Re: Some talk about FromNameSpoof and async coding

2022-05-05 Thread Henrik K
On Fri, May 06, 2022 at 07:49:14AM +0300, Henrik K wrote: > > I see now that we can keep adding evals into queue as they are called, and > keep the check_cleanup() around to make sure any pending queue is run. Now fixed and better commented in revision 1900613.

Re: Some talk about FromNameSpoof and async coding

2022-05-05 Thread Henrik K
On Fri, May 06, 2022 at 07:49:14AM +0300, Henrik K wrote: > > I was initially scared of users messing with priorities so that some of the > evals run at completely different times. This can happen for example with > meta rules and $pms->fix_priorities() which can adjust priorities if some of >

Some talk about FromNameSpoof and async coding

2022-05-05 Thread Henrik K
On Thu, May 05, 2022 at 10:25:07PM +0200, Michael Storz wrote: > > @Henrik: I was only surprised that you took the asynchronous method with > rule_pending/rule_ready + check_cleanup instead of the method for tags with > action_depends_on_tags. This separates the evaluation of the eval functions >

Re: svn commit: r1900597 - /spamassassin/trunk/t/mkrules_else.t

2022-05-05 Thread Sidney Markowitz
Thanks, Henrik for this and the other little test things. They were bothering me every time I ran tests the past few days while working on the big test issues, but I had to put off fixing them. Isn't nice when tests produce so little noise that you can finally see the cosmetic or little

Re: svn commit: r1900514 - /spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/FromNameSpoof.pm

2022-05-05 Thread Michael Storz
Am 2022-05-03 17:08, schrieb Henrik K: I really dislike these kinds of "band-aids" which really don't help the main cause: terribly convoluted plugin code. Why skip debug output if owner is missing? Is the a good reason for owner missing, what is the cause for that and can it be better

[Bug 7981] Remaining tasks for 4.0.0 release

2022-05-05 Thread bugzilla-daemon
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7981 --- Comment #3 from Sidney Markowitz --- A pending task for a full release is to write the description of new features and important changes in this version that will go into the build/announcements/4.0.0.txt file, and to look at the

[Bug 7986] spamd test fails when in directory with absolute pathname length greater than 50

2022-05-05 Thread bugzilla-daemon
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7986 Sidney Markowitz changed: What|Removed |Added Resolution|--- |FIXED Status|NEW

Rule updates are too old - 2022-05-05 3.3.0:2022-05-04 3.3.1:2022-05-04 3.3.2:2022-05-04

2022-05-05 Thread darxus
SpamAssassin version 3.3.0 has not had a rule update since 2022-05-04. SpamAssassin version 3.3.1 has not had a rule update since 2022-05-04. SpamAssassin version 3.3.2 has not had a rule update since 2022-05-04. 20220504: Spam and ham are above threshold of 150,000:

[Bug 7986] spamd test fails when in directory with absolute pathname length greater than 50

2022-05-05 Thread bugzilla-daemon
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7986 --- Comment #9 from Sidney Markowitz --- (In reply to Benny Pedersen from comment #8) > userprefs.db should not be in tempdir but in datadir This only has to do with the path used for sockets in the test scripts that test spamd. It isn't

[Bug 7986] spamd test fails when in directory with absolute pathname length greater than 50

2022-05-05 Thread bugzilla-daemon
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7986 Benny Pedersen changed: What|Removed |Added CC||m...@junc.eu --- Comment #8 from

[Bug 7986] spamd test fails when in directory with absolute pathname length greater than 50

2022-05-05 Thread bugzilla-daemon
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7986 Sidney Markowitz changed: What|Removed |Added Summary|spamd fails when in |spamd test fails when in