Request for bugzilla submitters/maintainers: "Reportad against" field

Request for bugzilla submitters/maintainers: "Reportad against" field

[Joe Buck]

The "Reported against:" field in Bugzilla is less useful than it could be;
it frequently seems to hold some random value.  For example, someone
reports a bug against 3.3.3, notices that it is also in the trunk, and
puts "3.5.0" in "Reported against:".  This makes it harder to do release
notes later, because I don't want release notes to contain mentions of
bugs that never existed in any released GCC (the release notes say what's
fixed as compared to former GCC releases).

So, if you know that a bug exists in a released GCC, please use its
version number in the "Reported against:" field.  You can use "Known to
work" and "Known to fail" to contain any info about which development
branches contain or don't contain the bug.

Thanks.

Re: Request for bugzilla submitters/maintainers: "Reportad against" field

[Daniel Berlin]

On Tue, 23 Mar 2004, Joe Buck wrote:

> The "Reported against:" field in Bugzilla is less useful than it could be;
> it frequently seems to hold some random value.  For example, someone
> reports a bug against 3.3.3, notices that it is also in the trunk, and
> puts "3.5.0" in "Reported against:".  This makes it harder to do release
> notes later, because I don't want release notes to contain mentions of
> bugs that never existed in any released GCC (the release notes say what's
> fixed as compared to former GCC releases).
By way of explanation of the purpose of "Reported Against"

Reported against was originally "Version", which makes no sense in the
face of known to work/known to fail.

Thus, it simply became a way to determine what version of gcc the bug was
reported against, as opposed to what versions it exists in.
In fact, unless the submitter chose the wrong version, i'm not sure it
should ever be changed during the life of the bug.

Re: Request for bugzilla submitters/maintainers: "Reportad against" field

[Joe Buck]

On Tue, Mar 23, 2004 at 01:48:42PM -0500, Daniel Berlin wrote:
> By way of explanation of the purpose of "Reported Against"
> 
> Reported against was originally "Version", which makes no sense in the
> face of known to work/known to fail.
> 
> Thus, it simply became a way to determine what version of gcc the bug was
> reported against, as opposed to what versions it exists in.
> In fact, unless the submitter chose the wrong version, i'm not sure it
> should ever be changed during the life of the bug.

In the couple of cases I looked at, the submitter himself mentioned a
released GCC version that had the bug, but still marked "3.4.0" or "3.5.0"
as "Reported against".

Re: Request for bugzilla submitters/maintainers: "Reportad against" field

[Giovanni Bajo]

Joe Buck wrote:

> In the couple of cases I looked at, the submitter himself mentioned a
> released GCC version that had the bug, but still marked "3.4.0" or
> "3.5.0" as "Reported against".

We can change the field values. For instance, we can remove "3.4.0" and "3.5.0"
until they are *really* released, and put "CVS 3.4 branch" and "CVS mainline",
and the same for known-to-work/fail. Eventually, you probably want to use
known-to-fail since reported against is not really meaningful anymore.

Giovanni Bajo