498 Messages in the trash folder can be listed like those in any other folder.
meillo@131	2499 Deleted messages can be displayed like any other messages.
meillo@131	2500 Restoring a deleted messages can be done with
meillo@131	2501 .Pn refile .
meillo@131	2502 All operations on deleted files are still covered by the MH tools.
meillo@131	2503 The trash folder is just like any other folder in the mail storage.
meillo@131	2504 .P
meillo@131	2505 Similar to the draft folder case, I dropped the old backup prefix approach
meillo@131	2506 in favor for replacing it by the better suiting trash folder system.
meillo@131	2507 Hence,
meillo@131	2508 .Pn rmm
meillo@131	2509 calls
meillo@131	2510 .Pn refile
meillo@131	2511 to move the to-be-removed message to the trash folder,
meillo@131	2512 .Fn +trash
meillo@131	2513 by default.
meillo@131	2514 To sweep it clean, one can use
meillo@131	2515 .Cl "rmm -unlink +trash a" ,
meillo@131	2516 where the
meillo@131	2517 .Sw -unlink
meillo@131	2518 switch causes the files to be unlinked.
meillo@131	2519 .P
meillo@131	2520 Dropping the legacy approach and completely converting to the new approach
meillo@131	2521 simplified the code base.
meillo@131	2522 The relationship between
meillo@131	2523 .Pn rmm
meillo@131	2524 and
meillo@131	2525 .Pn refile
meillo@131	2526 was inverted.
meillo@131	2527 In mmh,
meillo@131	2528 .Pn rmm
meillo@131	2529 invokes
meillo@131	2530 .Pn refile ,
meillo@131	2531 which used to be the other way round.
meillo@131	2532 Yet, the relationship is simpler now.
meillo@131	2533 No more can loops, like described in nmh's man page for
meillo@131	2534 .Mp refile (1),
meillo@131	2535 occur:
meillo@131	2536 .QS
meillo@131	2537 Since
meillo@131	2538 .Pn refile
meillo@131	2539 uses your
meillo@131	2540 .Pe rmmproc
meillo@131	2541 to delete the message, the
meillo@131	2542 .Pe rmmproc
meillo@131	2543 must NOT call
meillo@131	2544 .Pn refile
meillo@131	2545 without specifying
meillo@131	2546 .Sw -normmproc
meillo@131	2547 or you will create an infinite loop.
meillo@131	2548 .QE
meillo@131	2549 .LP
meillo@131	2550 .Pn rmm
meillo@131	2551 either unlinks a message with
meillo@131	2552 .Fu unlink()
meillo@131	2553 or invokes
meillo@131	2554 .Pn refile
meillo@131	2555 to move it to the trash folder.
meillo@131	2556 .Pn refile
meillo@131	2557 does not invoke any tools.
meillo@131	2558 .P
meillo@136	2559 By generalizing the message removal in the way that it became covered
meillo@136	2560 by the MH concepts made the whole system more powerful.
meillo@131	2561
meillo@131	2562
meillo@131	2563
meillo@131	2564
meillo@131	2565
meillo@133	2566 .H2 "Modern Defaults
meillo@133	2567 .P
meillo@133	2568 Nmh has a bunch of convenience-improving features inactive by default,
meillo@133	2569 although one can expect every new user wanting to have them active.
meillo@133	2570 The reason they are inactive by default is the wish to stay compatible
meillo@133	2571 with old versions.
meillo@136	2572 But what is the definition for old versions?
meillo@136	2573 Still, the highly useful draft folder facility has not been activated
meillo@136	2574 by default although it was introduced over twenty-five years ago.
meillo@133	2575 .[
meillo@133	2576 rose romine real work
meillo@133	2577 .]
meillo@136	2578 The community seems not to care.
meillo@136	2579 This is one of several examples that require new users to first build up
meillo@136	2580 a profile before they can access the modern features of nmh.
meillo@136	2581 Without an extensive profile, the setup is hardly usable
meillo@133	2582 for modern emailing.
meillo@133	2583 The point is not the customization of the setup,
meillo@136	2584 but the need to activate generally useful facilities.
meillo@133	2585 .P
meillo@133	2586 Yet, the real problem lies less in enabling the features, as this is
meillo@133	2587 straight forward as soon as one knows what he wants.
meillo@133	2588 The real problem is that new users need deep insights into the project
meillo@133	2589 before they find out what they are missing and that nmh actually
meillo@133	2590 provides it already, it just was not activated.
meillo@133	2591 To give an example, I needed one year of using nmh
meillo@133	2592 before I became aware of the existence of the attachment system.
meillo@133	2593 One could argue that this fact disqualifies my reading of the
meillo@133	2594 documentation.
meillo@133	2595 If I would have installed nmh from source back then, I could agree.
meillo@133	2596 Yet, I had used a prepackaged version and had expected that it would
meillo@133	2597 just work.
meillo@133	2598 Nevertheless, I had been convinced by the concepts of MH already
meillo@133	2599 and I am a software developer,
meillo@133	2600 still I required a lot of time to discover the cool features.
meillo@133	2601 How can we expect users to be even more advanced than me,
meillo@133	2602 just to allow them use MH in a convenient and modern way?
meillo@133	2603 Unless they are strongly convinced of the concepts, they will fail.
meillo@133	2604 I have seen friends of me giving up disappointed
meillo@133	2605 before they truly used the system,
meillo@133	2606 although they had been motivated in the beginning.
meillo@133	2607 They suffer hard enough to get used to the toolchest approach,
meillo@133	2608 we should spare them further inconveniences.
meillo@133	2609 .P
meillo@136	2610 Maintaining compatibility for its own sake is bad,
meillo@136	2611 because the code base collects more and more compatibility code.
meillo@136	2612 Sticking to the compatiblity code means remaining limited;
meillo@136	2613 not using it renders it unnecessary.
meillo@136	2614 Keeping unused alternative in the code is a bad choice as they likely
meillo@136	2615 gather bugs, by not being well tested.
meillo@136	2616 Also, the increased code size and the greater number of conditions
meillo@136	2617 increase the maintenance costs.
meillo@133	2618 If any MH implementation would be the back-end of widespread
meillo@133	2619 email clients with large user bases, compatibility would be more
meillo@133	2620 important.
meillo@133	2621 Yet, it appears as if this is not the case.
meillo@133	2622 Hence, compatibility is hardly important for technical reasons.
meillo@133	2623 Its importance originates rather from personal reasons.
meillo@133	2624 Nmh's user base is small and old.
meillo@133	2625 Changing the interfaces would cause inconvenience to long-term users of MH.
meillo@133	2626 It would force them to change their many years old MH configurations.
meillo@133	2627 I do understand this aspect, but it keeps new users from using MH.
meillo@133	2628 By sticking to the old users, new users are kept away.
meillo@133	2629 Yet, the future lies in new users.
meillo@133	2630 Hence, mmh invites new users by providing a convenient and modern setup,
meillo@133	2631 readily usable out-of-the-box.
meillo@133	2632 .P
meillo@136	2633 In mmh, all modern features are active by default and many previous
meillo@136	2634 approaches are removed or only accessible in manual ways.
meillo@136	2635 New default features include:
meillo@133	2636 .BU
meillo@133	2637 The attachment system (\c
meillo@133	2638 .Hd Attach ).
meillo@133	2639 .Ci 8ff284ff9167eff8f5349481529332d59ed913b1
meillo@133	2640 .BU
meillo@133	2641 The draft folder facility (\c
meillo@133	2642 .Fn +drafts ).
meillo@133	2643 .Ci 337338b404931f06f0db2119c9e145e8ca5a9860
meillo@133	2644 .BU
meillo@133	2645 The unseen sequence (`u')
meillo@133	2646 .Ci c2360569e1d8d3678e294eb7c1354cb8bf7501c1
meillo@133	2647 and the sequence negation prefix (`!').
meillo@133	2648 .Ci db74c2bd004b2dc9bf8086a6d8bf773ac051f3cc
meillo@133	2649 .BU
meillo@133	2650 Quoting the original message in the reply.
meillo@133	2651 .Ci 67411b1f95d6ec987b4c732459e1ba8a8ac192c6
meillo@133	2652 .BU
meillo@133	2653 Forwarding messages using MIME.
meillo@133	2654 .Ci 6e271608b7b9c23771523f88d23a4d3593010cf1
meillo@136	2655 .P
meillo@136	2656 In consequence, a setup with a profile that defines only the path to the
meillo@136	2657 mail storage, is already convenient to use.
meillo@136	2658 Again, Paul Vixie's ``edginess'' appeal supports the direction I took:
meillo@136	2659 ``the `main branch' should just be modern''.
meillo@136	2660 .[
meillo@136	2661 paul vixie edginess nmh-workers
meillo@136	2662 .]
meillo@131	2663
meillo@133	2664
meillo@133	2665
meillo@133	2666
meillo@133	2667
meillo@133	2668 .\" --------------------------------------------------------------
meillo@131	2669 .H1 "Styling
meillo@22	2670 .P
meillo@118	2671 Kernighan and Pike have emphasized the importance of style in the
meillo@118	2672 preface of their book:
meillo@118	2673 .[ [
meillo@118	2674 kernighan pike practice of programming
meillo@118	2675 .], p. x]
meillo@118	2676 .QS
meillo@118	2677 Chapter 1 discusses programming style.
