3 Rename S-nail to S-mailx in v15.0, change things i've messed with
4 a single, massively backward incompatible change.
6 Release S-mailx v20 on 2020-03-25, the 42nd anniversary of BSD Mail.
7 With a clean, conforming and efficient codebase, then.
9 In general the code is in a pretty bad shape due to the signal handling.
10 I should have sat back in 2012/13 and consider what i am doing.
11 My fault. If i would, we would have a blocked signal mask anywhere in
12 this software except in a few cases where it is necessary and/or
13 possible to deal with signals, and possibly we would not even have to
14 consider to switch the entire codebase to (the much superior, and the
15 only sane approach) SysV signal handling, without SA_RESTART.
17 But some things are already pretty good, except for normal iterations
18 and a review once we have a better signal handling, and can be taken
19 with us. For example termcap.c, most of tty.c, shexp.c, memory.c.
20 It all has to be converted to use a m_ prefix and follow the coding
21 style of the mentioned. Note n_(UN)?LIKELY() usage is yet a bit weird,
22 we should not use it for entire conditions but instead repeat its use
23 for each subcondition.
25 - We should have generic ENOMEM conditions, now that we have $!.
26 I.e., test overflow (e.g., nam-a-grp.c, whether an alias _can_ be
27 created / extended), like n_ENOMEM_CHECK(INTTYPE, SIZE1, SIZE2, NULL
28 or message), which returns m_bool (now bool_t).
29 Callers need to be aware of NULL returns and pass through errors,
32 - We need a "void" box that can be jumped to, i.e., a state in which no box
35 -- When a MBOX mailbox is removed while it is opened then changing the
36 folder is not possible. This is an inherent problem of the Berkeley
37 Mail codebase, and we need to have a fully functional intermediate
38 VOID box mechanism plus an object-based mailbox implementation to
41 -- Also, when the folder was modified concurrently we should bail, or,
42 in an interactive session, prompt the user what to do.
44 - IDNA decoding. Needs a complete design change.
45 (Unless wants to brute force decode anything before display, of course.)
47 - If pipes fail for part viewers then at least the usual PART X.Y should be
48 shown, maybe even including some error message.
49 I had 'set pipe-text/html="lynx -dump -force_html /dev/stdin"' but NetBSD
50 does not have lynx(1), and i thought i've found a S-nail(1) bug.
52 - It would be nice if it would be possible to define a format string for
53 *quote*, like 'set quote="format=some formats"'.
55 - Line editing should gain possibility of context sensitive tab completion.
57 - Maybe there should be an additional ZOMBIE directive that is served in
58 equal spirit to DEAD, but that could be a valid MBOX... ?
59 What i want is a *real* resend, best if possible from command line.
60 Meaning, also the possibility to postpone a message. In general.
62 - Having a newsreader would be a really cool thing. (RFC 977 and 2980)
64 - There should be a variable that controls whether leading and trailing
65 empty lines of parts and/or messages as such should be printed or not.
67 - printhead()/hprf(): support %n newline format (%t tab?).
68 Make it possible to use the *datefield* algorithm for plain From_ derived
69 dates (needs a From_ parser, i.e., strptime()-alike).
70 Once we have that, rename *datefield-markout-older* to
71 *date-markout-older* ??
72 Note that NetBSD's mail(1) has some other nice things.
73 Note also that our code is quite unflexible.
75 -- NetBSD's mail(1) has nice *indentprefix* and *indentpostscript*
76 variables (though prefix and appendix or prefix and suffix, but..).
77 Note that our code is quite unflexible.
79 - headerpick: add resend-retain/ignore! (Ralph Corderoy, Norman Shapiro)
80 (Delivered-To thread on nmh. Will be hard to do because of
86 - Improve name extraction rules. And field parsing. There
87 are structured and unstructured fields. There are quoted pairs and
88 comments etc. Rewrite the entire parsing mechanism to comply to RFC
89 5322, and try to merge all those many subparsers around in the codebase,
90 and accordingly. So much duplicated work ...
91 Name parsing has been improved a bit for v13, but it's still broken.
92 yankword(), *extract(), etc.: RFC 5322 says that comments in address
93 fields SHOULD NOT be used (mutt(1) maps them to full name-addr forms if
94 approbiate, even if that actually changes content!!?), and that full
95 name-addr SHOULD be used. Our functions are yet quite silly (i.e.,
96 leading comments remain, as in "(bier2) <a2@b2.de>", unless the address
97 doesn't come in angle brackets, trailing go away, as in "<a6@b6.de>
98 (bier6)", that becomes "<a6@b6.de>").
