1 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
2 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
3 <html xmlns="http://www.w3.org/1999/xhtml">
5 <title>Design notes on fetchmail</title>
6 <link rev="made" href="mailto:esr@snark.thyrsus.com" />
7 <meta name="description" content="Design notes on fetchmail." />
8 <meta name="keywords" content="fetchmail, POP, POP2, POP3, IMAP, remote mail" />
9 <style type="text/css">
11 h1.c1 {text-align: center}
16 <table width="100%" cellpadding="0" summary="Canned page header">
18 <td width="30%">Back to <a href="/~esr/index.html">Fetchmail Home Page</a></td>
19 <td width="30%" align="center">To <a href="/~esr/sitemap.html">Site Map</a></td>
20 <td width="30%" align="right">$Date: 2002/07/28 09:23:04 $</td>
25 <h1 class="c1">Design Notes On Fetchmail</h1>
27 <p>These notes are for the benefit of future hackers and
28 maintainers. The following sections are both functional and
29 narrative, read from beginning to end.</p>
33 <p>A direct ancestor of the fetchmail program was originally
34 authored (under the name popclient) by Carl Harris
35 <ceharris@mal.com>. I took over development in June 1996 and
36 subsequently renamed the program `fetchmail' to reflect the
37 addition of IMAP support and SMTP delivery. In early November 1996
38 Carl officially ended support for the last popclient versions.</p>
40 <p>Before accepting responsibility for the popclient sources from
41 Carl, I had investigated and used and tinkered with every other
42 UNIX remote-mail forwarder I could find, including fetchpop1.9,
43 PopTart-0.9.3, get-mail, gwpop, pimp-1.0, pop-perl5-1.2, popc,
44 popmail-1.6 and upop. My major goal was to get a header-rewrite
45 feature like fetchmail's working so I wouldn't have reply problems
48 <p>Despite having done a good bit of work on fetchpop1.9, when I
49 found popclient I quickly concluded that it offered the solidest
50 base for future development. I was convinced of this primarily by
51 the presence of multiple-protocol support. The competition didn't
52 do POP2/RPOP/APOP, and I was already having vague thoughts of maybe
53 adding IMAP. (This would advance two other goals: learn IMAP and
54 get comfortable writing TCP/IP client software.)</p>
56 <p>Until popclient 3.05 I was simply following out the implications
57 of Carl's basic design. He already had daemon.c in the
58 distribution, and I wanted daemon mode almost as badly as I wanted
59 the header rewrite feature. The other things I added were bug fixes
60 or minor extensions.</p>
62 <p>After 3.1, when I put in SMTP-forwarding support (more about
63 this below) the nature of the project changed -- it became a
64 carefully-thought-out attempt to render obsolete every other
65 program in its class. The name change quickly followed.</p>
67 <h1>The rewrite option</h1>
69 <p>MTAs ought to canonicalize the addresses of outgoing non-local
70 mail so that From:, To:, Cc:, Bcc: and other address headers
71 contain only fully qualified domain names. Failure to do so can
72 break the reply function on many mailers. (Sendmail has an option
75 <p>This problem only becomes obvious when a reply is generated on a
76 machine different from where the message was delivered. The two
77 machines will have different local username spaces, potentially
78 leading to misrouted mail.</p>
80 <p>Most MTAs (and sendmail in particular) do not canonicalize
81 address headers in this way (violating RFC 1123). Fetchmail
82 therefore has to do it. This is the first feature I added to the
83 ancestral popclient.</p>
85 <h1>Reorganization</h1>
87 <p>The second thing I did reorganize and simplify popclient a lot.