meillo@118	2678 Good style is so important to good programming that we have chose
meillo@118	2679 to cover it first.
meillo@118	2680 .QE
meillo@118	2681 This section covers changes in mmh that were motivated by the desire
meillo@118	2682 to improve on style.
meillo@118	2683 Many of them follow the rules given in the quoted book.
meillo@118	2684 .[
meillo@118	2685 kernighan pike practice of programming
meillo@118	2686 .]
meillo@118	2687
meillo@118	2688
meillo@127	2689
meillo@127	2690
meillo@127	2691 .H2 "Code Style
meillo@118	2692 .P
meillo@118	2693 .U3 "Indentation Style
meillo@118	2694 .P
meillo@118	2695 Indentation styles are the holy cow of programmers.
meillo@118	2696 Again Kernighan and Pike:
meillo@118	2697 .[ [
meillo@118	2698 kernighan pike practice of programming
meillo@118	2699 .], p. 10]
meillo@118	2700 .QS
meillo@118	2701 Programmers have always argued about the layout of programs,
meillo@118	2702 but the specific style is much less important than its consistent
meillo@118	2703 application.
meillo@121	2704 Pick one style, preferably ours, use it consistently, and don't waste
meillo@118	2705 time arguing.
meillo@118	2706 .QE
meillo@118	2707 .P
meillo@118	2708 I agree that the constant application is most important,
meillo@118	2709 but I believe that some styles have advantages over others.
meillo@118	2710 For instance the indentation with tab characters only.
meillo@118	2711 Tab characters directly map to the nesting level \(en
meillo@118	2712 one tab, one level.
meillo@118	2713 Tab characters are flexible because developers can adjust them to
meillo@118	2714 whatever width they like to have.
meillo@118	2715 There is no more need to run
meillo@118	2716 .Pn unexpand
meillo@118	2717 or
meillo@118	2718 .Pn entab
meillo@118	2719 programs to ensure the correct mixture of leading tabs and spaces.
meillo@118	2720 The simple rules are: (1) Leading whitespace must consist of tabs only.
meillo@118	2721 (2) Any other whitespace should consist of spaces.
meillo@121	2722 These two rules ensure the integrity of the visual appearance.
meillo@121	2723 Although reformatting existing code should be avoided, I did it.
meillo@136	2724 I did not waste time arguing; I just reformated the code.
meillo@118	2725 .Ci a485ed478abbd599d8c9aab48934e7a26733ecb1
meillo@118	2726
meillo@118	2727 .U3 "Comments
meillo@118	2728 .P
meillo@118	2729 Section 1.6 of
meillo@118	2730 .[ [
meillo@118	2731 kernighan pike practice of programming
meillo@118	2732 .], p. 23]
meillo@118	2733 demands: ``Don't belabor the obvious.''
meillo@122	2734 Hence, I simply removed all the comments in the following code excerpt:
meillo@118	2735 .VS
meillo@120	2736 context_replace(curfolder, folder); /* update current folder */
meillo@120	2737 seq_setcur(mp, mp->lowsel); /* update current message */
meillo@120	2738 seq_save(mp); /* synchronize message sequences */
meillo@120	2739 folder_free(mp); /* free folder/message structure */
meillo@120	2740 context_save(); /* save the context file */
meillo@120	2741
meillo@120	2742 [...]
meillo@120	2743
meillo@120	2744 int c; /* current character */
meillo@120	2745 char cp; / miscellaneous character pointer */
meillo@120	2746
meillo@120	2747 [...]
meillo@120	2748
meillo@120	2749 /* NUL-terminate the field */
meillo@120	2750 *cp = '\0';
meillo@118	2751 VE
meillo@120	2752 .Ci 426543622b377fc5d091455cba685e114b6df674
meillo@118	2753 .P
meillo@136	2754 The program code explains enough itself, already.
meillo@136	2755
meillo@118	2756
meillo@118	2757 .U3 "Names
meillo@118	2758 .P
meillo@118	2759 Kernighan and Pike suggest:
meillo@118	2760 ``Use active names for functions''.
meillo@118	2761 .[ [
meillo@118	2762 kernighan pike practice of programming
meillo@118	2763 .], p. 4]
meillo@118	2764 One application of this rule was the rename of
meillo@118	2765 .Fu check_charset()
meillo@118	2766 to
meillo@118	2767 .Fu is_native_charset() .
meillo@118	2768 .Ci 8d77b48284c58c135a6b2787e721597346ab056d
meillo@118	2769 The same change fixed a violation of ``Be accurate'' as well.
meillo@118	2770 The code did not match the expectation the function suggested,
meillo@118	2771 as it, for whatever reason, only compared the first ten characters
meillo@118	2772 of the charset name.
meillo@118	2773 .P
meillo@118	2774 More important than using active names is using descriptive names.
meillo@118	2775 Renaming the obscure function
meillo@118	2776 .Fu m_unknown()
meillo@118	2777 was a delightful event.
meillo@118	2778 .Ci 611d68d19204d7cbf5bd585391249cb5bafca846
meillo@118	2779 .P
meillo@118	2780 Magic numbers are generally considered bad style.
meillo@118	2781 Obviously, Kernighan and Pike agree:
meillo@118	2782 ``Give names to magic numbers''.
meillo@118	2783 .[ [
meillo@118	2784 kernighan pike practice of programming
meillo@118	2785 .], p. 19]
meillo@118	2786 One such change was naming the type of input \(en mbox or mail folder \(en
meillo@118	2787 to be scanned:
meillo@118	2788 .VS
meillo@118	2789 #define SCN_MBOX (-1)
meillo@118	2790 #define SCN_FOLD 0
meillo@118	2791 VE
meillo@118	2792 .Ci 7ffb36d28e517a6f3a10272056fc127592ab1c19
meillo@118	2793 .P
meillo@118	2794 The argument
meillo@118	2795 .Ar outnum
meillo@118	2796 of the function
meillo@118	2797 .Fu scan()
meillo@118	2798 in
meillo@118	2799 .Fn uip/scansbr.c
meillo@118	2800 defines the number of the message to be created.
meillo@118	2801 If no message is to be created, the argument is misused to transport
meillo@118	2802 program logic.
meillo@118	2803 This lead to obscure code.
meillo@118	2804 I improved the clarity of the code by introducing two variables:
meillo@118	2805 .VS
meillo@118	2806 int incing = (outnum > 0);
meillo@118	2807 int ismbox = (outnum != 0);
meillo@118	2808 VE
meillo@118	2809 They cover the magic values and are used for conditions.
meillo@118	2810 The variable
meillo@118	2811 .Ar outnum
meillo@118	2812 is only used when it holds an ordinary message number.
meillo@118	2813 .Ci b8b075c77be7794f3ae9ff0e8cedb12b48fd139f
meillo@118	2814 The clarity improvement of the change showed detours in the program logic
meillo@118	2815 of related code parts.
meillo@118	2816 Having the new variables with descriptive names, a more
meillo@121	2817 straight forward implementation became apparent.
meillo@118	2818 Before the clarification was done,
meillo@118	2819 the possibility to improve had not be seen.
meillo@118	2820 .Ci aa60b0ab5e804f8befa890c0a6df0e3143ce0723
meillo@118	2821
meillo@133	2822
meillo@133	2823
meillo@133	2824 .H2 "Structural Rework
meillo@133	2825 .P
meillo@136	2826 Although the stylistic changes described up to here improve the
meillo@136	2827 readability of the source code, all of them are changes ``in the small''.
meillo@136	2828 Structural changes affect a much larger area.
meillo@136	2829 They are more difficult to do but lead to larger improvements,
meillo@136	2830 especially as they influence the outer shape of the tools as well.
meillo@118	2831 .P
meillo@118	2832 At the end of their chapter on style,
meillo@118	2833 Kernighan and Pike ask: ``But why worry about style?''
meillo@136	2834 Following are two examples of structural rework that show
meillo@136	2835 why style is important in the first place.
meillo@136	2836
meillo@136	2837
meillo@136	2838 .U3 "Rework of \f(CWanno\fP
meillo@118	2839 .P
meillo@120	2840 Until 2002,
meillo@120	2841 .Pn anno
meillo@120	2842 had six functional command line switches,
meillo@120	2843 .Sw -component
meillo@120	2844 and
meillo@120	2845 .Sw -text ,
meillo@120	2846 which took an argument each,
meillo@120	2847 and the two pairs of flags,
meillo@120	2848 .Sw -[no]date
meillo@120	2849 and
meillo@120	2850 .Sw -[no]inplace.,
meillo@120	2851 .Sw -component
meillo@120	2852 and
meillo@120	2853 .Sw -text ,
meillo@120	2854 which took an argument each,
meillo@120	2855 and the two pairs of flags,
meillo@120	2856 .Sw -[no]date
meillo@120	2857 and
meillo@120	2858 .Sw -[no]inplace .
