Re: [PATCHv2 0/3] Automatic detection of test name problems

[Simon Marchi <simark@simark.ca>]

On 2020-04-29 5:02 a.m., Andrew Burgess wrote:
> * Keith Seitz <keiths@redhat.com> [2020-04-28 12:08:25 -0700]:
> 
>> On 4/27/20 3:01 PM, Andrew Burgess wrote:
>>> Changes since v1:
>>>
>>>   1. Original patch #1 is now merged.
>>>
>>>   2. Functionality is now placed inside a namespace.
>>>   
>>>   3. Better counting for multi-variant test runs, this is inline with
>>>      how Dejagnu's muti-variant result counting works.
>>>
>>>   4. Use 'string first' instead of 'regexp'.
>>>
>>>   5. Reworded commit message on what is now patch #2.
>>>
>>> Further feedback welcome.
>>
>> Wow, that's almost a complete rewrite! While not what I was really
>> suggesting, I have to admit, a /big/ smile came to my face while
>> reading this. It is so nicely done!
>>
>> Thank you!
>>
>> Keith
>>
>> PS. As you know, IANAM, but you are, so I encourage you to approve
>> your patch. ;-)
> 
> Thanks for your feedback, especially your TCL suggestions.
> 
> I'll let the patch sit for a while to see if anyone else has any
> input.
> 
> Thanks,
> Andrew
> 

Hi Andrew,

Thanks for doing this, it's nice.  And by testing this patchset, I've found an offender
and sent a patch for it :)

  https://sourceware.org/pipermail/gdb-patches/2020-April/168052.html

When you detect an offender, do you think you could print something?  A bit like a "FAIL"
is printed when a test fails.  For example:

  DUPLICATE: gdb.base/break.exp: set convenience variable $foo to 81.5

That would make it easier to spot the problematic test(s).  But even without that, your
patchset looks good and useful to me.

Simon