88 Carl Harris's implementation was very sound, but exhibited a kind
89 of unnecessary complexity common to many C programmers. He treated
90 the code as central and the data structures as support for the
91 code. As a result, the code was beautiful but the data structure
92 design ad-hoc and rather ugly (at least to this old LISP
95 <p>I was able to improve matters significantly by reorganizing most
96 of the program around the `query' data structure and eliminating a
97 bunch of global context. This especially simplified the main
98 sequence in fetchmail.c and was critical in enabling the daemon
101 <h1>IMAP support and the method table</h1>
103 <p>The next step was IMAP support. I initially wrote the IMAP code
104 as a generic query driver and a method table. The idea was to have
105 all the protocol-independent setup logic and flow of control in the
106 driver, and the protocol-specific stuff in the method table.</p>
108 <p>Once this worked, I rewrote the POP3 code to use the same
109 organization. The POP2 code kept its own driver for a couple more
110 releases, until I found sources of a POP2 server to test against
111 (the breed seems to be nearly extinct).</p>
113 <p>The purpose of this reorganization, of course, is to trivialize
114 the development of support for future protocols as much as
115 possible. All mail-retrieval protocols have to have pretty similar
116 logical design by the nature of the task. By abstracting out that
117 common logic and its interface to the rest of the program, both the
118 common and protocol-specific parts become easier to understand.</p>
120 <p>Furthermore, many kinds of new features can instantly be
121 supported across all protocols by modifying the one driver
124 <h1>Implications of smtp forwarding</h1>
126 <p>The direction of the project changed radically when Harry
127 Hochheiser sent me his scratch code for forwarding fetched mail to
128 the SMTP port. I realized almost immediately that a reliable
129 implementation of this feature would make all the other delivery
132 <p>Why mess with all the complexity of configuring an MDA or
133 setting up lock-and-append on a mailbox when port 25 is guaranteed
134 to be there on any platform with TCP/IP support in the first place?
135 Especially when this means retrieved mail is guaranteed to look
136 like normal sender- initiated SMTP mail, which is really what we
139 <p>Clearly, the right thing to do was (1) hack SMTP forwarding
140 support into the generic driver, (2) make it the default mode, and
141 (3) eventually throw out all the other delivery modes.</p>
143 <p>I hesitated over step 3 for some time, fearing to upset
144 long-time popclient users dependent on the alternate delivery
145 mechanisms. In theory, they could immediately switch to .forward
146 files or their non-sendmail equivalents to get the same effects. In
147 practice the transition might have been messy.</p>
149 <p>But when I did it (see the NEWS note on the great options
150 massacre) the benefits proved huge. The cruftiest parts of the
151 driver code vanished. Configuration got radically simpler -- no
152 more grovelling around for the system MDA and user's mailbox, no
153 more worries about whether the underlying OS supports file
156 <p>Also, the only way to lose mail vanished. If you specified
157 localfolder and the disk got full, your mail got lost. This can't
158 happen with SMTP forwarding because your SMTP listener won't return
159 OK unless the message can be spooled or processed.</p>
161 <p>Also, performance improved (though not so you'd notice it in a
162 single run). Another not insignificant benefit of this change was
163 that the manual page got a lot simpler.</p>
165 <p>Later, I had to bring --mda back in order to allow handling of
166 some obscure situations involving dynamic SLIP. But I found a much
167 simpler way to do it.</p>
169 <p>The moral? Don't hesitate to throw away superannuated features
170 when you can do it without loss of effectiveness. I tanked a couple
171 I'd added myself and have no regrets at all. As Saint-Exupery said,
172 "Perfection [in design] is achieved not when there is nothing more
173 to add, but rather when there is nothing more to take away." This
174 program isn't perfect, but it's trying.</p>
176 <h1>The most-requested features that I will never add, and why
179 <h2>Password encryption in .fetchmailrc</h2>
181 <p>The reason there's no facility to store passwords encrypted in
182 the .fetchmailrc file is because this doesn't actually add
185 <p>Anyone who's acquired the 0600 permissions needed to read your
186 .fetchmailrc file will be able to run fetchmail as you anyway --
187 and if it's your password they're after, they'd be able to rip the
188 necessary decoder out of the fetchmail code itself to get it.</p>
190 <p>All .fetchmailrc encryption would do is give a false sense of
191 security to people who don't think very hard.</p>
193 <h2>Truly concurrent queries to multiple hosts</h2>
195 <p>Occasionally I get a request for this on "efficiency" grounds.