100 (co$mm1) abc@däf.de (cö,mm,2) ('c'o"m"m.3)
101 Should eventually become
102 co$mm1 cö,mm,2 'c'o"m"m.3 <abc@xn--df-via.de>
103 on the display, or, with IDNA decoding (and thus rather unlikely)
104 co$mm1 cö,mm,2 'c'o"m"m.3 <abc@däf.de>
105 It should NOT become this mutt(1)ism:
106 "co$mm1 cö,mm,2 'c'omm.3" <abc@däf.de>
109 - many uses of whitechar() should be spacechar(), and spacechar() should
110 be blankspacechar(), and blankspacechar() should be dropped. then
111 drop onlywhitechar() again, too. maybe reduce char_class bits then.
113 - In v15.0, when we can address attachments of a message individually,
114 it would be nice to provide even more access, just like nmh(1) does
115 (Johan Commelin: Are s-nail and mh related?).
117 - I never used anything but the *datefield* option, and it would really be
118 nice if the date strings would be parsed off into some 16 byte or what
119 storage when about to producing the summary, so that it would be directly
120 available and there would be no need to reread the mail. Moreover, or
121 even more than that - the m_date field exists and should possibly simply
122 be init, at least in these cases. (P.S.: this doesn't contradict the
123 statement somewhere else in this file that the structure should be
124 slacked; simply use multiple thereof or so)
126 - After I/O layer rework we should optionally be able to read RSS
127 (Atom?) feeds -- Expat should be available almost everywhere and
128 should be able to parse that?
129 Atom is harder because it may support html+.
130 I mean, yeah, it's stupid, but we could fill in header fields with
131 dummies and still use S-nail to look into the separated feeds as if
132 they were mail messages; anyway i would like to save me from using too
133 many tools -- three seems reasonable.
135 -- `sync'hronize commando -- robin@stjerndorff.org (Robin Stjerndorff):
136 Wondering how to update back to my Maildir, moving new read mails
137 in ~/Maildir from new to cur, without exiting the application.
138 Automation available? [And simply re-`[Ff]i' involves a lot of
141 -- Provide sync'ing options -- Jacob Gelbman <gelbman@gmail.com>:
142 If I open two instances of mailx, I then delete a message and then
143 quit in one. Then in the other one I read a message and quit, mailx
144 saves the status of the read message and the fact that a message was
145 deleted, even though it was opened before the other instance deleted
146 it. How is it doing that? [Of course he was using Maildir]
148 - Add TODO notes for those RFCs:
149 RFC 977 -> 3977 - Network News Transfer Protocol
150 RFC 1036 - Standard for USENET Messages
151 RFC 1524 - True support for mailcap files?
152 TODO YES! We really need to replace the pipe-TYPE/SUBTYPE mechanism
153 with something real. When we can work on the parsed MAIL DOM.
154 RFC 1939 - Post Office Protocol v3
155 RFC 2017 - URL External-Body Access-Type
156 RFC 2183 - The Content-Disposition Header
157 RFC 2369 - The Use of URLs as Meta-Syntax for Core Mail List Commands
158 and their Transport through Message Header Fields
159 (RFC 6068 - The 'mailto' URL scheme)
160 RFC 2384,1738 - I.e., Much better URL support
161 RFC 2387 - multipart/related -- yet handled like /alternative
162 RFC 2392 - Content-ID and Message-ID Uniform Resource Locators
163 RFC 2405 - The format of MIME message bodies.
164 RFC 2406 - Common multimedia types.
165 RFC 2407 - Encoding of non-ASCII text in message headers.
166 RFC 2449 - POP3 Extensions (including SASL)
167 RFC 2595 - TLS for POP3 (among others)
168 RFC 2980 - Common NNTP Extensions
169 RFC 3156 - MIME Security with OpenPGP
170 RFC 3207 - SMTP over TLS
172 Simple Mail Transfer Protocol (SMTP) Service Extension for Delivery
173 Status Notifications (DSNs),
174 An Extensible Message Format for Delivery Status Notifications
175 RFC 3676 - Updates to the text/plain MIME type and extensions for flowed
176 text (format=flowed). (Martin Neitzel)
177 RFC 4422, 4505 - Simple Authentication and Security layer (SASL)
179 RFC 4880 - OpenPGP Message Format
180 RFC 4954 - SMTP Authentication
181 rfc5198.txt Unicode Format for Network Interchange
182 RFC 5246 - Transport Layer Security (TLS)
183 RFC 5321 - Simple Mail Transfer Protocol.
184 RFC 5322 - The basic format of email messages.
185 RFC 5598 - Internet Mail Architecture
186 RFC 5751 - Secure/Multipurpose Internet Mail Extensions (S/MIME)
187 TODO NOTE that our S/MIME support is extremely weak regarding
188 TODO understanding, we should not rely on OpenSSL but instead
189 TODO handle it ourselfs; the RFC says:
190 S/MIME is used to secure MIME entities. A MIME entity can be a sub-
191 part, sub-parts of a message, or the whole message with all its sub-
192 parts. A MIME entity that is the whole message includes only the
193 MIME message headers and MIME body, and does not include the RFC-822
194 header. Note that S/MIME can also be used to secure MIME entities
195 used in applications other than Internet mail. If protection of the
196 RFC-822 header is required, the use of the message/rfc822 media type
197 is explained later in this section.
