Suggestion for new keywords in bugzilla

Suggestion for new keywords in bugzilla

[Volker Reichelt]

Right now we have the following keywords in bugzilla:

accepts-invalid 
ice-on-invalid-code
ice-on-valid-code
pessimizes-code
rejects-valid
wrong-code

I'd like to add the following keywords:

* diagnostic (for error messages/warnings that are misleading or broken
  and superfluous/missing warnings)
* error-recovery (we get an ICE/bailing-out-message after a sensible
  error message). This should be specified in addition to
  ice-on-invalid-code.

What do you think?

Regards,
Volker

Re: Suggestion for new keywords in bugzilla

[Giovanni Bajo]

Volker Reichelt <reichelt@igpm.rwth-aachen.de> wrote:

> I'd like to add the following keywords:
>
> * diagnostic (for error messages/warnings that are misleading or broken
>   and superfluous/missing warnings)

Maybe this should be a new component (like, c, c++, etc). On the other hand,
it's clear that e.g. some diagnosis bugs are stricly C++ related. I'm not
sure what it's best option. Of course, everything is better than the
[diagnosis] stamp in the summary.

> * error-recovery (we get an ICE/bailing-out-message after a sensible
>   error message). This should be specified in addition to
>   ice-on-invalid-code.

Yes, agreed. Is it possible to have some dropbox with keywords instead of
being forced to type them in manually?

Giovanni Bajo

Re: Suggestion for new keywords in bugzilla

[Giovanni Bajo]

Nathanael Nerode <neroden@twcny.rr.com> wrote:

>> I think the biggest difference is between diagnostic bugs which are
request
>> for enhancements and not strictly necessary, and diagnostic bugs which
are
>> really bugs (or at least, can be considered as such). And this difference
>> is already taken into account by the Severity of the bug. In other words
I
>> suggest that your missing-diagnostic be simply a diagnostic keyword with
>> Severity = enhancement, and you misleading-diagnostic be a diagnostic
>> keyword with Severity != enhancement.

> This isn't *exactly* what I was thinking of.  A missing diagnostic might
> be a warning which is supposed to trigger and doesn't.  A misleading
> diagnostic would be one which triggers but shouldn't.
>
> It may be that the distinction is not valuable.  I defer to others.


Since we got no more comments on this issue, I'm creating a new keyword
"diagnostic" for now, and removing all the [diagnostic] stamps. We can still
split diagnostic into smaller parts later if we need. OTOH, Gaby never asked
for it.

Giovanni Bajo

Re: Suggestion for new keywords in bugzilla

[Nathanael Nerode]

Giovanni Bajo wrote:
> Nathanael Nerode wrote:
> 
> 
>>I actually thought that there should be two keywords for this:
> 
> 
>>missing-diagnostic (There should be a warning, but there isn't.)
>>misleading-diagnostic (There's a warning which shouldn't be there, or
>>should be replaced by a different warning.)
> 
> 
> The latter (misleading-diagnostic) can apply also to error messages. There
> are situations where error messages are totally wrong (for instance, they
> show unreadable names because they can't pretty-print correctly them, or
> they are totally unrelated to the real error), or could be improved (for
> instance, a simply "parse error", or "expected primary-expression" instead
> of something more detailed which would help the user understanding the
> problem).
> 
> But in the first place, why do you think that two keywords could be useful?
> I think the biggest difference is between diagnostic bugs which are request
> for enhancements and not strictly necessary, and diagnostic bugs which are
> really bugs (or at least, can be considered as such). And this difference is
> already taken into account by the Severity of the bug. In other words I
> suggest that your missing-diagnostic be simply a diagnostic keyword with
> Severity = enhancement, and you misleading-diagnostic be a diagnostic
> keyword with Severity != enhancement.
This isn't *exactly* what I was thinking of.  A missing diagnostic might 
be a warning which is supposed to trigger and doesn't.  A misleading 
diagnostic would be one which triggers but shouldn't.

It may be that the distinction is not valuable.  I defer to others.

> 
> Giovanni Bajo
> 
> 

Re: Suggestion for new keywords in bugzilla

[Nathanael Nerode]

Volker Reichelt wrote:
> Nathanael Nerode wrote:
> 
> 
>>>>* diagnostic (for error messages/warnings that are misleading or broken
>>>>  and superfluous/missing warnings)
>>
>>I actually thought that there should be two keywords for this:
>>
>>missing-diagnostic (There should be a warning, but there isn't.)
>>misleading-diagnostic (There's a warning which shouldn't be there, or 
>>should be replaced by a different warning.)
> 
> 
> I don't think that we really need two keywords, because the volume is low:
> We now have thirty-something PRs with the stamp "[diagnostic]" in the
> summary line. They all fit on a single screen. And even if the number
> doubled, the number would still be small compared to "ice-on-(in)valid-code"
> for example.
Missing diagnostics and misleading diagnostics are often very different 
kinds of issues requiring different skills to fix, which is why I 
thought the distinction might be useful.

I defer to the opinion of this list. :-)

> And we still have the summary line where we can give information like
> "misleading".
> 
> Btw, thanks for adding the keyword "error-recovery"!
I saw consensus. :-)

> Regards,
> Volker

Re: Suggestion for new keywords in bugzilla

[Volker Reichelt]

Nathanael Nerode wrote:

> >> * diagnostic (for error messages/warnings that are misleading or broken
> >>   and superfluous/missing warnings)
> 
> I actually thought that there should be two keywords for this:
> 
> missing-diagnostic (There should be a warning, but there isn't.)
> misleading-diagnostic (There's a warning which shouldn't be there, or 
> should be replaced by a different warning.)