196 These people aren't thinking either. True concurrency would do
197 nothing to lessen fetchmail's total IP volume. The best it could
198 possibly do is change the usage profile to shorten the duration of
199 the active part of a poll cycle at the cost of increasing its
200 demand on IP volume per unit time.</p>
202 <p>If one could thread the protocol code so that fetchmail didn't
203 block on waiting for a protocol response, but rather switched to
204 trying to process another host query, one might get an efficiency
205 gain (close to constant loading at the single-host level).</p>
207 <p>Fortunately, I've only seldom seen a server that incurred
208 significant wait time on an individual response. I judge the gain
209 from this not worth the hideous complexity increase it would
210 require in the code.</p>
212 <h2>Multiple concurrent instances of fetchmail</h2>
214 <p>Fetchmail locking is on a per-invoking-user because
215 finer-grained locks would be really hard to implement in a portable
216 way. The problem is that you don't want two fetchmails querying the
217 same site for the same remote user at the same time.</p>
219 <p>To handle this optimally, multiple fetchmails would have to
220 associate a system-wide semaphore with each active pair of a remote
221 user and host canonical address. A fetchmail would have to block
222 until getting this semaphore at the start of a query, and release
223 it at the end of a query.</p>
225 <p>This would be way too complicated to do just for an "it might be
226 nice" feature. Instead, you can run a single root fetchmail polling
227 for multiple users in either single-drop or multidrop mode.</p>
229 <p>The fundamental problem here is how an instance of fetchmail
230 polling host foo can assert that it's doing so in a way visible to
231 all other fetchmails. System V semaphores would be ideal for this
232 purpose, but they're not portable.</p>
234 <p>I've thought about this a lot and roughed up several designs.
235 All are complicated and fragile, with a bunch of the standard
236 problems (what happens if a fetchmail aborts before clearing its
237 semaphore, and how do we recover reliably?).</p>
239 <p>I'm just not satisfied that there's enough functional gain here
240 to pay for the large increase in complexity that adding these
241 semaphores would entail.</p>
243 <h1>Multidrop and alias handling</h1>
245 <p>I decided to add the multidrop support partly because some users
246 were clamoring for it, but mostly because I thought it would shake
247 bugs out of the single-drop code by forcing me to deal with
248 addressing in full generality. And so it proved.</p>
250 <p>There are two important aspects of the features for handling
251 multiple-drop aliases and mailing lists which future hackers should
252 be careful to preserve.</p>
256 <p>The logic path for single-recipient mailboxes doesn't involve
257 header parsing or DNS lookups at all. This is important -- it means
258 the code for the most common case can be much simpler and more
263 <p>The multidrop handing does <em>not</em> rely on doing the
264 equivalent of passing the message to sendmail -oem -t. Instead, it
265 explicitly mines members of a specified set of local usernames out
270 <p>We do <em>not</em> attempt delivery to multidrop mailboxes in
271 the presence of DNS errors. Before each multidrop poll we probe DNS
272 to see if we have a nameserver handy. If not, the poll is skipped.
273 If DNS crashes during a poll, the error return from the next
274 nameserver lookup aborts message delivery and ends the poll. The
275 daemon mode will then quietly spin until DNS comes up again, at
276 which point it will resume delivering mail.</p>
280 <p>When I designed this support, I was terrified of doing anything
281 that could conceivably cause a mail loop (you should be too).
