Re: Suggested Bugzilla improvements

Re: Suggested Bugzilla improvements

[Volker Reichelt]

On 13 May, Daniel Berlin wrote:

>> * What's the purpose of the URL field?
>>   Sending an URL instead of a preprocessed file is a no-no.
>>   Doesn't that encourage users to send us URLs?
> It was there, i never removed it, and probably won't (though i could 
> make it invisible if you like).

Please make it invisible (that's at least my opinion).

>> * I'd like to propose the keyword "diagnostic" for reports that deal
>>   with malformed/strange error messages and warnings.
>>   There are more than 20 bugs that have the stamp "[diagnostic]" in the
>>   synopsis line right now.
> 
> People who can edit bugs can edit/add/remove keywords, specifically so 
> you don't need my approval to do this type of thing.

Great.

>> * The page "bug_status.html" needs updating, especially the parts about
>>   the Platform/Operating System.
>>
> Thanks, i'll do it.

The page should probably be merged with "bugs/management.html" somehow
when Bugzilla goes online. One page of documentation should be enough.

Regards,
Volker

Re: Suggested Bugzilla improvements

[Daniel Berlin]

On Tuesday, May 13, 2003, at 11:58  AM, Volker Reichelt wrote:

> On 13 May, Daniel Berlin wrote:
>
>>> * What's the purpose of the URL field?
>>>   Sending an URL instead of a preprocessed file is a no-no.
>>>   Doesn't that encourage users to send us URLs?
>> It was there, i never removed it, and probably won't (though i could
>> make it invisible if you like).
>
> Please make it invisible (that's at least my opinion).
Done. It's now invisible.

>>> * The page "bug_status.html" needs updating, especially the parts 
>>> about
>>>   the Platform/Operating System.
>>>
>> Thanks, i'll do it.
>
> The page should probably be merged with "bugs/management.html" somehow
> when Bugzilla goes online. One page of documentation should be enough.
>
Yup.
I seem to have lost the message saying what other docs on the gcc 
website i should update when it goes live.
Can someone give me the list again, if it's not too much trouble?
I'm not talking about Bugzilla's docs, i mean the gcc web pages.


I seem to remember that we already took care of the internal-only docs 
related to branch and release creation.
--Dan
> Regards,
> Volker
>
>

Re: Suggested Bugzilla improvements

[Joseph S. Myers]

On Tue, 13 May 2003, Daniel Berlin wrote:

> I seem to have lost the message saying what other docs on the gcc 
> website i should update when it goes live.
> Can someone give me the list again, if it's not too much trouble?
> I'm not talking about Bugzilla's docs, i mean the gcc web pages.

Everywhere on the website mentioning GNATS, as found by grep.

Everywhere in the manual mentioning GNATS, likewise.  In particular the
checklist in sourcebuild.texi of what's needed when adding a new front
end.

> I seem to remember that we already took care of the internal-only docs 
> related to branch and release creation.

They describe updating versions, milestones and bug_email.pl.

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: Suggested Bugzilla improvements

[Gerald Pfeifer]

Daniel,

here is some further information in addition to what Joseph wrote.

On Tue, 13 May 2003, Joseph S. Myers wrote:
>> I seem to have lost the message saying what other docs on the gcc
>> website i should update when it goes live.
>> Can someone give me the list again, if it's not too much trouble?
>> I'm not talking about Bugzilla's docs, i mean the gcc web pages.
> Everywhere on the website mentioning GNATS, as found by grep.

Most of that is gnats.html, which we should replace by a suitable page
for bugzilla (note that in bugs/ we have some further information on
our bug tracking effort, some of which we already generalized from GNATS),
and some references in bugs.html.

> Everywhere in the manual mentioning GNATS, likewise.

Wolfgang will soon commit a patch for the main manual, and already has
submitted one for the Fortran manual, so you can safely ignore these two.

Gerald
-- 
Gerald "Jerry"   pfeifer@dbai.tuwien.ac.at   http://www.pfeifer.com/gerald/

Suggested Bugzilla improvements

[Volker Reichelt]

Hi Daniel,

thanks for your work with Bugzilla!

Alas I wasn't able to take part in the discussion yesterday since our
mailserver broke shortly after my initial posting of the "Suggestion
for a new GNATS policy" thread.

But it looks like most of yesterdays problems (real ones or just
misunderstandings) are resolved now.

There are, however, some points that still need to be addressed IMHO.

* Version: By now, one can only select one version. The bugs, however,
  usually appear in more than one compiler version.
  Would it be possible to change that into a text field or to rename it
  into "Initial version"? Or are there better suggestions?

* What's the purpose of the URL field?
  Sending an URL instead of a preprocessed file is a no-no.
  Doesn't that encourage users to send us URLs?

* I'd like to propose the keyword "diagnostic" for reports that deal
  with malformed/strange error messages and warnings.
  There are more than 20 bugs that have the stamp "[diagnostic]" in the
  synopsis line right now.

* The page "bug_status.html" needs updating, especially the parts about
  the Platform/Operating System.