meillo@120	2859 Then Jon Steinhart introduced his attachment system.
meillo@120	2860 In need for more advanced annotation handling, he extended
meillo@120	2861 .Pn anno .
meillo@120	2862 He added five more switches:
meillo@120	2863 .Sw -draft ,
meillo@120	2864 .Sw -list ,
meillo@120	2865 .Sw -delete ,
meillo@120	2866 .Sw -append ,
meillo@120	2867 and
meillo@120	2868 .Sw -number ,
meillo@120	2869 the last one taking an argument.
meillo@121	2870 .Ci 7480dbc14bc90f2d872d434205c0784704213252
meillo@120	2871 Later,
meillo@120	2872 .Sw -[no]preserve
meillo@120	2873 was added.
meillo@121	2874 .Ci d9b1d57351d104d7ec1a5621f090657dcce8cb7f
meillo@120	2875 Then, the Synopsis section of the man page
meillo@120	2876 .Mp anno (1)
meillo@120	2877 read:
meillo@120	2878 .VS
meillo@120	2879 anno [+folder] [msgs] [-component field] [-inplace \| -noinplace]
meillo@120	2880 [-date \| -nodate] [-draft] [-append] [-list] [-delete]
meillo@120	2881 [-number [num\|all]] [-preserve \| -nopreserve] [-version]
meillo@120	2882 [-help] [-text body]
meillo@120	2883 VE
meillo@120	2884 .LP
meillo@120	2885 The implementation followed the same structure.
meillo@120	2886 Problems became visible when
meillo@120	2887 .Cl "anno -list -number 42
meillo@120	2888 worked on the current message instead on message number 42,
meillo@120	2889 and
meillo@120	2890 .Cl "anno -list -number l:5
meillo@124	2891 did not work on the last five messages but failed with the mysterious
meillo@120	2892 error message: ``anno: missing argument to -list''.
meillo@121	2893 Yet, the invocation matched the specification in the man page.
meillo@120	2894 There, the correct use of
meillo@120	2895 .Sw -number
meillo@120	2896 was defined as being
meillo@120	2897 .Cl "[-number [num\|all]]
meillo@120	2898 and the textual description for the combination with
meillo@120	2899 .Sw -list
meillo@120	2900 read:
meillo@120	2901 .QS
meillo@120	2902 The -list option produces a listing of the field bodies for
meillo@120	2903 header fields with names matching the specified component,
meillo@120	2904 one per line. The listing is numbered, starting at 1, if
meillo@120	2905 the -number option is also used.
meillo@120	2906 .QE
meillo@120	2907 .LP
meillo@120	2908 The problem was manifold.
meillo@120	2909 The code required a numeric argument to the
meillo@120	2910 .Sw -number
meillo@120	2911 switch.
meillo@120	2912 If it was missing or non-numeric,
meillo@120	2913 .Pn anno
meillo@120	2914 aborted with an error message that had an off-by-one error,
meillo@120	2915 printing the switch one before the failing one.
meillo@120	2916 Semantically, the argument to the
meillo@120	2917 .Sw -number
meillo@120	2918 switch is only necessary in combination with
meillo@120	2919 .Sw -delete ,
meillo@120	2920 but not with
meillo@120	2921 .Sw -list .
meillo@120	2922 In the former case it is even necessary.
meillo@120	2923 .P
meillo@120	2924 Trying to fix these problems on the surface would not have solved it truly.
meillo@120	2925 The problems discovered originate from a discrepance between the semantic
meillo@120	2926 structure of the problem and the structure implemented in the program.
meillo@120	2927 Such structural differences can not be cured on the surface.
meillo@120	2928 They need to be solved by adjusting the structure of the implementation
meillo@120	2929 to the structure of the problem.
meillo@120	2930 .P
meillo@120	2931 In 2002, the new switches
meillo@120	2932 .Sw -list
meillo@120	2933 and
meillo@120	2934 .Sw -delete
meillo@120	2935 were added in the same way, the
meillo@120	2936 .Sw -number
meillo@120	2937 switch for instance had been added.
meillo@120	2938 Yet, they are of structural different type.
meillo@120	2939 Semantically,
meillo@120	2940 .Sw -list
meillo@120	2941 and
meillo@120	2942 .Sw -delete
meillo@120	2943 introduce modes of operation.
meillo@120	2944 Historically,
meillo@120	2945 .Pn anno
meillo@120	2946 had only one operation mode: adding header fields.
meillo@120	2947 With the extension, it got two moder modes:
meillo@120	2948 listing and deleting header fields.
meillo@120	2949 The structure of the code changes did not pay respect to this
meillo@120	2950 fundamental change to
meillo@120	2951 .Pn anno 's
meillo@120	2952 behavior.
meillo@120	2953 Neither the implementation nor the documentation did clearly
meillo@120	2954 define them as being exclusive modes of operation.
meillo@120	2955 Having identified the problem, I solved it by putting structure into
meillo@120	2956 .Pn anno
meillo@120	2957 and its documentation.
meillo@120	2958 .Ci d54c8db8bdf01e8381890f7729bc0ef4a055ea11
meillo@120	2959 .P
meillo@120	2960 The difference is visible in both, the code and the documentation.
meillo@121	2961 The following code excerpt:
meillo@120	2962 .VS
meillo@120	2963 int delete = -2; /* delete header element if set */
meillo@120	2964 int list = 0; /* list header elements if set */
meillo@120	2965 [...]
meillo@121	2966 case DELETESW: /* delete annotations */
meillo@121	2967 delete = 0;
meillo@121	2968 continue;
meillo@121	2969 case LISTSW: /* produce a listing */
meillo@121	2970 list = 1;
meillo@121	2971 continue;
meillo@120	2972 VE
meillo@121	2973 .LP
meillo@121	2974 was replaced by:
meillo@120	2975 .VS
meillo@120	2976 static enum { MODE_ADD, MODE_DEL, MODE_LIST } mode = MODE_ADD;
meillo@120	2977 [...]
meillo@121	2978 case DELETESW: /* delete annotations */
meillo@121	2979 mode = MODE_DEL;
meillo@121	2980 continue;
meillo@121	2981 case LISTSW: /* produce a listing */
meillo@121	2982 mode = MODE_LIST;
meillo@121	2983 continue;
meillo@120	2984 VE
meillo@120	2985 .LP
meillo@121	2986 The replacement code does not only reflect the problem's structure better,
meillo@121	2987 it is easier to understand as well.
meillo@121	2988 The same applies to the documentation.
meillo@120	2989 The man page was completely reorganized to propagate the same structure.
meillo@121	2990 This is visible in the Synopsis section:
meillo@120	2991 .VS
meillo@120	2992 anno [+folder] [msgs] [-component field] [-text body]
meillo@120	2993 [-append] [-date \| -nodate] [-preserve \| -nopreserve]
meillo@120	2994 [-Version] [-help]
meillo@120	2995
meillo@120	2996 anno -delete [+folder] [msgs] [-component field] [-text
meillo@120	2997 body] [-number num \| all ] [-preserve \| -nopreserve]
meillo@120	2998 [-Version] [-help]
meillo@120	2999
meillo@120	3000 anno -list [+folder] [msgs] [-component field] [-number]
meillo@120	3001 [-Version] [-help]
meillo@120	3002 VE
meillo@121	3003 .\" XXX think about explaining the -preserve rework?
meillo@118	3004
meillo@58	3005
meillo@58	3006
meillo@133	3007 .U3 "Path Conversion
meillo@133	3008 .P
meillo@134	3009 Four kinds of path names can appear in MH:
meillo@134	3010 .IP (1)
meillo@134	3011 Absolute Unix directory paths, like
meillo@134	3012 .Fn /etc/passwd .
meillo@134	3013 .IP (2)
meillo@134	3014 Relative Unix directory paths, like
meillo@134	3015 .Fn ./foo/bar .
meillo@134	3016 .IP (3)
meillo@134	3017 Absolute MH folder paths, like
meillo@134	3018 .Fn +friends/phil .
meillo@134	3019 .IP (4)
meillo@134	3020 Relative MH folder paths, like
meillo@134	3021 .Fn @subfolder .
meillo@134	3022 .P
meillo@134	3023 The last type, relative MH folder paths, are hardly documented.
meillo@134	3024 Nonetheless, they are useful for large mail storages.
meillo@134	3025 The current mail folder is specified as `\c
meillo@134	3026 .Fn @ ',
meillo@134	3027 just like the current directory is specified as `\c
meillo@134	3028 .Fn . '.
meillo@134	3029 .P
meillo@134	3030 To allow MH tools to understand all four notations,
meillo@134	3031 they need to convert between them.
meillo@134	3032 In nmh, these path name conversion functions were located in the files
meillo@134	3033 .Fn sbr/path.c
meillo@134	3034 (``return a pathname'') and
meillo@134	3035 .Fn sbr/m_maildir.c
meillo@134	3036 (``get the path for the mail directory'').
