https://bugs.exim.org/show_bug.cgi?id=1661
jasen Betts <eximbugzilla@???> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |eximbugzilla@???.
| |org
--- Comment #4 from jasen Betts <eximbugzilla@???> ---
(In reply to Jeremy Harris from comment #3)
> Since you've obfuscated, and omitted a lot, including that you're running
> with modified source, not easy to guess what's gone wrong.
Only names and comments have been changed to protect the guilty. The ACL
really was structured like that with large fully commented-out warn sections.
I'm not sure how much I'm allowed to reveal, I don't want to inconvenience you
unnecessarily.
> Did any of your other event handlers create (by using for the first time) any
> new acl variables? This might be a source of corruption in the variables
> data structure.
the only variables I'm modifying in the handlers are the imp[licit acl_arg[123]
(when I do the the ${acl expansion these variables are created, I don't
explcitly modify them).
I once tried modifying other variables in the handler but it didn't work for
me so I went to a funtional model.
> I'd be wary of your n-way branch mux also, as being vulnerable to event names
> it does not code - but that probably isn't the issue here.
yes, I should probably re-do that perhaps using ${extract from a string
literal instead of ${sg: that way I can have a do-nothing default
--
You are receiving this mail because:
You are on the CC list for the bug.