Regards,
Volker

Re: Suggested Bugzilla improvements

[Joseph S. Myers]

On Tue, 13 May 2003, Volker Reichelt wrote:

> * Version: By now, one can only select one version. The bugs, however,
>   usually appear in more than one compiler version.
>   Would it be possible to change that into a text field or to rename it
>   into "Initial version"? Or are there better suggestions?

We have at present a full version number field (with date), which is
converted into the version number only field (for searching) and the full
version number inserted in the text (so we can tell which of the very wide
range of compilers calling themselves "3.3" the bug was reported against).  
But tracking which set of versions a bug applies in seems to be a
difficult problem for bug tracking systems in general.

Also, what's happened to the various minor versions that were in the
version list?  We did have 2.95.[1234], 3.0.[12345], 3.2.[123] included as
possible short version numbers - and there is documentation in
releasing.html / branching.html of when and how to add new numbers when
releasing and branching - but they are no longer available, although the 
minor versions 3.1.1 and 3.1.2 are.

Didn't someone request that the keywords with legal/illegal in them should
be changed to use the GNUically correct terminology valid/invalid as part
of the conversion?

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: Suggested Bugzilla improvements

[Daniel Berlin]

On Tuesday, May 13, 2003, at 06:17  AM, Joseph S. Myers wrote:

> On Tue, 13 May 2003, Volker Reichelt wrote:
>
>> * Version: By now, one can only select one version. The bugs, however,
>>   usually appear in more than one compiler version.
>>   Would it be possible to change that into a text field or to rename 
>> it
>>   into "Initial version"? Or are there better suggestions?
>
> We have at present a full version number field (with date), which is
> converted into the version number only field (for searching) and the 
> full
> version number inserted in the text (so we can tell which of the very 
> wide
> range of compilers calling themselves "3.3" the bug was reported 
> against).
> But tracking which set of versions a bug applies in seems to be a
> difficult problem for bug tracking systems in general.
>
Correct.

> Also, what's happened to the various minor versions that were in the
> version list?  We did have 2.95.[1234], 3.0.[12345], 3.2.[123] 
> included as
> possible short version numbers - and there is documentation in
> releasing.html / branching.html of when and how to add new numbers when
> releasing and branching - but they are no longer available,
This is my fault, it's checking them in the wrong order, so it checks 
for 2.95 before 2.95.3, etc.

>  although the
> minor versions 3.1.1 and 3.1.2 are.
>
> Didn't someone request that the keywords with legal/illegal in them 
> should
> be changed to use the GNUically correct terminology valid/invalid as 
> part
> of the conversion?

Nope.

>
> -- 
> Joseph S. Myers
> jsm28@cam.ac.uk

Re: Suggested Bugzilla improvements

[Joseph S. Myers]

On Tue, 13 May 2003, Daniel Berlin wrote:

> > Didn't someone request that the keywords with legal/illegal in them 
> > should
> > be changed to use the GNUically correct terminology valid/invalid as 
> > part
> > of the conversion?
> 
> Nope.

The request in question was
<http://gcc.gnu.org/ml/gcc/2003-02/msg01688.html>.

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: Suggested Bugzilla improvements

[Daniel Berlin]

On Tuesday, May 13, 2003, at 10:52  AM, Joseph S. Myers wrote:

> On Tue, 13 May 2003, Daniel Berlin wrote:
>
>>> Didn't someone request that the keywords with legal/illegal in them
>>> should
>>> be changed to use the GNUically correct terminology valid/invalid as
>>> part
>>> of the conversion?
>>
>> Nope.
>
> The request in question was
> <http://gcc.gnu.org/ml/gcc/2003-02/msg01688.html>.
Must have missed that followup.
I'll make it happen.

Re: Suggested Bugzilla improvements

[Daniel Berlin]

On Tuesday, May 13, 2003, at 05:12  AM, Volker Reichelt wrote:

> Hi Daniel,
>
> thanks for your work with Bugzilla!
>
> Alas I wasn't able to take part in the discussion yesterday since our
> mailserver broke shortly after my initial posting of the "Suggestion
> for a new GNATS policy" thread.
>
> But it looks like most of yesterdays problems (real ones or just
> misunderstandings) are resolved now.
>
> There are, however, some points that still need to be addressed IMHO.
>
> * Version: By now, one can only select one version. The bugs, however,
>   usually appear in more than one compiler version.
>   Would it be possible to change that into a text field or to rename it
>   into "Initial version"? Or are there better suggestions?
>
> * What's the purpose of the URL field?
>   Sending an URL instead of a preprocessed file is a no-no.
>   Doesn't that encourage users to send us URLs?
It was there, i never removed it, and probably won't (though i could 
make it invisible if you like).

>
> * I'd like to propose the keyword "diagnostic" for reports that deal
>   with malformed/strange error messages and warnings.
>   There are more than 20 bugs that have the stamp "[diagnostic]" in the
>   synopsis line right now.

People who can edit bugs can edit/add/remove keywords, specifically so 
you don't need my approval to do this type of thing.

>
> * The page "bug_status.html" needs updating, especially the parts about
>   the Platform/Operating System.
>
Thanks, i'll do it.

> Regards,
> Volker
>
>