I don't think that we really need two keywords, because the volume is low:
We now have thirty-something PRs with the stamp "[diagnostic]" in the
summary line. They all fit on a single screen. And even if the number
doubled, the number would still be small compared to "ice-on-(in)valid-code"
for example.

And we still have the summary line where we can give information like
"misleading".

Btw, thanks for adding the keyword "error-recovery"!

Regards,
Volker

Re: Suggestion for new keywords in bugzilla

[Giovanni Bajo]

Nathanael Nerode wrote:

>I actually thought that there should be two keywords for this:

>missing-diagnostic (There should be a warning, but there isn't.)
>misleading-diagnostic (There's a warning which shouldn't be there, or
>should be replaced by a different warning.)

The latter (misleading-diagnostic) can apply also to error messages. There
are situations where error messages are totally wrong (for instance, they
show unreadable names because they can't pretty-print correctly them, or
they are totally unrelated to the real error), or could be improved (for
instance, a simply "parse error", or "expected primary-expression" instead
of something more detailed which would help the user understanding the
problem).

But in the first place, why do you think that two keywords could be useful?
I think the biggest difference is between diagnostic bugs which are request
for enhancements and not strictly necessary, and diagnostic bugs which are
really bugs (or at least, can be considered as such). And this difference is
already taken into account by the Severity of the bug. In other words I
suggest that your missing-diagnostic be simply a diagnostic keyword with
Severity = enhancement, and you misleading-diagnostic be a diagnostic
keyword with Severity != enhancement.

Giovanni Bajo

Re: Suggestion for new keywords in bugzilla

[Joseph S. Myers]

On Thu, 22 May 2003, Nathanael Nerode wrote:

> missing-diagnostic (There should be a warning, but there isn't.)
> misleading-diagnostic (There's a warning which shouldn't be there, or 
> should be replaced by a different warning.)

FWIW, in GNATS rejects-legal and accepts-illegal were sometimes overloaded
to refer to wrong warnings on correct code and warnings that would be
desirable even if the code wasn't actually invalid per the standard.  
With new keywords we can be more specific and avoid this overloading.

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

Re: Suggestion for new keywords in bugzilla

[Michael S. Zick]

On Thursday 22 May 2003 12:13 pm, Nathanael Nerode wrote:
> On 22 May, Giovanni Bajo wrote:
> > Volker Reichelt <reichelt@igpm.rwth-aachen.de> wrote:
>
> <snip>
>
> >> * diagnostic (for error messages/warnings that are misleading or
>
> broken
>
> >>   and superfluous/missing warnings)
>
> I actually thought that there should be two keywords for this:
>
> missing-diagnostic (There should be a warning, but there isn't.)
diagnostic-missing

> misleading-diagnostic (There's a warning which shouldn't be there, or
diagnostic-wrong

> should be replaced by a different warning.)
>
> Better names are solicited.
>
> --Nathanael

Re: Suggestion for new keywords in bugzilla

[Nathanael Nerode]

>* error-recovery (we get an ICE/bailing-out-message after a sensible
>  error message). This should be specified in addition to
>  ice-on-invalid-code.

I added this; I think we have consensus.

--Nathanael

Re: Suggestion for new keywords in bugzilla

[Nathanael Nerode]

On 22 May, Giovanni Bajo wrote:
> Volker Reichelt <reichelt@igpm.rwth-aachen.de> wrote:
<snip>
>> * diagnostic (for error messages/warnings that are misleading or 
broken
>>   and superfluous/missing warnings)

I actually thought that there should be two keywords for this:

missing-diagnostic (There should be a warning, but there isn't.)
misleading-diagnostic (There's a warning which shouldn't be there, or 
should be replaced by a different warning.)

Better names are solicited.

--Nathanael

Re: Suggestion for new keywords in bugzilla

[Daniel Berlin]

>
>> I'm not sure what it's best option. Of course, everything is better 
>> than the
>> [diagnosis] stamp in the summary.
>
>>> * error-recovery (we get an ICE/bailing-out-message after a sensible
>>>   error message). This should be specified in addition to
>>>   ice-on-invalid-code.
>>
>> Yes, agreed. Is it possible to have some dropbox with keywords 
>> instead of
>> being forced to type them in manually?
This is being worked on by someone else (If you look at 
bugzilla.mozilla.org's bugzilla bugs, you'll see one for this 
enhancement).
When they get it done, i'll add it to our bugzilla.

Re: Suggestion for new keywords in bugzilla

[Volker Reichelt]

On 22 May, Giovanni Bajo wrote:
> Volker Reichelt <reichelt@igpm.rwth-aachen.de> wrote:
> 
>> I'd like to add the following keywords:
>>
>> * diagnostic (for error messages/warnings that are misleading or broken
>>   and superfluous/missing warnings)
> 
> Maybe this should be a new component (like, c, c++, etc). On the other hand,
> it's clear that e.g. some diagnosis bugs are stricly C++ related.

That was my intention for making it a keyword.

> I'm not sure what it's best option. Of course, everything is better than the
> [diagnosis] stamp in the summary.

>> * error-recovery (we get an ICE/bailing-out-message after a sensible
>>   error message). This should be specified in addition to
>>   ice-on-invalid-code.
> 
> Yes, agreed. Is it possible to have some dropbox with keywords instead of
> being forced to type them in manually?

Me, too!

Regards,
Volker