198 RFC 6125 - Representation and Verification of Domain-Based Application
199 Service Identity within Internet Public Key Infrastructure Using
200 X.509 (PKIX) Certificates in the Context of Transport Layer Security
202 RFC 6152 - SMTP Service Extension for 8-bit MIME Transport
203 RFC 6409 - Message Submission for Mail
204 rfc6530.txt Overview and Framework for Internationalized Email
205 rfc6531.txt SMTP Extension for Internationalized Email
206 rfc6532.txt Internationalized Email Headers
207 rfc6854.txt Update to Internet Message Format to Allow Group Syntax in
208 the "From:" and "Sender:" Header Fields
209 rfc6855.txt IMAP Support for UTF-8
210 rfc6856.txt Post Office Protocol Version 3 (POP3) Support for UTF-8
211 rfc6857.txt Post-Delivery Message Downgrading for Internationalized
213 rfc6858.txt Simplified POP and IMAP Downgrading for Internationalized Email
214 RFC 8058 Signaling One-Click Functionality for List Email Headers
216 draft-ietf-uta-email-tls-certs-01.txt
217 SMTP security via opportunistic DANE TLS draft-ietf-dane-smtp-with-dane-15
218 draft-melnikov-smime-header-signing
219 Considerations for protecting Email header with S/MIME
220 Read https://tools.ietf.org/html/draft-ietf-uta-tls-bcp-07.
221 Can we implement OCSP (see RFC 6066; -> RFC 6960)????
223 - This is how the codebase has to be reworked in respect to signals and
226 1. We introduce some environment/carrier structs: struct eval_ctx,
227 struct cmd_ctx, (struct send_ctx). All of these form lists.
228 eval_ctx gets a new instance every time evaluate() is entered; for
229 the interactive mode, commands() instantiates an outermost eval_ctx
230 that "cannot be left".
232 cmd_ctx knows about the eval_ctx in which it is was created; it is
233 created for each command that has an entry in cmd_tab and is passed
234 as the new argument of these kind of functions.
235 (send_ctx is the carrier for the MIME and send layer rewrite.)
236 2. cmd_tab handling becomes more intelligent: add bit fields so that
237 for some arguments (i.e., the first two or three) automatic
238 normalization can be specified, like whitespace trimming, like
239 automatic expansion of variables etc. For getrawlist.
240 It knows about the short synopsis and longer help of its program.
242 - and: commands that work only in interactive mode, and are
243 _silently_ ignored otherwise (no error).
244 I.e., commands which work on only with true user interaction
245 should not need to take care no more, but simply assume that
246 there will be true user interaction. And like that.
248 E.g., customhdr header field is always to be WS trimmed, just the
249 same as the body. shortcut at least should have trimmed name, etc.
250 To do that right, properly parse from left to right, take care
251 about backslash escaping, and (optionally) expand shell variables
252 and backtick substituations.
254 The cmd_ctx needs to carry around the cmd_tab structure so that the
255 commands can themselves print the synopsis in case of errors --
256 alternatively we need a return bit which should cause the command
257 callee to emit the synopsis as an error message, whatever is best.
258 Some (hopefully) duplicate occurrances of such strings can vanish.
259 2.1 We have some commands which offer "show" subcommands.
260 Those should only work in interactive mode OR SO.
261 This could be done with a flag, too.
262 X. Offer a central "`[un]onevent' EVENT MACRO [conditions]" register.
263 Change all hooks to use that one, optimize the case where a single
264 macro is registered for a single event but with different
267 E.g., "on_interactive_mode_enter" could then be hooked to call
268 `bind' and set `colour's, for example. In conjunction with 2.
269 above those commands could simply be (silent, successful) no-ops
270 before we reach that state (and again after
271 on_interactive_mode_leave is processed).
273 X.x We should make the message accessible at least a bit on macro level.
274 I.e., so that message specification checks can be applied on the
276 if `message match-spec '@~t@bank.com'`
277 set smime-sign customhdr='X-BlahBlahBlah:Banks are monsters'
279 where `message' (or so) would refer to the currently active
280 message (in compose mode) or the "dot" (otherwise).