meillo@134	3037 The seven functions in the two files were documented with no more
meillo@134	3038 than two comments, which described obvious information.
meillo@134	3039 The function signatures were neither explaining:
meillo@134	3040 .VS
meillo@134	3041 char path(char , int);
meillo@134	3042 char pluspath(char );
meillo@134	3043 char m_mailpath(char );
meillo@134	3044 char m_maildir(char );
meillo@134	3045 VE
meillo@134	3046 .P
meillo@134	3047 My investigation provides the following description:
meillo@134	3048 .BU
meillo@134	3049 The second parameter of
meillo@134	3050 .Fu path()
meillo@134	3051 defines the type of path given as first parameter.
meillo@134	3052 Directory paths are converted to absolute directory paths.
meillo@134	3053 Folder paths are converted to absolute folder paths.
meillo@134	3054 Folder paths must not include a leading `@' character.
meillo@134	3055 Leading plus characters are preserved.
meillo@134	3056 The result is a pointer to newly allocated memory.
meillo@134	3057 .BU
meillo@134	3058 .Fu pluspath()
meillo@134	3059 is a convenience-wrapper to
meillo@134	3060 .Fu path() ,
meillo@134	3061 to convert folder paths only.
meillo@134	3062 This function can not be used for directory paths.
meillo@134	3063 An empty string parameter causes a buffer overflow.
meillo@134	3064 .BU
meillo@134	3065 .Fu m_mailpath()
meillo@134	3066 converts directory paths to absolute directory paths.
meillo@134	3067 The characters `+' or `@' at the beginning of the path name are
meillo@134	3068 treated literal, i.e. as the first character of a relative directory path.
meillo@134	3069 Hence, this function can not be used for folder paths.
meillo@134	3070 In any case, the result is an absolute directory path.
meillo@134	3071 The result is a pointer to newly allocated memory.
meillo@134	3072 .BU
meillo@134	3073 .Fu m_maildir()
meillo@134	3074 returns the parameter unchanged if it is an absolute directory path
meillo@134	3075 or begins with the entry `.' or `..'.
meillo@134	3076 All other strings are prepended with the current working directory.
meillo@134	3077 Hence, this functions can not be used for folder paths.
meillo@134	3078 The result is either an absolute directory path or a relative
meillo@134	3079 directory path, starting with a dot.
meillo@134	3080 In contrast to the other functions, the result is a pointer to
meillo@134	3081 static memory.
meillo@134	3082 .P
meillo@134	3083 The situation was obscure, irritating, error-prone, and non-orthogonal.
meillo@134	3084 No clear terminology was used to name the different kinds of path names.
meillo@134	3085 The first argument of
meillo@134	3086 .Fu m_mailpath() ,
meillo@134	3087 for instance, was named
meillo@134	3088 .Ar folder ,
meillo@134	3089 though
meillo@134	3090 .Fu m_mailpath()
meillo@134	3091 can not be used for MH folders.
meillo@134	3092 .P
meillo@134	3093 I reworked the path name conversion completely, introducing clarity.
meillo@134	3094 First of all, the terminology needed to be defined.
meillo@134	3095 A path name is either in the Unix domain, then it is called
meillo@134	3096 \fIdirectory path\fP, `dirpath' for short, or it is in the MH domain,
meillo@134	3097 then it is called \fIfolder path\fP, `folpath' for short.
meillo@134	3098 The two terms need to be used with strict distinction.
meillo@134	3099 Having a clear terminology is often an indicator of having understood
meillo@134	3100 the problem itself.
meillo@134	3101 Second, I exploited the concept of path type indicators.
meillo@134	3102 By requesting every path name to start with a clear type identifier,
meillo@134	3103 conversion between the types can be fully automated.
meillo@134	3104 Thus the tools can accept paths of any type from the user.
meillo@134	3105 Therefore, it was necessary to require relative directory paths to be
meillo@134	3106 prefixed with a dot character.
meillo@134	3107 In consequence, the dot character could no longer be an alias for the
meillo@134	3108 current message.
meillo@134	3109 .Ci cff0e16925e7edbd25b8b9d6d4fbdf03e0e60c01
meillo@134	3110 Third, I created three new functions to replace the previous mess:
meillo@134	3111 .BU
meillo@134	3112 .Fu expandfol()
meillo@134	3113 converts folder paths to absolute folder paths,
meillo@134	3114 without the leading plus character.
meillo@134	3115 Directory paths are simply passed through.
meillo@134	3116 This function is to be used for folder paths only, thus the name.
meillo@134	3117 The result is a pointer to static memory.
meillo@134	3118 .BU
meillo@134	3119 .Fu expanddir()
meillo@134	3120 converts directory paths to absolute directory paths.
meillo@134	3121 Folder paths are treated as relative directory paths.
meillo@134	3122 This function is to be used for directory paths only, thus the name.
meillo@134	3123 The result is a pointer to static memory.
meillo@134	3124 .BU
meillo@134	3125 .Fu toabsdir()
meillo@134	3126 converts any type of path to an absolute directory path.
meillo@134	3127 This is the function of choice for path conversion.
meillo@134	3128 Absolute directory paths are the most general representation of a
meillo@134	3129 path name.
meillo@134	3130 The result is a pointer to static memory.
meillo@134	3131 .P
meillo@134	3132 The new functions have names that indicate their use.
meillo@134	3133 Two of the functions convert relative to absolute path names of the
meillo@134	3134 same type.
meillo@134	3135 The third function converts any path name type to the most general one,
meillo@134	3136 the absolute directory path.
meillo@134	3137 All of the functions return pointers to static memory.
meillo@134	3138 All three functions are implemented in
meillo@134	3139 .Fn sbr/path.c .
meillo@134	3140 .Fn sbr/m_maildir.c
meillo@134	3141 is removed.
meillo@134	3142 .P
meillo@134	3143 Along with the path conversion rework, I also replaced
meillo@134	3144 .Fu getfolder(FDEF)
meillo@134	3145 with
meillo@134	3146 .Fu getdeffol()
meillo@134	3147 and
meillo@134	3148 .Fu getfolder(FCUR)
meillo@134	3149 with
meillo@134	3150 .Fu getcurfol() ,
meillo@134	3151 which is only a convenience wrapper for
meillo@134	3152 .Fu expandfol("@") .
meillo@134	3153 This code was moved from
meillo@134	3154 .Fn sbr/getfolder.c
meillo@134	3155 to
meillo@134	3156 .Fn sbr/path.c .
meillo@134	3157 .P
meillo@134	3158 The related function
meillo@134	3159 .Fu etcpath()
meillo@134	3160 was moved to
meillo@134	3161 .Fn sbr/path.c ,
meillo@134	3162 too.
meillo@134	3163 Previously, it had been located in
meillo@134	3164 .Fn config/config.c ,
meillo@134	3165 for whatever reasons.
meillo@134	3166 .P
meillo@134	3167 .Fn sbr/path.c
meillo@134	3168 now contains all path handling code.
meillo@134	3169 Only 173 lines of code were needed to replace the previous 252 lines.
meillo@134	3170 The readability of the code is highly improved.
meillo@134	3171 Additionally, each of the six exported and one static functions
meillo@134	3172 is introduced by an explaining comment.
meillo@134	3173 .Ci d39e2c447b0d163a5a63f480b23d06edb7a73aa0
meillo@133	3174
meillo@133	3175
meillo@133	3176
meillo@133	3177
meillo@133	3178 .H2 "Profile Reading
meillo@133	3179 .P
meillo@138	3180 The MH profile contains the configuration for the user-specific MH setup.
meillo@138	3181 MH tools read the profile right after starting up,
meillo@138	3182 as it contains the location of the user's mail storage
meillo@138	3183 and similar settings that influence the whole setup.
meillo@138	3184 Further more, the profile contains the default switches for the tools,
meillo@138	3185 hence, it must be read before the command line switches are processed.
meillo@138	3186 .P
meillo@138	3187 For historic reasons, some MH tools did not read the profile and context.
meillo@138	3188 Among them were
meillo@138	3189 .Pn post /\c
meillo@138	3190 .Pn spost ,
meillo@138	3191 .Pn mhmail ,
meillo@138	3192 and
meillo@138	3193 .Pn slocal .
meillo@138	3194 The reason why these tools ignored the profile were not clearly stated.
meillo@138	3195 During the discussion on the nmh-workers mailing list,
meillo@138	3196 .[
meillo@138	3197 nmh-workers levine post profile
meillo@138	3198 .]
meillo@138	3199 David Levine posted an explanation, quoting John Romine:
meillo@138	3200 .QS
meillo@138	3201 I asked John Romine and here's what he had to say, which
meillo@138	3202 agrees and provides an example that convinces me:
meillo@138	3203 .QS
meillo@138	3204 My take on this is that post should not be called by
meillo@138	3205 users directly, and it doesn't read the .mh_profile
meillo@138	3206 (only front-end UI programs read the profile).
meillo@138	3207 .QP
meillo@138	3208 For example, there can be contexts where post is called
meillo@138	3209 by a helper program (like 'mhmail') which may be run by
meillo@138	3210 a non-MH user. We don't want this to prompt the user
meillo@138	3211 to create an MH profile, etc.
