This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Future of gccbug
Daniel Berlin <dberlin@dberlin.org> writes:
>> Hmm, then something is wrong somewhere. Everything that is entered in
>> bugzilla gets posted to me with a To: gcc-bugs and Reply-To:
>> gcc-bugzilla.
>
> Which is correct, don't wide-reply.
>
>> It happens that a wide-reply sends mails to both addresses and things
>> get duplicated. My suggestion is to have a separate list for
>> gcc-bugzilla traffic.
>
> This is why the top of the email says what it says :P
>
> If i had any clue how our list software worked (i don't even know what
> the heck we use), i would make it ignore any replies where
> gcc-bugzilla is copied.
>
> Remember, only copying gcc-bugzilla actually gets the comment into the
> database.
>
> And we want all bug discussion to be in the bug db.
> So don't copy gcc-bugs, only gcc-bugzilla.
Adding Mail-Copies-To: and Mail-Followup-To: headers to bugzilla
generated mail may help with this, at least with some MUAs. See
http://cr.yp.to/proto/replyto.html [I think this made it into
one of the DRUMS proposed RFCs but I can't find them right now].
Since you're soliciting suggestions:
It would help if all the messages sent out in response to an update of
a single bug had the same message-id so that duplicate message filters
would work on them. Right now if someone adds a comment to a bug that
I reported or is assigned to me, I get one copy of the comment in my
inbox (which is what I want) and another copy in the gcc-bugs folder.
The [Bug foo/nnn] tags tend to shove the actual subject off the
right-hand side of the gcc-bugs summary window. Could these be moved
to the end of the line? And could we dispense with the foo/ bit,
which AFAIunderstand is a relict of GNATS?
The [3.x regression] tags also do this - aren't these supposed to be
keywords?
The From: munging causes BBDB to go nuts.
The text of a bugzilla mailing has an immense amount of "noise".
Consider:
| From: "pinskia at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org>
| Subject: [Bug debug/12860] [3.4 regression] stabs register number out of range
| To: gcc-bugs@gcc.gnu.org
| Date: 7 Nov 2003 18:16:12 -0000
|
|
| PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
|
| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12860
|
|
| pinskia at gcc dot gnu dot org changed:
|
| What |Removed |Added
| ----------------------------------------------------------------------------
| Keywords| |wrong-code
| Target Milestone|--- |3.4
|
|
| ------- Additional Comments From pinskia at gcc dot gnu dot org 2003-11-07 18:16 -------
| Looks like the psedu registers from gcc are being used for the debug info.
The actual content of interest is lost in all this boilerplate.
I would far rather see something like this instead:
| From: Andrew Pinski <pinskia@gcc.gnu.org>
| Subject: stabs register number out of range [Bug 12860]
| To: gcc-bugs@gcc.gnu.org
| Date: 7 Nov 2003 18:16:12 -0000
|
| Looks like the psedu registers from gcc are being used for the debug info.
|
| --
| Keywords: +wrong-code
| Target Milestone: ->3.4
|
| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12860
where that -- is a proper sig delimiter (dash dash space newline).
The munging of the From: line *should* be obviated by proper MCT/MRT
headers.
zw