282 That's why the code as written can only append <em>local</em> names
283 (never @-addresses) to the recipients list.</p>
285 <p>The code in mxget.c is nasty, no two ways about it. But it's
286 utterly necessary, there are a lot of MX pointers out there. It
287 really ought to be a (documented!) entry point in the bind
290 <h1>DNS error handling</h1>
292 <p>Fetchmail's behavior on DNS errors is to suppress forwarding and
293 deletion of the individual message that each occurs in, leaving it
294 queued on the server for retrieval on a subsequent poll. The
295 assumption is that DNS errors are transient, due to temporary
298 <p>Unfortunately this means that if a DNS error is permanent a
299 message can be perpetually stuck in the server mailbox. We've had a
300 couple bug reports of this kind due to subtle RFC822 parsing errors
301 in the fetchmail code that resulted in impossible things getting
302 passed to the DNS lookup routines.</p>
304 <p>Alternative ways to handle the problem: ignore DNS errors
305 (treating them as a non-match on the mailserver domain), or forward
306 messages with errors to fetchmail's invoking user in addition to
307 any other recipients. These would fit an assumption that DNS lookup
308 errors are likely to be permanent problems associated with an
311 <h1>IPv6 and IPSEC</h1>
313 <p>The IPv6 support patches are really more protocol-family
314 independence patches. Because of this, in most places, "ports"
315 (numbers) have been replaced with "services" (strings, that may be
316 digits). This allows us to run with certain protocols that use
317 strings as "service names" where we in the IP world think of port
318 numbers. Someday we'll plumb strings all over and then, if inet6 is
319 not enabled, do a getservbyname() down in SocketOpen. The IPv6
320 support patches use getaddrinfo(), which is a POSIX p1003.1g
321 mandated function. So, in the not too distant future, we'll zap the
322 ifdefs and just let autoconf check for getaddrinfo. IPv6 support
323 comes pretty much automatically once you have protocol family
326 <h1>Internationalization</h1>
328 <p>Internationalization is handled using GNU gettext (see the file
329 ABOUT_NLS in the source distribution). This places some minor
330 constraints on the code.</p>
332 <p>Strings that must be subject to translation should be wrapped
333 with GT_() or N_() -- the former in function arguments, the latter
334 in static initializers and other non-function-argument
337 <h1>Checklist for Adding Options</h1>
339 <p>Adding a control option is not complicated in principle, but
340 there are a lot of fiddly details in the process. You'll need to do
341 the following minimum steps.</p>
344 <li>Add a field to represent the control in <code>struct
345 run</code>, <code>struct query</code>, or <code>struct
346 hostdata</code>.</li>
348 <li>Go to <code>rcfile_y.y</code>. Add the token to the grammar.
349 Don't forget the <code>%token</code> declaration.</li>
351 <li>Pick an actual string to declare the option in the .fetchmailrc
352 file. Add the token to <code>rcfile_l</code>.</li>
354 <li>Pick a long-form option name, and a one-letter short option if
355 any are left. Go to <code>options.c</code>. Pick a new
356 <code>LA_</code> value. Hack the <code>longoptions</code> table to
357 set up the association. Hack the big switch statement to set the
358 option. Hack the `?' message to describe it.</li>
360 <li>If the default is nonzero, set it in <code>def_opts</code> near
361 the top of <code>load_params</code> in
362 <code>fetchmail.c</code>.</li>
364 <li>Add code to dump the option value in
365 <code>fetchmail.c:dump_params</code>.</li>
367 <li>For a per-site or per-user option, add proper
368 <code>FLAG_MERGE</code> actions in fetchmail.c's optmerge()
369 function. For a global option, add an override at the end of
370 load_params; this will involve copying a "cmd_run." field to a
371 corresponding "run." field, see the existing code for models.</li>
373 <li>Document the option in fetchmail.man. This will require at
374 least two changes; one to the collected table of options, and one
375 full text description of the option.</li>
377 <li>Hack fetchmailconf to configure it. Bump the fetchmailconf
380 <li>Hack conf.c to dump the option so we won't have a version-skew
383 <li>Add an entry to NEWS.</li>
385 <li>If the option implements a new feature, add a note to the
389 <p>There may be other things you have to do in the way of logic, of
392 <p>Before you implement an option, though, think hard. Is there any
393 way to make fetchmail automatically detect the circumstances under
394 which it should change its behavior? If so, don't write an option.