meillo@138	3212 .QP
meillo@138	3213 My suggestion would be to have send pass a (hidden)
meillo@138	3214 `\-fileproc proc' option to post if needed. You could also
meillo@138	3215 use an environment variable (I think send/whatnow do
meillo@138	3216 this).
meillo@138	3217 .QE
meillo@138	3218 I think that's the way to go. My personal preference is to use a command line option, not an environment variable.
meillo@138	3219 .QE
meillo@138	3220 .P
meillo@138	3221 To solve the problem of
meillo@138	3222 .Pn post
meillo@138	3223 not honoring the
meillo@138	3224 .Pe fileproc
meillo@138	3225 profile entry,
meillo@138	3226 the community roughly agreed that a switch
meillo@138	3227 .Sw -fileproc
meillo@138	3228 should be added to
meillo@138	3229 .Pn post
meillo@138	3230 to be able to pass a different fileproc.
meillo@138	3231 I strongly disagree with this approach because it does not solve
meillo@138	3232 the problem; it only removes a single symptom.
meillo@138	3233 The problem is that
meillo@138	3234 .Pn post
meillo@138	3235 does not behave as expected.
meillo@138	3236 But all programs should behave as expected.
meillo@138	3237 Clear and simple concepts are a precondition for this.
meillo@138	3238 Hence, the real solution is having all MH tools read the profile.
meillo@138	3239 .P
meillo@138	3240 Yet, the problem has a further aspect.
meillo@138	3241 It mainly originates in
meillo@138	3242 .Pn mhmail .
meillo@138	3243 .Pn mhmail
meillo@138	3244 was intended to be a replacement for
meillo@138	3245 .Pn mailx
meillo@138	3246 on systems with MH installations.
meillo@138	3247 .Pn mhmail
meillo@138	3248 should have been able to use just like
meillo@138	3249 .Pn mailx ,
meillo@138	3250 but sending the message via MH's
meillo@138	3251 .Pn post
meillo@138	3252 instead of
meillo@138	3253 .Pn sendmail .
meillo@138	3254 Using
meillo@138	3255 .Pn mhmail
meillo@138	3256 should not be influenced by the question whether the user had
meillo@138	3257 MH set up for himself or not.
meillo@138	3258 .Pn mhmail
meillo@138	3259 did not read the profile as this requests the user to set up MH
meillo@138	3260 if not done yet.
meillo@138	3261 As
meillo@138	3262 .Pn mhmail
meillo@138	3263 used
meillo@138	3264 .Pn post ,
meillo@138	3265 .Pn post
meillo@138	3266 could not read the profile neither.
meillo@138	3267 This is the reason why
meillo@138	3268 .Pn post
meillo@138	3269 does not read the profile.
meillo@138	3270 This is the reason for the actual problem.
meillo@138	3271 It was not much of a problem because
meillo@138	3272 .Pn post
meillo@138	3273 was not intended to be used by users directly.
meillo@138	3274 .Pn send
meillo@138	3275 is the interactive front-end to
meillo@138	3276 .Pn post .
meillo@138	3277 .Pn send
meillo@138	3278 read the profile and passed all relevant values on the command line to
meillo@138	3279 .Pn post
meillo@138	3280 \(en an awkward solution.
meillo@138	3281 .P
meillo@138	3282 The important insight is that
meillo@138	3283 .Pn mhmail
meillo@138	3284 is no true MH tool.
meillo@138	3285 The concepts broke because this outlandish tool was treated as any other
meillo@138	3286 MH tool.
meillo@138	3287 Instead it should have been treated accordingly to its foreign style.
meillo@138	3288 The solution is not to prevent the tools reading the profile but
meillo@138	3289 to instruct them reading a different profile.
meillo@138	3290 .Pn mhmail
meillo@138	3291 could have set up a well-defined profile and caused all MH tools
meillo@138	3292 in the session use it by exporting an environment variable.
meillo@138	3293 With this approach, no special cases would have been introduced,
meillo@138	3294 no surprises would have been caused.
meillo@138	3295 By writing a clean-profile-wrapper, the concept could have been
meillo@138	3296 generalized orthogonally to the whole MH toolchest.
meillo@138	3297 Then Rose's motivation behind the decision that
meillo@138	3298 .Pn post
meillo@138	3299 ignores the profile, as quoted by Jeffrey Honig,
meillo@138	3300 .[
meillo@138	3301 nmh-workers post profile
meillo@138	3302 .]
meillo@138	3303 would have become possible:
meillo@138	3304 .QS
meillo@138	3305 when you run mh commands in a script, you want all the defaults to be
meillo@138	3306 what the man page says.
meillo@138	3307 when you run a command by hand, then you want your own defaults...
meillo@138	3308 .QE
meillo@138	3309 .LP
meillo@138	3310 Yet, I consider this explanation short-sighted.
meillo@138	3311 We should rather regard theses two cases as just two different MH setups,
meillo@138	3312 based on two different profiles.
meillo@138	3313 Mapping such problems on the concepts of switching between different
meillo@138	3314 profiles, solves them once for all.
meillo@138	3315 .P
meillo@138	3316 In mmh, the wish to have
meillo@138	3317 .Pn mhmail
meillo@138	3318 as as replacement for
meillo@138	3319 .Pn mailx
meillo@138	3320 is considered obsolete.
meillo@138	3321 Mmh's
meillo@138	3322 .Pn mhmail
meillo@138	3323 does no longer cover this use-case.
meillo@138	3324 Currently,
meillo@138	3325 .Pn mhmail
meillo@138	3326 is in a transition state.
meillo@138	3327 .Ci 32d4f9daaa70519be3072479232ff7be0500d009
meillo@138	3328 It may become a front-end to
meillo@138	3329 .Pn comp ,
meillo@138	3330 which provides an interface more convenient in some cases.
meillo@138	3331 In this case,
meillo@138	3332 .Pn mhmail
meillo@138	3333 will become an ordinary MH tool, reading the profile.
meillo@138	3334 If, however, this idea will not convince, then
meillo@138	3335 .Pn mhmail
meillo@138	3336 will be removed.
meillo@138	3337 .P
meillo@138	3338 Every program in the mmh toolchest reads the profile.
meillo@138	3339 The only exception is
meillo@138	3340 .Pn slocal ,
meillo@138	3341 which is not considered part of the mmh toolchest.
meillo@138	3342 This MDA is only distributed with mmh, currently.
meillo@138	3343 Mmh has no
meillo@138	3344 .Pn post
meillo@138	3345 program, but
meillo@138	3346 .Pn spost ,
meillo@138	3347 which now reads the profile.
meillo@138	3348 .Ci 3e017a7abbdf69bf0dff7a4073275961eda1ded8
meillo@138	3349 With this change,
meillo@138	3350 .Pn send
meillo@138	3351 and
meillo@138	3352 .Pn spost
meillo@138	3353 can be considered to be merged.
meillo@138	3354 Direct invocations of
meillo@138	3355 .Pn spost
meillo@138	3356 are only done by the to-be-changed
meillo@138	3357 .Pn mhmail
meillo@138	3358 implementation and by
meillo@138	3359 .Pn rcvdist ,
meillo@138	3360 which will require rework.
meillo@138	3361 .P
meillo@138	3362 The
meillo@138	3363 .Fu context_foil()
meillo@138	3364 function to pretend to have read an empty profile was removed.
meillo@138	3365 .Ci 68af8da96bea87a5541988870130b6209ce396f6
meillo@138	3366 All mmh tools read the profile.
meillo@133	3367
meillo@133	3368
meillo@127	3369
meillo@121	3370 .H2 "Standard Libraries
meillo@22	3371 .P
meillo@121	3372 MH is one decade older than the POSIX and ANSI C standards.
meillo@121	3373 Hence, MH included own implementations of functions
meillo@121	3374 that are standardized and thus widely available today,
meillo@121	3375 but were not back then.
meillo@121	3376 Today, twenty years after the POSIX and ANSI C were published,
meillo@121	3377 developers can expect system to comply with these standards.
meillo@121	3378 In consequence, MH-specific replacements for standard functions
meillo@121	3379 can and should be dropped.
meillo@121	3380 Kernighan and Pike advise: ``Use standard libraries.''
meillo@121	3381 .[ [
meillo@121	3382 kernighan pike practice of programming
meillo@121	3383 .], p. 196]
meillo@121	3384 Actually, MH had followed this advice in history,
meillo@121	3385 but it had not adjusted to the changes in this field.
meillo@121	3386 The
meillo@121	3387 .Fu snprintf()
meillo@121	3388 function, for instance, was standardized with C99 and is available
meillo@121	3389 almost everywhere because of its high usefulness.
meillo@123	3390 In project's own implementation of
meillo@121	3391 .Fu snprintf()
meillo@123	3392 was dropped in March 2012 in favor for using the one of the
meillo@123	3393 standard library.
meillo@123	3394 .Ci 0052f1024deb0a0a2fc2e5bacf93d45a5a9c9b32
meillo@123	3395 Such decisions limit the portability of mmh
meillo@121	3396 if systems don't support these standardized and widespread functions.
