http://issues.apache.org/SpamAssassin/show_bug.cgi?id=5419
[EMAIL PROTECTED] changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|3.2.3 |3.2.2
------- Additional Comments From [EMAIL PROTECTED] 2007-06-29 05:04 -------
while waiting for 3.2.2 patch reviews, I decided to have a go at this; I think
this may fix it.
Unfortunately it's not possible to find out exactly what the "real" ARGV array
was when the script was started; instead, what the change does is check the #!
line on the perl script, and if the 2 perl interpreters are the same, it just
execs the script instead of "perl -T -w /the/script". I think that should work.
btw, this is safer than the alternative approach: 'Can the correct perl
interpreter not be specified with PERL_BIN when making the
Makefile, negating the need for r490760?'
the problem with this is that the user could install using /usr/bin/perl, then
later decide to run spamd using /usr/local/bin/perl (or whatever). there's
nothing forbidding them from doing this! when SIGHUP'ed, the spamd may not
restart with the same behaviour since it'd then wind up using a different perl
interpreter. we want to avoid that.
btw -- I have a minor problem trying to repro a test case. In
http://issues.apache.org/SpamAssassin/show_bug.cgi?id=5255#c2 , Martin Lathoud
noted:
'The fix [for bug 5255] creates a side effect at least on Linux, /proc/$PID/exe
points to /usr/bin/perl instead of /usr/bin/spamd, so some proc tools won't find
spamd anymore after a -HUP (i.e pstree, pidof), showing perl instead.'
however when I run spamd, even before SIGHUP, "pidof spamd" only ever lists the
children -- not the parent spamd process! (That's listed under "pidof perl").
But given that pstree works though, and reports what Martin wanted, I think
that's good enough anyway. let's see how it does in trunk against the buildbots
before I make a patch...
: jm 185...; svn commit -m "bug 5419: when re-execing the perl interpreter in
response to a SIGHUP, the fix for bug 5255 caused the ARGV[0] to change from
'spamd' to 'perl' under certain circumstances. fix"
Sending spamd/spamd.raw
Transmitting file data .
Committed revision 551862.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.