395 Just do the check!</p>
397 <h1>Lessons learned</h1>
399 <h3>1. Server-side state is essential</h3>
401 <p>The person(s) responsible for removing LAST from POP3 deserve to
402 suffer. Without it, a client has no way to know which messages in a
403 box have been read by other means, such as an MUA running on the
406 <p>The POP3 UID feature described in RFC1725 to replace LAST is
407 insufficient. The only problem it solves is tracking which messages
408 have been read <em>by this client</em> -- and even that requires
409 tricky, fragile implementation.</p>
411 <p>The underlying lesson is that maintaining accessible server-side
412 `seen' state bits associated with Status headers is indispensible
413 in a Unix/RFC822 mail server protocol. IMAP gets this right.</p>
415 <h3>2. Readable text protocol transactions are a Good Thing</h3>
417 <p>A nice thing about the general class of text-based protocols
418 that SMTP, POP2, POP3, and IMAP belongs to is that client/server
419 transactions are easy to watch and transaction code correspondingly
420 easy to debug. Given a decent layer of socket utility functions
421 (which Carl provided) it's easy to write protocol engines and not
422 hard to show that they're working correctly.</p>
424 <p>This is an advantage not to be despised! Because of it, this
425 project has been interesting and fun -- no serious or persistent
426 bugs, no long hours spent looking for subtle pathologies.</p>
428 <h3>3. IMAP is a Good Thing.</h3>
430 <p>Now that there is a standard IMAP equivalent of the POP3 APOP
431 validation in CRAM-MD5, POP3 is completely obsolete.</p>
433 <h3>4. SMTP is the Right Thing</h3>
435 <p>In retrospect it seems clear that this program (and others like
436 it) should have been designed to forward via SMTP from the
437 beginning. This lesson may be applicable to other Unix programs
438 that now call the local MDA/MTA as a program.</p>
440 <h3>5. Syntactic noise can be your friend</h3>
442 <p>The optional `noise' keywords in the rc file syntax started out
443 as a late-night experiment. The English-like syntax they allow is
444 considerably more readable than the traditional terse keyword-value
445 pairs you get when you strip them all out. I think there may be a
446 wider lesson here.</p>
448 <h1>Motivation and validation</h1>
450 <p>It is truly written: the best hacks start out as personal
451 solutions to the author's everyday problems, and spread because the
452 problem turns out to be typical for a large class of users. So it
453 was with Carl Harris and the ancestral popclient, and so with me
456 <p>It's gratifying that fetchmail has become so popular. Until just
457 before 1.9 I was designing strictly to my own taste. The multi-drop
458 mailbox support and the new --limit option were the first features
459 to go in that I didn't need myself.</p>
461 <p>By 1.9, four months after I started hacking on popclient and a
462 month after the first fetchmail release, there were literally a
463 hundred people on the fetchmail-friends contact list. That's pretty
464 powerful motivation. And they were a good crowd, too, sending fixes
465 and intelligent bug reports in volume. A user population like that
466 is a gift from the gods, and this is my expression of
469 <p>The beta testers didn't know it at the time, but they were also
470 the subjects of a sociological experiment. The results are
471 described in my paper, <a
472 href="//www.tuxedo.org/~esr/writings/cathedral-bazaar/">The
473 Cathedral And The Bazaar</a>.</p>
477 <p>Special thanks go to Carl Harris, who built a good solid code
478 base and then tolerated me hacking it out of recognition. And to
479 Harry Hochheiser, who gave me the idea of the SMTP-forwarding
482 <p>Other significant contributors to the code have included Dave
483 Bodenstab (error.c code and --syslog), George Sipe (--monitor and
484 --interface), Gordon Matzigkeit (netrc.c), Al Longyear (UIDL
485 support), Chris Hanson (Kerberos V4 support), and Craig Metz (OPIE,
490 <p>At this point, the fetchmail code appears to be pretty stable.