meillo@123	3397 This compromise is made because mmh focuses on the future.
meillo@121	3398 .P
meillo@123	3399 I am not yet thirty years old and my C and Unix experience comprises
meillo@123	3400 only half a dozen years.
meillo@121	3401 Hence, I need to learn about the history in retrospective.
meillo@121	3402 I have not used those ancient constructs myself.
meillo@121	3403 I have not suffered from their incompatibilities.
meillo@121	3404 I have not longed for standardization.
meillo@121	3405 All my programming experience is from a time when ANSI C and POSIX
meillo@121	3406 were well established already.
meillo@121	3407 I have only read a lot of books about the (good) old times.
meillo@121	3408 This puts me in a difficult positions when working with old code.
meillo@123	3409 I need to freshly acquire knowledge about old code constructs and ancient
meillo@123	3410 programming styles, whereas older programmers know these things by
meillo@123	3411 heart from their own experience.
meillo@121	3412 .P
meillo@123	3413 Being aware of the situation, I rather let people with more historic
meillo@123	3414 experience replace ancient code constructs with standardized ones.
meillo@121	3415 Lyndon Nerenberg covered large parts of this task for the nmh project.
meillo@121	3416 He converted project-specific functions to POSIX replacements,
meillo@121	3417 also removing the conditionals compilation of now standardized features.
meillo@123	3418 Ken Hornstein and David Levine had their part in the work, too.
meillo@121	3419 Often, I only needed to pull over changes from nmh into mmh.
meillo@121	3420 These changes include many commits; these are among them:
meillo@121	3421 .Ci 768b5edd9623b7238e12ec8dfc409b82a1ed9e2d
meillo@121	3422 .Ci 0052f1024deb0a0a2fc2e5bacf93d45a5a9c9b32 .
meillo@102	3423 .P
meillo@123	3424 During my own work, I tidied up the \fIMH standard library\fP,
meillo@123	3425 .Fn libmh.a ,
meillo@123	3426 which is located in the
meillo@123	3427 .Fn sbr
meillo@123	3428 (``subroutines'') directory in the source tree.
meillo@123	3429 The MH library includes functions that mmh tools usually need.
meillo@123	3430 Among them are MH-specific functions for profile, context, sequence,
meillo@123	3431 and folder handling, but as well
meillo@123	3432 MH-independent functions, such as auxiliary string functions,
meillo@123	3433 portability interfaces and error-checking wrappers for critical
meillo@123	3434 functions of the standard library.
meillo@123	3435 .P
meillo@123	3436 I have replaced the
meillo@121	3437 .Fu atooi()
meillo@121	3438 function with calls to
meillo@123	3439 .Fu strtoul()
meillo@139	3440 with the third parameter, the base, set to eight.
meillo@121	3441 .Fu strtoul()
meillo@123	3442 is part of C89 and thus considered safe to use.
meillo@121	3443 .Ci c490c51b3c0f8871b6953bd0c74551404f840a74
meillo@102	3444 .P
meillo@121	3445 I did remove project-included fallback implementations of
meillo@121	3446 .Fu memmove()
meillo@121	3447 and
meillo@121	3448 .Fu strerror() ,
meillo@121	3449 although Peter Maydell had re-included them into nmh in 2008
meillo@121	3450 to support SunOS 4.
meillo@121	3451 Nevertheless, these functions are part of ANSI C.
meillo@121	3452 Systems that do not even provide full ANSI C support should not
meillo@121	3453 put a load on mmh.
meillo@121	3454 .Ci b067ff5c465a5d243ce5a19e562085a9a1a97215
meillo@121	3455 .P
meillo@121	3456 The
meillo@121	3457 .Fu copy()
meillo@121	3458 function copies the string in argument one to the location in two.
meillo@121	3459 In contrast to
meillo@121	3460 .Fu strcpy() ,
meillo@121	3461 it returns a pointer to the terminating null-byte in the destination area.
meillo@123	3462 The code was adjusted to replace
meillo@121	3463 .Fu copy()
meillo@123	3464 with
meillo@121	3465 .Fu strcpy() ,
meillo@121	3466 except within
meillo@121	3467 .Fu concat() ,
meillo@121	3468 where
meillo@121	3469 .Fu copy()
meillo@123	3470 was more convenient.
meillo@123	3471 Therefore, the definition of
meillo@121	3472 .Fu copy()
meillo@123	3473 was moved into the source file of
meillo@121	3474 .Fu concat()
meillo@123	3475 and its visibility is now limited to it.
meillo@121	3476 .Ci 552fd7253e5ee9e554c5c7a8248a6322aa4363bb
meillo@121	3477 .P
meillo@121	3478 The function
meillo@121	3479 .Fu r1bindex()
meillo@121	3480 had been a generalized version of
meillo@121	3481 .Fu basename()
meillo@121	3482 with minor differences.
meillo@121	3483 As all calls to
meillo@121	3484 .Fu r1bindex()
meillo@121	3485 had the slash (`/') as delimiter anyway,
meillo@121	3486 replacing
meillo@121	3487 .Fu r1bindex()
meillo@121	3488 with the more specific and better-named function
meillo@121	3489 .Fu basename()
meillo@121	3490 became desirable.
meillo@121	3491 Unfortunately, many of the 54 calls to
meillo@121	3492 .Fu r1bindex()
meillo@123	3493 depended on a special behavior,
meillo@121	3494 which differed from the POSIX specification for
meillo@121	3495 .Fu basename() .
meillo@121	3496 Hence,
meillo@121	3497 .Fu r1bindex()
meillo@121	3498 was kept but renamed to
meillo@123	3499 .Fu mhbasename() ,
meillo@123	3500 fixing the delimiter to the slash.
meillo@121	3501 .Ci 240013872c392fe644bd4f79382d9f5314b4ea60
meillo@121	3502 For possible uses of
meillo@121	3503 .Fu r1bindex()
meillo@121	3504 with a different delimiter,
meillo@121	3505 the ANSI C function
meillo@121	3506 .Fu strrchr()
meillo@121	3507 provides the core functionality.
meillo@121	3508 .P
meillo@121	3509 The
meillo@121	3510 .Fu ssequal()
meillo@121	3511 function \(en apparently for ``substring equal'' \(en
meillo@121	3512 was renamed to
meillo@121	3513 .Fu isprefix() ,
meillo@121	3514 because this is what it actually checks.
meillo@121	3515 .Ci c20b4fa14515c7ab388ce35411d89a7a92300711
meillo@121	3516 Its source file had included the following comments, no joke.
meillo@121	3517 .VS
meillo@121	3518 /*
meillo@121	3519 * THIS CODE DOES NOT WORK AS ADVERTISED.
meillo@121	3520 * It is actually checking if s1 is a PREFIX of s2.
meillo@121	3521 * All calls to this function need to be checked to see
meillo@121	3522 * if that needs to be changed. Prefix checking is cheaper, so
meillo@121	3523 * should be kept if it's sufficient.
meillo@121	3524 */
meillo@121	3525
meillo@121	3526 /*
meillo@121	3527 * Check if s1 is a substring of s2.
meillo@121	3528 * If yes, then return 1, else return 0.
meillo@121	3529 */
meillo@121	3530 VE
meillo@123	3531 Two months later, it was completely removed by replacing it with
meillo@123	3532 .Fu strncmp() .
meillo@123	3533 .Ci b0b1dd37ff515578cf7cba51625189eb34a196cb
meillo@121	3534
meillo@102	3535
meillo@102	3536
meillo@102	3537
meillo@133	3538
meillo@133	3539 .H2 "User Data Locations
meillo@133	3540 .P
meillo@133	3541 In nmh, a personal setup consists of the MH profile and the MH directory.
meillo@133	3542 The profile is a file named
meillo@133	3543 .Fn \&.mh_profile
meillo@133	3544 in the user's home directory.
meillo@133	3545 It contains the static configuration.
meillo@133	3546 It also contains the location of the MH directory in the profile entry
meillo@133	3547 .Pe Path .
meillo@133	3548 The MH directory contains the mail storage and is the first
meillo@133	3549 place to search for personal forms, scan formats, and similar
meillo@133	3550 configuration files.
meillo@133	3551 The location of the MH directory can be chosen freely by the user.
meillo@133	3552 The default and usual name is a directory named
meillo@133	3553 .Fn Mail
meillo@133	3554 in the home directory.
meillo@133	3555 .P
meillo@133	3556 The way MH data is splitted between profile and MH directory is a legacy.
meillo@133	3557 It is only sensible in a situation where the profile is the only
meillo@133	3558 configuration file.
meillo@133	3559 Why else should the mail storage and the configuration files be intermixed?
meillo@133	3560 They are different kinds of data:
meillo@133	3561 The data to be operated on and the configuration to change how
meillo@133	3562 tools operate.
meillo@133	3563 Splitting the configuration between the profile and the MH directory
meillo@133	3564 is bad.
meillo@133	3565 Merging the mail storage and the configuration in one directory is bad
meillo@133	3566 as well.
