| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I have a misconfigured MTA server to which I transfer mail from an IMAP
account using fetchmail. Although I use the --softbounce switch,
fetchmail still deletes the email even in case of unsuccessful delivery.
The problem is caused by the fact, that the close_sink() call returns
with the value TRUE not only in case of successful delivery, but also in
case of a permanent (code 5xx) error. Then the fetch_messages()
procedure will no longer know, whether the message can be deleted or
not, because it does not have the information that the close_sink() call
returned TRUE because the delivery was successful or because it is
expected that the delivery will not be successful later either, so the
value of the softbounce switch cannot be taken into account at this
point.
Solution that in the case of a permanent error, the close_sink() call
does not return a fixed TRUE value, but the negated value of the
softbounce switch.
|
|
|
|
|
|
| |
* Normalize include order.
* Backport missed bug fixes.
* Remove dead code.
|
|
|
|
| |
Reported by: clang static analyzer (clang-8)
|
|
|
|
| |
(patch re-indented)
|
| |
|
|
|
|
| |
Reported by Gonzalo Pérez de Olaguer Córdoba, Debian Bug#744907.
|
| |
|
|
|
|
|
|
|
|
| |
This appears to be similar to the recent --nosoftbounce and --nobounce
issue where command line options are not handled exactly the same way as
the rcfile parser handles them.
Reported by Kees Bakker, BerliOS bug #17599.
|
| |
|
| |
|
|
|
|
|
|
|
| |
...that were converted to char* when they should have been converted to
const char *.
Use braces for empty if/else statements.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
* Do not close stdout when using mda and "bsmtp -" at the same time.
* Log operating system errors when BSMTP writes fail.
|
| |
|
|
|
|
|
|
|
| |
There is currently no user outside the compilation unit, so there is no need
to have it visible.
svn path=/branches/BRANCH_6-3/; revision=5464
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Apply patch from Sunil Shetye to fix a problem reported by James Moe.
Before this fix, fetchmail's SMTP client would not recover from errors
such as lost connections that were encountered when fetchmail had sent
RSET, for instance, after an anti-spam filter dropped the connection
after detecting spam. Fetchmail then tried to send subsequent mail
through this broken connection and deferred retrieval until the next
poll. Now, if RSET fails, fetchmail closes the connection and reopens
it for the next message to be delivered.
svn path=/branches/BRANCH_6-3/; revision=5463
|
|
|
|
|
|
|
| |
These are usually configuration errors (missing TLS/SSL). Patch
partially taken from Petr Cerny, Novell's Bugzilla 246829.
svn path=/branches/BRANCH_6-3/; revision=5339
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Short timeouts could cause fetchmail to not wait long enough for the
"250 Ok" after shipping a long message, particularly with synchronous
mailers and extensive spam filtering. This caused fetchmail to re-fetch
long messages.
While the actual fix is making sure that the timeout is no shorter than
the time the SMTP server takes to process the message, we now enforce
the minimum RFC-5321 recommended timeouts even if the user configures a
lower timeout.
This is to fix Berlios Bug #10972, reported by Viktor Binzberger.
NOTE: it is untested whether we will properly delete the message from
the POP3/IMAP server or mark it as seen, as the upstream server may
close the connection sooner.
svn path=/branches/BRANCH_6-3/; revision=5338
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=5333
|
|
|
|
|
|
|
|
| |
Non-delivery messages now mention the original reason for the bounce
message again. It was lost in merging Holger Mauermann's patch before
6.3.0, and caused a sink.c compiler warning ever since.
svn path=/branches/BRANCH_6-3/; revision=5332
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=5331
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=5261
|
|
|
|
|
|
|
| |
when the local SMTP listener rejects a warning message of fetchmail's.
Report and suggested fix by Earl Chew.
svn path=/branches/BRANCH_6-3/; revision=5119
|
|
|
|
|
|
| |
after writing the first header.
svn path=/branches/BRANCH_6-3/; revision=5071
|
|
|
|
|
|
|
| |
fails. Improve error checking and reporting. Reported by Reto Schttel,
Debian Bug#416625. Fix based on a patch by Nico Golde.
svn path=/branches/BRANCH_6-3/; revision=5070
|
|
|
|
|
|
| |
or other precautions, as it's relaying to the outside.
svn path=/branches/BRANCH_6-3/; revision=5042
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=5041
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4989
|
|
|
|
|
|
|
| |
malformed message. Report by Neil Hoggart (also closes BerliOS bugs #9364,
#9412, #9449).
svn path=/branches/BRANCH_6-3/; revision=4940
|
|
|
|
|
|
| |
This might fix Debian Bug#294547 and Bug#377135.
svn path=/branches/BRANCH_6-3/; revision=4880
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4878
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4780
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4753
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4744
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4736
|
|
|
|
| |
svn path=/branches/BRANCH_6-3/; revision=4655
|
|
|
|
|
|
|
| |
introduced into 6.3.0 when removing alloca(); it caused fetchmail to
free random memory. Reported by Nathaniel W. Turner, Debian Bug#348747.
svn path=/branches/BRANCH_6-3/; revision=4652
|
|
|
|
|
|
|
|
|
| |
- switch back to SMTP after having tried LMTP hosts (multiple smtphost hosts)
- switch back to LMTP after sending a bounce.
The patch removes the global state variable that was the root of this problem.
Patch by Sunil Shetye.
svn path=/branches/BRANCH_6-3/; revision=4581
|
|
|
|
|
|
| |
same SMTP/LMTP connection. Reported by Joachim Feise, Berlios Bug #5849.
svn path=/branches/BRANCH_6-3/; revision=4547
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1. parsed_host is not freed in some cases. This happens when the first
smtp server is down in this setup:
poll mailserver
...
smtphost "smtpserver1" "smtpserver2"
...
2. parsed_host is being initialized for UNIX socket also. For UNIX
socket, parsed_host should be NULL.
3. If EHLO fails on a UNIX socket, it tries HELO on a network socket!
4. ctl->destaddr is allocated memory in two cases. This memory is
never freed.
5. ctl->destaddr was being assigned in a very convoluted manner.
Since, parsed_host is already set correctly now, it can be used
directly.
svn path=/trunk/; revision=4443
|
|
|
|
| |
svn path=/trunk/; revision=4382
|
|
|
|
|
|
| |
Darwin 7.9 build fails in resource.h with ru_utime/ru_stime incomplete types.
svn path=/trunk/; revision=4296
|
|
|
|
| |
svn path=/trunk/; revision=4237
|
|
|
|
| |
svn path=/trunk/; revision=4219
|
|
|
|
| |
svn path=/trunk/; revision=4209
|
|
|
|
|
|
| |
library is no longer available.
svn path=/trunk/; revision=4192
|
|
|
|
|
|
| |
Add a few comments.
svn path=/trunk/; revision=4187
|
|
|
|
| |
svn path=/trunk/; revision=4169
|