491 It will probably undergo substantial change only if and when
492 support for a new retrieval protocol or authentication method is
495 <h1>Relevant RFCS</h1>
497 <p>Not all of these describe standards explicitly used in
498 fetchmail, but they all shaped the design in one way or
502 <dt><a href="ftp://ftp.isi.edu/in-notes/rfc821.txt">RFC821</a></dt>
504 <dd>SMTP protocol</dd>
506 <dt><a href="ftp://ftp.isi.edu/in-notes/rfc822.txt">RFC822</a></dt>
508 <dd>Mail header format</dd>
510 <dt><a href="ftp://ftp.isi.edu/in-notes/rfc937.txt">RFC937</a></dt>
512 <dd>Post Office Protocol - Version 2</dd>
514 <dt><a href="ftp://ftp.isi.edu/in-notes/rfc974.txt">RFC974</a></dt>
518 <dt><a href="ftp://ftp.isi.edu/in-notes/rfc976.txt">RFC976</a></dt>
520 <dd>UUCP mail format</dd>
523 href="ftp://ftp.isi.edu/in-notes/rfc1081.txt">RFC1081</a></dt>
525 <dd>Post Office Protocol - Version 3</dd>
528 href="ftp://ftp.isi.edu/in-notes/rfc1123.txt">RFC1123</a></dt>
530 <dd>Host requirements (modifies 821, 822, and 974)</dd>
533 href="ftp://ftp.isi.edu/in-notes/rfc1176.txt">RFC1176</a></dt>
535 <dd>Interactive Mail Access Protocol - Version 2</dd>
538 href="ftp://ftp.isi.edu/in-notes/rfc1203.txt">RFC1203</a></dt>
540 <dd>Interactive Mail Access Protocol - Version 3</dd>
543 href="ftp://ftp.isi.edu/in-notes/rfc1225.txt">RFC1225</a></dt>
545 <dd>Post Office Protocol - Version 3</dd>
548 href="ftp://ftp.isi.edu/in-notes/rfc1344.txt">RFC1344</a></dt>
550 <dd>Implications of MIME for Internet Mail Gateways</dd>
553 href="ftp://ftp.isi.edu/in-notes/rfc1413.txt">RFC1413</a></dt>
555 <dd>Identification server</dd>
558 href="ftp://ftp.isi.edu/in-notes/rfc1428.txt">RFC1428</a></dt>
560 <dd>Transition of Internet Mail from Just-Send-8 to 8-bit
564 href="ftp://ftp.isi.edu/in-notes/rfc1460.txt">RFC1460</a></dt>
566 <dd>Post Office Protocol - Version 3</dd>
569 href="ftp://ftp.isi.edu/in-notes/rfc1508.txt">RFC1508</a></dt>
571 <dd>Generic Security Service Application Program Interface</dd>
574 href="ftp://ftp.isi.edu/in-notes/rfc1521.txt">RFC1521</a></dt>
576 <dd>MIME: Multipurpose Internet Mail Extensions</dd>
579 href="ftp://ftp.isi.edu/in-notes/rfc1869.txt">RFC1869</a></dt>
581 <dd>SMTP Service Extensions (ESMTP spec)</dd>
584 href="ftp://ftp.isi.edu/in-notes/rfc1652.txt">RFC1652</a></dt>
586 <dd>SMTP Service Extension for 8bit-MIMEtransport</dd>
589 href="ftp://ftp.isi.edu/in-notes/rfc1725.txt">RFC1725</a></dt>
591 <dd>Post Office Protocol - Version 3</dd>
594 href="ftp://ftp.isi.edu/in-notes/rfc1730.txt">RFC1730</a></dt>
596 <dd>Interactive Mail Access Protocol - Version 4</dd>
599 href="ftp://ftp.isi.edu/in-notes/rfc1731.txt">RFC1731</a></dt>
601 <dd>IMAP4 Authentication Mechanisms</dd>
604 href="ftp://ftp.isi.edu/in-notes/rfc1732.txt">RFC1732</a></dt>
606 <dd>IMAP4 Compatibility With IMAP2 And IMAP2bis</dd>
609 href="ftp://ftp.isi.edu/in-notes/rfc1734.txt">RFC1734</a></dt>
611 <dd>POP3 AUTHentication command</dd>
614 href="ftp://ftp.isi.edu/in-notes/rfc1870.txt">RFC1870</a></dt>
616 <dd>SMTP Service Extension for Message Size Declaration</dd>
619 href="ftp://ftp.isi.edu/in-notes/rfc1891.txt">RFC1891</a></dt>