281 `message' would gain some keywords. Something like that.
282 8. The line buffer used in evaluate() that is passed through to
283 commands (thus: in cmd_ctx, then) needs to become `const'.
284 (I tried to do so in the past, but some commands write into it,
285 thus i stopped and iirc even added some changes on my own which
286 take favour of reusing that buffer.)
287 + Macro execution then no longer needs to clone the macro content
288 lines before executing then.
289 + The command context also takes preparsed command array if so
290 specified in cmd_tab, and entries are cleaned up (see 2.)
291 9. We should unite all the un*() commands with their non-un*
292 versions, if they have one. They may take a different argument
293 list etc., but only one entry in the command table for such.
294 - The POSIX standard command abbreviations must remain, so maybe
295 outsource those in a hashtable or whatever that is checked first,
296 but detach command table order from that anyway.
297 - Offer a(n optional, and on/off switchable) Damerau-Levenshtein
298 mode for command completion;
299 10. We MUST switch the entire codebase to use SysV signal handling, don't
300 do the BSDish SA_RESTART, which is why we still suffer the way we
301 do and need jumps. I can't dig BSD signal handling, and never ever
302 did so myself until i got here.
303 11. Macro execution is potentially recursive. Meaning that
304 `undefine', etc. can occur while macros are executing.
305 The simplemost approach would be to have some recursion counter for
306 each macro and a delete_later flag that gets honoured when the
307 recursion counter gets zero. It would be already possible to
308 immediately remove the macro from the hashtable, so that deeper
309 levels wouldn't find it anymore. To avoid leaks (which *are*) we
310 need to have a jump location for our upcoming signal handler
311 anyway. (Also to get rid of the temporary_localopts_free() hack.
312 + The same is true for `account's. Here things are complicated by
313 the global `account_name', i.e., the account could be the current
315 + That also is: redefinition of names that have yet-pending
316 deletion requests are possible.
317 20. The attachment charset selection loop can then be rewritten to
318 check whether an ^C occurred and treat that as end-of-loop
319 condition. In v14.6.3 this was introduced, but it should act
320 differently depending on whether the interrupt occurred during
321 character set selection or attachment filename input.
322 Also in respect whether the interrupt is "propagated" or not.
323 It's ugly, and documented accordingly.
324 30. Mail protocols and mail messages are accessed through a "VFS".
325 URL should then support file:// and maildir:// etc. Update manual!
327 . It should be considered to drop many variables in favour of
328 ?SEARCH usage for keys, or key=value pairs, e.g.
329 smtp://exam.ple?starttls=yes;smtp-hostname=;
330 etc. USER@HOST is neat, but that is possibly preferable?
331 We could then also extend *mta* and drop *smtp* as such, e.g.,
332 mta=smtp://dSDAS, mta=builtin://XZ, mta=bltin://XZ.
333 Question: what is with smtp-hostname and such??
334 31. Flag updates of individual messages must find their way through to
336 32. Use deque (on partial views).
337 34. We need a new abstraction: `vie[ws]'. I.e, viewset, viewclear,
338 view(show|look)? We will have (possibly readonly) boxes, a summary
339 cache file, which is created when a mailbox is read in, and all
340 that crap that we currently have (setptr(), setmsize(), etc.!) must
341 vanish. Instead there is another, in-memory abstraction, the view.
342 Some views are built-in and are somehow selectable (the "all" view,
343 for example, and the "new" view).
344 It is possible to make a view persistent by giving it a name, e.g.,
345 'viewset NAME MSG-SPEC' -- 'viewset allnew :n' (and 'viewset XY `'
346 or something must be capable to tag the last a.k.a current).
347 Switching to a named view would thus look over the entire current
348 view (!) for all messages that comply to the message-spec of the
349 view, then create a sorted/threaded display of that subset and
350 create a new anonymous "result" view. It must be possible to
351 specify that a view is to be applied to the entire mailbox instead
352 of the current view, via a simple easy understandable syntax.
355 We won't extend macros that much because it would require much too
356 much logic for no purpose, instead we'll (hopefully) add some
357 scriptable abstraction, with an optional built-in Lua binding.
358 E.g., it would be too complicated to create a language to access
359 individual message parts (headers etc.), add error handling etc.
360 Better to make it all directly accessible via language binding, and
361 offer commands which relinquish execution to such a binding, maybe
362 even with some arguments, as in "? lua ARG" - it should be possible
363 that ARG is a msgset, then.
364 On the other hand, ~^ text-protocol stuff isn't that bad, and if
365 a text-protocol command mode could be added, that would be cool.