meillo@133	3567 As the mail storage and the configuration were not separated sensibly
meillo@133	3568 in the first place, I did it now.
meillo@133	3569 .P
meillo@133	3570 Personal mmh data is grouped by type, resulting in two distinct parts:
meillo@133	3571 The mail storage and the configuration.
meillo@133	3572 In mmh, the mail storage directory still contains all the messages,
meillo@133	3573 but, in exception of public sequences files, nothing else.
meillo@133	3574 In difference to nmh, the auxiliary configuration files are no longer
meillo@133	3575 located there.
meillo@133	3576 Therefore, the directory is no longer called the user's \fIMH directory\fP
meillo@133	3577 but his \fImail storage\fP.
meillo@133	3578 Its location is still user-chosen, with the default name
meillo@133	3579 .Fn Mail ,
meillo@133	3580 in the user's home directory.
meillo@133	3581 In mmh, the configuration is grouped together in
meillo@133	3582 the hidden directory
meillo@133	3583 .Fn \&.mmh
meillo@133	3584 in the user's home directory.
meillo@133	3585 This \fImmh directory\fP contains the context file, personal forms,
meillo@133	3586 scan formats, and the like, but also the user's profile, now named
meillo@133	3587 .Fn profile .
meillo@133	3588 The location of the profile is no longer fixed to
meillo@133	3589 .Fn $HOME/.mh_profile
meillo@133	3590 but to
meillo@133	3591 .Fn $HOME/.mmh/profile .
meillo@133	3592 Having both, the file
meillo@133	3593 .Fn $HOME/.mh_profile
meillo@133	3594 and the configuration directory
meillo@133	3595 .Fn $HOME/.mmh
meillo@133	3596 appeared to be inconsistent.
meillo@133	3597 The approach chosen for mmh is consistent, simple, and familiar to
meillo@133	3598 Unix users.
meillo@133	3599 .P
meillo@133	3600 MH allows users to have multiiple MH setups.
meillo@133	3601 Therefore, it is necessary to select a different profile.
meillo@133	3602 The profile is the single entry point to access the rest of a
meillo@133	3603 personal MH setup.
meillo@133	3604 In nmh, the environment variable
meillo@133	3605 .Ev MH
meillo@133	3606 could be used to specifiy a different profile.
meillo@133	3607 To operate in the same MH setup with a separate context,
meillo@133	3608 the
meillo@133	3609 .Ev MHCONTEXT
meillo@133	3610 environment variable could be used.
meillo@133	3611 This allows having own current folders and current messages in
meillo@133	3612 each terminal, for instance.
meillo@133	3613 In mmh, three environment variables are used.
meillo@133	3614 .Ev MMH
meillo@133	3615 overrides the default location of the mmh directory (\c
meillo@133	3616 .Fn .mmh ).
meillo@133	3617 .Ev MMHP
meillo@133	3618 and
meillo@133	3619 .Ev MMHC
meillo@133	3620 override the paths to the profile and context files, respectively.
meillo@133	3621 This approach allows the set of personal configuration files to be chosen
meillo@133	3622 independently from the profile, context, and mail storage.
meillo@133	3623 .P
meillo@133	3624 The separation of the files by type is sensible and convenient.
meillo@133	3625 The new approach has no functional disadvantages,
meillo@133	3626 as every setup I can imagine can be implemented with both approaches,
meillo@133	3627 possibly even easier with the new approach.
meillo@133	3628 The main achievement of the change is the clear and sensible split
meillo@133	3629 between mail storage and configuration.
meillo@133	3630
meillo@133	3631
meillo@133	3632
meillo@133	3633
meillo@133	3634
meillo@118	3635 .H2 "Modularization
meillo@118	3636 .P
meillo@123	3637 The source code of the mmh tools is located in the
meillo@122	3638 .Fn uip
meillo@123	3639 (``user interface programs'') directory.
meillo@123	3640 Each tools has a source file with the same name.
meillo@122	3641 For example,
meillo@122	3642 .Pn rmm
meillo@122	3643 is built from
meillo@122	3644 .Fn uip/rmm.c .
meillo@123	3645 Some source files are used for multiple programs.
meillo@122	3646 For example
meillo@122	3647 .Fn uip/scansbr.c
meillo@123	3648 is used for both,
meillo@122	3649 .Pn scan
meillo@122	3650 and
meillo@122	3651 .Pn inc .
meillo@122	3652 In nmh, 49 tools were built from 76 source files.
meillo@123	3653 This is a ratio of 1.6 source files per program.
meillo@123	3654 32 programs depended on multiple source files;
meillo@123	3655 17 programs depended on one source file only.
meillo@122	3656 In mmh, 39 tools are built from 51 source files.
meillo@123	3657 This is a ratio of 1.3 source files per program.
meillo@123	3658 18 programs depend on multiple source files;
meillo@123	3659 21 programs depend on one source file only.
meillo@123	3660 (These numbers and the ones in the following text ignore the MH library
meillo@123	3661 as well as shell scripts and multiple names for the same program.)
meillo@122	3662 .P
meillo@123	3663 Splitting the source code of a large program into multiple files can
meillo@122	3664 increase the readability of its source code.
meillo@124	3665 Most of the mmh tools, however, are simple and straight-forward programs.
meillo@122	3666 With the exception of the MIME handling tools,
meillo@122	3667 .Pn pick
meillo@122	3668 is the largest tools.
meillo@122	3669 It contains 1\\|037 lines of source code (measured with
meillo@122	3670 .Pn sloccount ), excluding the MH library.
meillo@122	3671 Only the MIME handling tools (\c
meillo@122	3672 .Pn mhbuild ,
meillo@122	3673 .Pn mhstore ,
meillo@122	3674 .Pn show ,
meillo@122	3675 etc.)
meillo@122	3676 are larger.
meillo@122	3677 Splitting programs with less than 1\\|000 lines of code into multiple
meillo@123	3678 source files seldom leads to better readability.
meillo@123	3679 For such tools, splitting makes sense
meillo@122	3680 when parts of the code are reused in other programs,
meillo@122	3681 and the reused code fragment is not general enough
meillo@122	3682 for including it in the MH library,
meillo@124	3683 or, if the code has dependencies on a library that only few programs need.
meillo@122	3684 .Fn uip/packsbr.c ,
meillo@122	3685 for instance, provides the core program logic for the
meillo@122	3686 .Pn packf
meillo@122	3687 and
meillo@122	3688 .Pn rcvpack
meillo@122	3689 programs.
meillo@122	3690 .Fn uip/packf.c
meillo@122	3691 and
meillo@122	3692 .Fn uip/rcvpack.c
meillo@122	3693 mainly wrap the core function appropriately.
meillo@122	3694 No other tools use the folder packing functions.
meillo@123	3695 As another example,
meillo@123	3696 .Fn uip/termsbr.c
meillo@123	3697 provides termcap support, which requires linking with a termcap or
meillo@123	3698 curses library.
meillo@123	3699 Including
meillo@123	3700 .Fn uip/termsbr.c
meillo@123	3701 into the MH library would require every program to be linked with
meillo@123	3702 termcap or curses, although only few of the programs require it.
meillo@122	3703 .P
meillo@122	3704 The task of MIME handling is complex enough that splitting its code
meillo@122	3705 into multiple source files improves the readability.
meillo@122	3706 The program
meillo@122	3707 .Pn mhstore ,
meillo@122	3708 for instance, is compiled out of seven source files with 2\\|500
meillo@122	3709 lines of code in summary.
meillo@122	3710 The main code file
meillo@122	3711 .Fn uip/mhstore.c
meillo@123	3712 consists of 800 lines; the other 1\\|700 lines of code are reused in
meillo@123	3713 other MIME handling tools.
meillo@123	3714 It seems to be worthwhile to bundle the generic MIME handling code into
meillo@123	3715 a MH-MIME library, as a companion to the MH standard library.
meillo@122	3716 This is left open for the future.
meillo@122	3717 .P
meillo@123	3718 The work already done, focussed on the non-MIME tools.
meillo@122	3719 The amount of code compiled into each program was reduced.
meillo@123	3720 This eases the understanding of the code base.
meillo@122	3721 In nmh,
meillo@122	3722 .Pn comp
meillo@122	3723 was built from six source files:
meillo@122	3724 .Fn comp.c ,
meillo@122	3725 .Fn whatnowproc.c ,
meillo@122	3726 .Fn whatnowsbr.c ,
meillo@122	3727 .Fn sendsbr.c ,
meillo@122	3728 .Fn annosbr.c ,
meillo@122	3729 and
meillo@122	3730 .Fn distsbr.c .
meillo@122	3731 In mmh, it builds from only two:
meillo@122	3732 .Fn comp.c
meillo@122	3733 and
meillo@122	3734 .Fn whatnowproc.c .
meillo@123	3735 In nmh's
meillo@123	3736 .Pn comp ,
meillo@123	3737 the core function of
meillo@122	3738 .Pn whatnow ,
meillo@122	3739 .Pn send ,
meillo@122	3740 and
meillo@122	3741 .Pn anno
meillo@123	3742 were compiled into
meillo@122	3743 .Pn comp .
