| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
artifacts are removed, as is the C files which we will no longer need.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
uncleanly (e.g. segfaults the Python interpreter), messages being processed
will not be lost.
The vulnerability, ideas, and patches are credited to Richard Barrett and Mark
Sapiro. Their original work was modified by Barry for this commit and any
bugs are his fault.
The basic idea is that instead of unlinking a .pck file in dequeue(), the file
is renamed to a .bak file. The Switchboard grows a finish() method which then
unlinks the .bak file. That class's constructor also grows a 'restore'
argument (defaulting to false), which when true moves all .bak files it finds
in its hash space to .pck, thereby restoring a file lost while "in flight".
This relies on the fact that even with multiple qrunners, exactly one process
will be responsible for one hash space slice, so it's never possible (under
normal operation) for a .bak file to be renamed to .pck by some other process.
Test cases for both the new Switchboard behavior and the use of that by Runner
subclasses has been added.
There are two things to watch out for, either of which may require some
additional changes. There is some small potential to duplicate messages in
various queues, if say 'mailmanctl' were improperly started more than once by
a site admin. This usually won't happen unless an admin is overly eager with
the mailmanctl -s switch, so we can chalk this one up to operator error. I'm
not sure what more we can do about that.
There's also a possibility that if we're processing a message that continually
causes the Python interpreter to crash, we could end up duplicating messages
endlessly. This is especially troublesome for the Outgoing runner which could
conceivably cause a mail flood. I consider this the more critical issue to
defend against, probably by adding a numbering scheme to the .bak file names
and refusing to restore a .bak file more than say 3 times without human
intervention.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Jr, after much checking and rechecking (and some massaging) by me. Checking
in now before I fall asleep and forget what all this was for.
This patch adds the ability to 'make DESTDIR=/some/dir/tree install' that
doesn't influence the paths stored in e.g. Defaults.py at all, which is of
good use for various package managers. It is not all that package managers
must do, however! Running make install with DESTDIR set means bin/update is
never run, and mm_cfg.py is always written; package managers should make
sure the appropriate post-installation is done, and that mm_cfg.py is
treated as a config file.
This patch inadvertently fixes some bogus whitespace: 8-spaces where
surrounding code used tabs. The difference was harmless because the 8-spaces
were used inside shell-continued-oneliners, but it is confusing and could
lead to future harm. I'm too tired to make those two or three changes in a
separate checkin, sorry.
This patch also assumes the various packages that are installed using
distutils do not record (or rather, use) their installation paths anywhere,
but this seems to hold true at least for the moment.
Also, I've done so many slow cvs diff's, I'm wondering when we'll switch to
Subversion. Unfortunately, I've also done so many 'cvs diff -c > file; patch
-p0 -R < file's to switch back and forth between patches and change sets,
I'm wondering when we'll switch to Aegis as well. :-P
|
| | |
|
| | |
|
| | |
|
|
|
Runner is the base class providing the framework for other qrunner
specializations.
IncomingRunner runs through qfiles/in taking messages from the smtpd,
sending them through the standard message handler pipeline and
potentially writing the message to qfiles/out or qfiles/news.
OutgoingRunner takes message from qfiles/out and hands them over to
the MTA.
NewsRunner takes message from qfiles/news and hands them over to the
news server.
|