621 <dd>SMTP Service Extension for Delivery Status Notifications</dd>
624 href="ftp://ftp.isi.edu/in-notes/rfc1892.txt">RFC1892</a></dt>
626 <dd>The Multipart/Report Content Type for the Reporting of Mail
627 System Administrative Messages</dd>
630 href="ftp://ftp.isi.edu/in-notes/rfc1894.txt">RFC1894</a></dt>
632 <dd>An Extensible Message Format for Delivery Status
636 href="ftp://ftp.isi.edu/in-notes/rfc1893.txt">RFC1893</a></dt>
638 <dd>Enhanced Mail System Status Codes</dd>
641 href="ftp://ftp.isi.edu/in-notes/rfc1894.txt">RFC1894</a></dt>
643 <dd>An Extensible Message Format for Delivery Status
647 href="ftp://ftp.isi.edu/in-notes/rfc1938.txt">RFC1938</a></dt>
649 <dd>A One-Time Password System</dd>
652 href="ftp://ftp.isi.edu/in-notes/rfc1939.txt">RFC1939</a></dt>
654 <dd>Post Office Protocol - Version 3</dd>
657 href="ftp://ftp.isi.edu/in-notes/rfc1957.txt">RFC1957</a></dt>
659 <dd>Some Observations on Implementations of the Post Office
663 href="ftp://ftp.isi.edu/in-notes/rfc1985.txt">RFC1985</a></dt>
665 <dd>SMTP Service Extension for Remote Message Queue Starting</dd>
668 href="ftp://ftp.isi.edu/in-notes/rfc2033.txt">RFC2033</a></dt>
670 <dd>Local Mail Transfer Protocol</dd>
673 href="ftp://ftp.isi.edu/in-notes/rfc2060.txt">RFC2060</a></dt>
675 <dd>Internet Message Access Protocol - Version 4rev1</dd>
678 href="ftp://ftp.isi.edu/in-notes/rfc2061.txt">RFC2061</a></dt>
680 <dd>IMAP4 Compatibility With IMAP2bis</dd>
683 href="ftp://ftp.isi.edu/in-notes/rfc2062.txt">RFC2062</a></dt>
685 <dd>Internet Message Access Protocol - Obsolete Syntax</dd>
688 href="ftp://ftp.isi.edu/in-notes/rfc2195.txt">RFC2195</a></dt>
690 <dd>IMAP/POP AUTHorize Extension for Simple Challenge/Response</dd>
693 href="ftp://ftp.isi.edu/in-notes/rfc2177.txt">RFC2177</a></dt>
695 <dd>IMAP IDLE command</dd>
698 href="ftp://ftp.isi.edu/in-notes/rfc2449.txt">RFC2449</a></dt>
700 <dd>POP3 Extension Mechanism</dd>
703 href="ftp://ftp.isi.edu/in-notes/rfc2554.txt">RFC2554</a></dt>
705 <dd>SMTP Service Extension for Authentication</dd>
708 href="ftp://ftp.isi.edu/in-notes/rfc2595.txt">RFC2595</a></dt>
710 <dd>Using TLS with IMAP, POP3 and ACAP</dd>
713 href="ftp://ftp.isi.edu/in-notes/rfc2645.txt">RFC2645</a></dt>
715 <dd>On-Demand Mail Relay: SMTP with Dynamic IP Addresses</dd>
718 href="ftp://ftp.isi.edu/in-notes/rfc2683.txt">RFC2683</a></dt>
720 <dd>IMAP4 Implementation Recommendations</dd>
723 href="ftp://ftp.isi.edu/in-notes/rfc2821.txt">RFC2821</a></dt>
725 <dd>Simple Mail Transfer Protocol</dd>
728 href="ftp://ftp.isi.edu/in-notes/rfc2822.txt">RFC2822</a></dt>
730 <dd>Internet Message Format</dd>
734 RFC2192 IMAP URL Scheme
735 RFC2193 IMAP4 Mailbox Referrals
736 RFC2221 IMAP4 Login Referrals
739 <table width="100%" cellpadding="0" summary="Canned page footer">
741 <td width="30%">Back to <a href="index.html">Fetchmail Home Page</a></td>
742 <td width="30%" align="center">To <a href="/~esr/sitemap.html">Site Map</a></td>
743 <td width="30%" align="right">$Date: 2002/07/28 09:23:04 $</td>
748 <address>Eric S. Raymond <a
749 href="mailto:esr@thyrsus.com"><esr@snark.thyrsus.com></a></address>