366 Even things like 'create temporary file', write data until EOT in
367 descriptor / temporary file / etc. and such!
369 But what macros should be able to do is iterating over such
370 a msgset / view, as in
371 msgset create NAME :SPEC:
372 msgset add NAME :SPEC:
375 normal-cmd SOMEHOW-REFER-TO-ITER
376 if SOMEHOW-ACCESS-LAST-STATUS
380 msgset cmd NAME normal-cmd
383 in which case NAME should simply be treated as a SPEC, likely that
384 this requires a new trigger, though, e.g. {NAME}, [NAME] or !NAME...
385 I.e., the new namelist (likely a deque) should contain MsgRef
386 objects which point to a Message and a Mailbox (for cross-mailbox
387 msgset's and without the need for a Message to store a pointer to
388 the owning mailbox? Or make MsgRef a superclass that may have
389 a subclass which offers such cross-refs).
391 50. Support SASL, unite all GSS-API etc. under an abstraction!
392 Maybe even drop direct GSS-API and support only through SASL.
393 That is, we can very well provide our own little SASL-client
394 abstraction with what we have already by simply defining some
395 "readline" abstraction plus struct ccred for use by the
396 authentication layer: the protocols must set it up by passing in
397 a line of authentication mechanisms and a callback mechanism.
398 Possibly the user should be able to permit or forbid automatic
399 selection of GSS-API (to avoid useless round-trips) etc. etc.
400 80. The MIME rewrite: mime_parser <-> mime "DOM" analyzer <->
401 selectively create filter chains per part and do XY.
403 This also affects sending, and it will allow us to dig MIME
404 (multipart) mail for -t/-m _correctly_. Also in sofar as we can
405 hook a content-decoder before diving into the MIME structure, and
406 with a DOM, we can re-encode such things properly as we (re)send
407 such mails. All this is wrong at the time of this writing!
408 We still need to special treat things like, e.g., RFC 2046, 5.2.1.
409 But on top of we-can, as opposed to the opposite.
410 99. Now i'm dreaming some more: with the new object-based approach
411 multiple mailboxes could be in an open state. And it should be
412 possible to do so for the user (`file' and `folder' are required to
413 quit the current mailbox [first -- this not yet]), which is why we
414 either need new trigger characters or new commands.
415 The absolute sensation would be joinable operations over multiple
416 open mailboxes, e.g., views over multiple such!
417 100. If i say `p 3 2 1' then i mean `3 2 1' not `1 2 3'.
419 - The thread sort doesn't get
428 The current sort fails to recognize that F and the thread starting at
429 B are related, which results in a mess.
431 -- Being able to sort the outermost level of threads was a suggestion
432 of Rudolf Sykora, especially being able to sort the outermost level
433 according to the date of the newest message in a thread.
435 - Drop **use-starttls* in favour of something better: support 'auto',
436 'no' and 'yes' and act accordingly. For the former be smart enough on
437 the protocol side. (RFC 3207 describes man-in-the-middle attacks due
438 to 'auto' TLS, so explicit 'yes' should be favoured).
440 - NOTE: we do not really support IPv6 sofar in that we are not prepared to
441 deal with IPv6 addresses (as in '[ADDR]:PORT'). Pimp url_parse().
444 - I had a connection collapse during a POP3 download, and neither was
445 there a chance to get access to the 22 yet downloaded mails (after
446 five minutes of waiting followed by CNTRL-C), nor did the layer
447 recognize this very well (got myriads of `POP3 connection already
448 closed.' messages, btw., the thirty-something messages which were not
449 yet downloaded caused (after CNTRL-C) this: ETC. ETC.
451 - I got an email in base64 that obviously used CRNL line endings, and once
452 i've replied the CR where quoted as *control* characters.
453 Get rid of those (kwcrtest.mbox; may be hard to do everywhere for some
454 time, due to how we deal with I/O and Send layer etc).
456 - edit.c doesn't do NEED_BODY (but IMAP won't work anyway).
459 .. s-nail </dev/null should work interactively when STDERR_FILENO is
460 a terminal! (Built-in editor; how do editline and readline work?
461 should this be documented? What does NetBSD Mail do? Should we NOT
462 be interactive?? POSIX says for sh(1) (APPLICATION USAGE): 'sh
463 2>FILE' is not interactive, even though it accepts terminal input.)
464 . It would be cool if ghosts, shortcuts, alternates could
465 (optionally?) be tracked via localopts. And macros. And inner macros.