meillo@123	3744 This saved the need to execute these programs with
meillo@122	3745 .Fu fork()
meillo@122	3746 and
meillo@122	3747 .Fu exec() ,
meillo@122	3748 two expensive system calls.
meillo@122	3749 Whereis this approach improved the time performance,
meillo@122	3750 it interweaved the source code.
meillo@122	3751 Core functionalities were not encapsulated into programs but into
meillo@122	3752 function, which were then wrapped by programs.
meillo@122	3753 For example,
meillo@122	3754 .Fn uip/annosbr.c
meillo@122	3755 included the function
meillo@122	3756 .Fu annotate() .
meillo@122	3757 Each program that wanted to annotate messages, included the source file
meillo@123	3758 .Fn uip/annosbr.c
meillo@123	3759 and called
meillo@123	3760 .Fu annotate() .
meillo@123	3761 Because the function
meillo@123	3762 .Fu annotate()
meillo@123	3763 was used like the tool
meillo@123	3764 .Pn anno ,
meillo@123	3765 it had seven parameters, reflecting the command line switches of the tool.
meillo@122	3766 When another pair of command line switches was added to
meillo@122	3767 .Pn anno ,
meillo@122	3768 a rather ugly hack was implemented to avoid adding another parameter
meillo@122	3769 to the function.
meillo@122	3770 .Ci d9b1d57351d104d7ec1a5621f090657dcce8cb7f
meillo@122	3771 .P
meillo@122	3772 Separation simplifies the understanding of program code
meillo@122	3773 because the area influenced by any particular statement is smaller.
meillo@122	3774 The separating on the program-level is more strict than the separation
meillo@122	3775 on the function level.
meillo@122	3776 In mmh, the relevant code of
meillo@122	3777 .Pn comp
meillo@122	3778 comprises the two files
meillo@122	3779 .Fn uip/comp.c
meillo@122	3780 and
meillo@122	3781 .Fn uip/whatnowproc.c ,
meillo@123	3782 together 210 lines of code.
meillo@122	3783 In nmh,
meillo@122	3784 .Pn comp
meillo@122	3785 comprises six files with 2\\|450 lines.
meillo@123	3786 Not all of the code in these six files was actually used by
meillo@122	3787 .Pn comp ,
meillo@123	3788 but the code reader needed to read all of the code first to know which
meillo@123	3789 parts were used.
meillo@122	3790 .P
meillo@123	3791 As I have read a lot in the code base during the last two years,
meillo@123	3792 I learned about the easy and the difficult parts.
meillo@123	3793 Code is easy to understand if:
meillo@123	3794 .BU
meillo@139	3795 The influenced code area is small.
meillo@123	3796 .BU
meillo@139	3797 The boundaries are strictly defined.
meillo@123	3798 .BU
meillo@139	3799 The code is written straight-forward.
meillo@123	3800 .P
meillo@123	3801 .\" XXX move this paragraph somewhere else?
meillo@123	3802 Reading
meillo@122	3803 .Pn rmm 's
meillo@122	3804 source code in
meillo@122	3805 .Fn uip/rmm.c
meillo@122	3806 is my recommendation for a beginner's entry point into the code base of nmh.
meillo@122	3807 The reasons are that the task of
meillo@122	3808 .Pn rmm
meillo@122	3809 is straight forward and it consists of one small source code file only,
meillo@122	3810 yet its source includes code constructs typical for MH tools.
meillo@122	3811 With the introduction of the trash folder in mmh,
meillo@122	3812 .Pn rmm
meillo@122	3813 became a bit more complex, because it invokes
meillo@122	3814 .Pn refile .
meillo@122	3815 Still, it is a good example for a simple tool with clear sources.
meillo@122	3816 .P
meillo@122	3817 Understanding
meillo@122	3818 .Pn comp
meillo@122	3819 requires to read 210 lines of code in mmh, but ten times as much in nmh.
meillo@123	3820 Due to the aforementioned hack in
meillo@122	3821 .Pn anno
meillo@122	3822 to save the additional parameter, information passed through the program's
meillo@122	3823 source base in obscure ways.
meillo@123	3824 Thus, understanding
meillo@122	3825 .Pn comp ,
meillo@123	3826 required understanding the inner workings of
meillo@122	3827 .Fn uip/annosbr.c
meillo@122	3828 first.
meillo@123	3829 To be sure to fully understand a program, its whole source code needs
meillo@122	3830 to be examined.
meillo@123	3831 Not doing so is a leap of faith, assuming that the developers
meillo@122	3832 have avoided obscure programming techniques.
meillo@122	3833 By separating the tools on the program-level, the boundaries are
meillo@122	3834 clearly visible and technically enforced.
meillo@122	3835 The interfaces are calls to
meillo@122	3836 .Fu exec()
meillo@122	3837 rather than arbitrary function calls.
meillo@123	3838 .P
meillo@123	3839 But the real problem is another:
meillo@123	3840 Nmh violates the golden ``one tool, one job'' rule of the Unix philosophy.
meillo@123	3841 Understanding
meillo@122	3842 .Pn comp
meillo@123	3843 requires understanding
meillo@123	3844 .Fn uip/annosbr.c
meillo@123	3845 and
meillo@123	3846 .Fn uip/sendsbr.c
meillo@123	3847 because
meillo@123	3848 .Pn comp
meillo@123	3849 does annotate and send messages.
meillo@123	3850 In nmh, there surely exists the tool
meillo@122	3851 .Pn send ,
meillo@123	3852 which does (almost) only send messages.
meillo@123	3853 But
meillo@122	3854 .Pn comp
meillo@123	3855 and
meillo@122	3856 .Pn repl
meillo@122	3857 and
meillo@122	3858 .Pn forw
meillo@122	3859 and
meillo@122	3860 .Pn dist
meillo@122	3861 and
meillo@122	3862 .Pn whatnow
meillo@122	3863 and
meillo@123	3864 .Pn viamail ,
meillo@123	3865 they all (!) have the same message sending function included, too.
meillo@123	3866 In result,
meillo@123	3867 .Pn comp
meillo@123	3868 sends messages without using
meillo@123	3869 .Pn send .
meillo@123	3870 The situation is the same as if
meillo@123	3871 .Pn grep
meillo@123	3872 would page without
meillo@123	3873 .Pn more
meillo@123	3874 just because both programs are part of the same code base.
meillo@123	3875 .P
meillo@122	3876 The clear separation on the surface \(en the toolchest approach \(en
meillo@123	3877 is violated on the level below.
meillo@122	3878 This violation is for the sake of time performance.
meillo@122	3879 On systems where
meillo@122	3880 .Fu fork()
meillo@122	3881 and
meillo@122	3882 .Fu exec()
meillo@122	3883 are expensive, the quicker response might be noticable.
meillo@124	3884 In the old times, sacrificing readability and conceptional beauty for
meillo@124	3885 speed might even have been a must to prevent MH from being unusably slow.
meillo@122	3886 Whatever the reasons had been, today they are gone.
meillo@123	3887 No longer should we sacrifice readability or conceptional beauty.
meillo@122	3888 No longer should we violate the Unix philosophy's ``one tool, one job''
meillo@122	3889 guideline.
meillo@123	3890 No longer should we keep speed improvements that became unnecessary.
meillo@122	3891 .P
meillo@123	3892 Therefore, mmh's
meillo@123	3893 .Pn comp
meillo@123	3894 does no longer send messages.
meillo@123	3895 In mmh, different jobs are divided among separate programs that
meillo@122	3896 invoke each other as needed.
meillo@123	3897 In consequence,
meillo@123	3898 .Pn comp
meillo@123	3899 invokes
meillo@123	3900 .Pn whatnow
meillo@123	3901 which thereafter invokes
meillo@123	3902 .Pn send .
meillo@123	3903 The clear separation on the surface is maintained on the level below.
meillo@123	3904 Human users and the tools use the same interface \(en
meillo@123	3905 annotations, for example, are made by invoking
meillo@123	3906 .Pn anno ,
meillo@123	3907 no matter if requested by programs or by human beings.
meillo@123	3908 The decrease of tools built from multiple source files and thus
meillo@123	3909 the decrease of
meillo@123	3910 .Fn uip/*sbr.c
meillo@123	3911 files confirm the improvement.
meillo@123	3912 .P
meillo@123	3913 One disadvantage needs to be taken with this change:
meillo@123	3914 The compiler can no longer check the integrity of the interfaces.
meillo@123	3915 By changing the command line interfaces of tools, it is
meillo@123	3916 the developer's job to adjust the invocations of these tools as well.
meillo@123	3917 As this is a manual task and regression tests, which could detect such
meillo@124	3918 problems, are not available yet, it is prone to errors.
meillo@123	3919 These errors will not be detected at compile time but at run time.
meillo@123	3920 Installing regression tests is a task left to do.
meillo@123	3921 In the best case, a uniform way of invoking tools from other tools
meillo@123	3922 can be developed to allow automated testing at compile time.