466 (Additional entry on xy-local xy somewhere above)
467 And / or local to a macro/account if defined in one.
468 Should be per-carrier copy-on-write environments.
469 Just like TeX with def / gdef ...
470 . Just like the RFC 3676 link above, it would be nice if it would be
471 somehow possible to recognize links in a document; i don't know yet
472 how this could be achieved without loosing formatting information (i
473 mean, we could enable this and inject terminal colour sequences, but
474 one should be able to say 'follow link x', starting an action
475 handler, and the 'x' must come from somwhere - simply injecting
476 '[NUMBER]' references distorts visual). Anyway, it's just a filter
477 that recognized the usual <SCHEME:/> stuff, and of course we can
478 simply have a buffer which records all such occurrences, so that
479 user can say '? xy NUMBER', but without the context it soon gets
481 . TTY layer: the tc*() family may fail with EINTR, which MUST be
482 handled; setting also generates SIGTTOU when we're not in foreground
483 pgrp, so we better deal with all that and ENSURE WE GET THROUGH when
484 resetting terminal attributes!
485 .. TTY "I guess it would be much better to create our own session via
486 setpgid(2) and then tcsetpgrp(3) any processes we run synchronously,
487 and properly deal with SIGTTOU, but it always has been like that and
488 i won't do that before other things have been changed.
489 .. NOTE: TTY: place (at least n_child_run()) childs which go over
490 terminal into own group.
491 Introduce global "terminal state" manager which tracks who
492 currently owns the terminal, so that we can gracefully switch it
493 on/off, check in main loop whether restore is necessary.
494 It that can deal with multiple "windows" then we could have open
495 multiple of those, i.e., multiple PAGER instances, and non-PAGER
496 instances (fflush() if there is FILE before switch off).
498 TTY thus: "needsterminal" could be driven gracefully EVEN IF
499 a PAGER is open because the current user action is "print*", since
500 we KNOW that there is a PAGER and can temporarily lay it down to
501 sleep, fflush()ing/adjusting as necessary, and wake it up again
502 afterwards (or, with some work, gracefully shut it down if ^C ^C).
503 This is of course also true for user questions ("really display
504 part XY?", "need decryption password:", etc.!!!)!!
505 . Remove all occurrences of mbtowc() with mbrtowc(); temporarily add (some)
506 global mbstate_t objects until the send / MIME layer rewrite is done and
507 has the carrier. Use flip states and add aux funs with only update the
508 state+toggle on success -- CURRENTLY MBTOWC FAILURES ARE PRACTICALLY NOT
510 . pop3,mime_cte +++: \r,\n -> \015,\012, to avoid ANY problems..
511 Maybe our passed carrier should pass desired output newline (as
512 opposed to data-embedded newlines)
513 . which_protocol(), *newmail* mechanism, displayname, mailname: all of
514 this <rude>SHIT</rude> must vanish and be replaced by a URL, and
515 a nice "VFS" mailbox object that carries all necessary state so that
516 one can work with it.
518 If not mentioned somewhere else: struct message should be splitted
519 into a tree of objects, with a base class that has as few fields as
520 possible; the global *message should be a deque, only accessible via
521 iterator; it should store pointers to (the actually used subtype of)
522 message structures instead; i.e., for maildir boxes the path is yet
523 allocated separately, then it could be part of the message object,
525 It should contain a ui8_t that tracks the number of contained parts,
526 so that the "fits-onto-the-screen" tests are more useful than today;
527 i think 8-bit is sufficient, with 0xFF meaning more-than-fits-here.
528 . Given how many temporary files we use, it would make sense to
529 support a reusable single temporary file, as in singletmp_take() and
530 singletmp_release(), where singletmp_release() would close and thus
531 drop the file if it excesses a specific (configurable) size, and the
532 mainloop tick would close it (after X (configurable) unused ticks))
533 otherwise. I guess this would improve performance for searching
536 use GNU tools for extraction etc., and write a simple helper program
537 which converts these files to a serialized hashmap, just like we did
538 for the okeys (and *exactly* so); add a config check whether the ({})
539 extension is supported and finally use that for some ({static char
540 const *tr_res;}) injection optimization, then. (Think SFSYS)
541 . Searching body/text yet includes headers from attachments and
542 attachment data. This is shit. :)
543 . The "nifty" unregister_file()->_compress() mechanism that even
544 shovels '-Sfolder=imaps://user1@localhost -Srecord="+Sent Items"'
545 *records* calls clearerr() on the descriptor before performing it's
546 action anyway. when we really make it even to the I/O rewrite, it
547 should be possible to dis-/allow such -- it doesn't make sense to
548 add something faulty to whatever was not faulty before!
549 . `dp' prints EOF at the end of a thread even if unread messages
551 . `resend' doesn't smime-sign.
552 . Really do extend the test already today; test S/MIME
553 signing/encryption/decryption with two pairs of identities, instead
555 . RFC 5751 describes a message multipart layout that also includes the
556 headers in the signature; it would be nice (for completeness sake)
557 to be able to support that. Note shutup@ietf.org.
558 . The capability to save a message under the name of a recipient is in
559 the standard etc., but i've never used it.
560 What would be cool, otoh, would be if there would be the possibility
561 to register a regular expression, and if just *any* recipient of
562 a message matches, store the message in the given folder instead.
563 I.e., if i send a message to s-nail-users@ then i most likely want
564 to get a copy to the corresponding box, regardless of whoever the
565 message was sent To: Cc: or Bcc: else..
566 . why not simply sucking in complete MIME messages via -t? In a way
567 that parses it as a MIME message, that is!
568 (Brezn Stangl, brezn DOT stangl AT yandex DOT com)
569 . mutt list handling (`~') is very powerful
570 . We have some use of *at() functions, especially anything which
571 temporarily switches cwd.
572 . *newmail* is terrible. At some later time we need to do somethings
573 with timeouts etc. (for MBOX and Maildir it's not that bad, but for
574 anything over the network, yet the mentioned may come in over NFS).
575 Remove it until we have something better?
576 . The RFC 3798 *disposition-notification-send* mechanism is yet not
577 truly conforming (and works with *from*). Also, this is only the
578 sender side, there should be support for creating the MDN response.
579 (Maybe ternary option: off (default),
580 create-when-unread-flag-goes-away, ditto-but-also-strip-header)
581 .. Also, there is DSN as a SMTP extension, see the RFCs 3461, 346 (as
582 above) and 6522 (Wikipedia).
583 . The var_* series should return "const char*" not "char*".
584 This should already work today because otherwise we would get SEGV
586 .. While here: rename enum okeys to enum internal_variables, and the
587 ok_*() series to iv_(). And see below for env_*() series.
588 . fexpand() the 2nd: it should return structure because we need to
589 check for FEDIT_SYSBOX, which currently only checks whether the first
590 character of a file name is '%', not whether it is '%', '%:FILEPATH'
591 or '%VALIDUSER', because that is impossible to do!
592 . On the long run in-memory password storage should be zeroed after
593 use, possibly even encoded *during* use. After v15.
594 . We need a `spamcheck' command that is like `spamrate' but updates
595 the mail in-place, i.e., with the headers that the spam engine adds.
596 . __narrow_suffix() is wrong (for stateful encodings that we
597 don't support yet) and should inject a reset sequence if it shortens
599 .. THAT IS TO SAY: the entire codebase doesn't really support stateful
600 encodings, including the bidi_ things that i've done (but the MLE
601 does iirc? what is this??). We should have a global string that
602 has the multibyte reset sequence plus length available for easy
604 . When a user edits a specific header, it should no longer be
605 modified. (Do not loose knowledge that collect() edited it.)
606 . Regular expression list resorting is no good; the user should be
607 able to specify a match order weight, as in:
608 mlist 10 a@b.org 8 c@d.org .*@else@org 0 almost@never.com
609 So: optional digit 0-10, where 0-4 are never relinked and always
610 placed at the tail, 6-10 are never relinked and always placed at
611 head (all in decreasing order, head to tail), and 5 is the implicit
612 value, placed in between and automatically resorted just as is the
613 sole algorithm we currently have.
614 .. And maybe we should have an event mechanism with one-shot etc..
615 Then install a resorter function when we actually have lookups and
616 one-shot sort the entire thing once (when the loop ticks).
617 Instead of busy resorting, that is.
618 . We need more hooks: on-leave, on-connect.. whatever
619 . The new internal ~/$ expansion mechanism should possibly get support
620 for POSIX parameter expansions ${[:]-} and ${[:]+} (and ${[:]?}).
621 There is no real way to get the functionality otherwise...
622 . Make S/MIME an option separate of SSL/TLS, i.e., optional.
623 . With very long input Heirloom mailx(1) / S-nail(1) can produce
624 encoded-words (RFC 2047) with incomplete multibyte sequences (i.e.,
625 non self-contained encoded-words).
626 . Group addresses, especially the undisclosed recipients but also
627 "Bla": addresses; are missing.
628 . Cleanup: n_ is anything public / external,
629 _n_ is public/e but not really, FILENAME/ABBREV_symbol are statics,
630 FILENAME/ABBREV__symbol are helpers of a single other static
631 (function group) (-> colour.c, lex_input.c: done).
632 . Per-folder (S/MIME) en- and decryption key (Tarqi Kazan): if a xy
633 variable is set (that points to a key) add a transparent en- and
634 decryption layer on top of any per-message operation (for boxes for
635 which the variable is set).
636 . For v15.0: remember private thread with Tarqi Kazan (2015-05) and
637 try to improve situation with *record*, so that only messages enter
638 it which have really been sent. If we support postponing and have
639 a multi-process layout and add an intermediate *record-queue* we
640 may be able to improve the situation.
641 . [Dd]ecrypt should transport decryption errors, not silently be like
642 copy and copy undecrypted content, because this is what it's for?
643 ..We need atomic operations with rollback support in order to make
644 this happen, but i think maybe file truncation (decryption always
645 appends?) is enough provided that files are locked?
646 WE NEED ATOMIC OPERATION SUPPORT for quite some operations.
647 Man, are we far from that.
648 . `pipe' is total shit regarding MIME. We need some defined and
649 documented method to configure which parts are displayed and/or how
650 they are visually separated.
651 . The new n_err() facility is nice, but we need n_msg() too, maybe
652 more. Maybe à la Log:: and macro wrappers with priority unrolled,
653 then add user-settable treshold, too. Btw. C99 adds vararg macros.
654 .. We must handle embedded newlines properly
655 .. In PARTICULAR we MUST NOT use stdout when in batch mode:
656 ssh X "echo 'move * |cat' | s-nail -#f %" >> download
657 should do the right thing, but can't like that due to unwanted
658 noise in the stdout output! Best would be if it would be
659 possible to explicitly define a file/dev to be used, but falling
660 back to stderr in batch mode otherwise. (think S-Web42)
661 . Exit status handling is sick.
662 . Especially in interactive mode MIME classification should count
663 (non-NUL) control characters and the number of different thereof,
664 giving users an option to treat something as text nonetheless.
665 E.g., roff files often use controls as separators in conditionals,
666 some shell scripts use controls for $IFS field separation, and that
667 will end up with charset=binary.
668 .. *mime-allow-text-controls* is a no-brainer: instead we should
669 introduce something that allows us to switch and detect UTF-16 once
670 we run into the problematic situation, then start all over in an
671 Unicode mode? I.e.: continue to force the user to set such
672 a switch, but do it in a sensible fashion, because the UTF-16 data
673 stream may nonetheless contain control characters??
676 . We can "steal" features from msmtp(1) that make sense: SOCKS support
677 (primitive) and /etc/aliases ($mta_alias_file). At least postfix(1)
678 supports file and pipe addressees in the latter... It also
679 supports include files via :include:/filename but which i think
680 should be supported in a second step. Ditto caching (timestamp
681 check and a mechanism to support/disable caching.)
685 . smime_verify(): only dump the multipart that is signed into the file for
686 verification purposes. DOCUMENT that only the FIRST such part is verified.
687 Ditto, we don't decrypt but on toplevel. Sic.
689 . convert iconv so that it always "places the reset sequence" i.e.
690 finalizes the string properly. we don't do this at all right now!
692 . wysh: support a -- terminator, also in getmsglist();
693 then: `pipe' etc.: add support for wysh prefix, which uses the new
694 syntax: like this we can obsolete the ridiculous lastring() hack of nail
695 in v15! (Stanley Lieber, Gavin Troy)
697 . -:, *mimetypes-load-control*, ?, should honour the given load order; as
698 appropriate, add a "b" for built-in!
699 It happened to me that i searched for at least 30 minutes for a bug
700 that resulted in text/plain not text/x-diff only to find out that this
701 was because of ArchLinux's /etc/mime.types!
703 . getapproval() should support a TRUM1 return, meaning "cancel",
704 to be understood as appropriate.
706 . `mbox' _can_ be made usable anywhere with yet another PS_MBOX global
707 bypass! ditto touch,save,Save
709 . with *mime-alternative-favour-rich* it is possibly time to support GMAIL
710 style quotes, as well as class="quote" and such.
712 . can we introduce a BUILD_DIR thing? or OBJ_DIR OBJDIR (what was the BSD
713 name for this again?) I really would like to be able to have separate
714 build directories! (PREFIXDEV, however, won't work out for a while.)
716 . We should be much smarter regarding when we allow a PAGER
717 etc. to be used, which is supposed to be a possibly useful thing in
718 $ s-nail -Scrt=0 >LOG 2>&1
720 . when doing Lreply we may ask for Reply-To:, but strip out the address
721 actively even if user said yes to the question. That should not
722 happen? It somehow matches the documentation however. unsure.
724 . if -t is used and the file includes Mail-Followup-To:, then we should
725 NOT add to it, OR we need to offer a way to get there!
727 . stale dotlock files should be removed - postfix has
728 a stale_lock_time=500s variable to configure that