[FFmpeg-devel] [PATCH v3] avutil/avstring: improve av_strreplace implement

[James Almer]

On 4/5/2017 2:38 PM, Marton Balint wrote:
> 
> 
> On Wed, 5 Apr 2017, James Almer wrote:
> 
>> On 4/5/2017 1:49 PM, Marton Balint wrote:
>>>
>>>
>>> On Wed, 5 Apr 2017, James Almer wrote:
>>>
>>>> On 4/5/2017 11:05 AM, Steven Liu wrote:
>>>>> 2017-04-03 20:32 GMT+08:00 Nicolas George <george at nsup.org>:
>>>>>
>>>>>> Le quartidi 14 germinal, an CCXXV, Steven Liu a écrit :
>>>>>>> change name from av_strreplace to av_strireplace
>>>>>>> Use AVBprint to implement av_strireplace
>>>>>>> add av_strireplace test case TEST_STRIREPLACE
>>>>>>>
>>>>>>> Signed-off-by: Steven Liu <lq at chinaffmpeg.org>
>>>>>>> ---
>>>>>>>  libavutil/avstring.c       | 76 +++++++-----------------------
>>>>>> ----------------
>>>>>>>  libavutil/avstring.h       |  2 +-
>>>>>>>  libavutil/tests/avstring.c | 16 ++++++++++
>>>>>>>  3 files changed, 28 insertions(+), 66 deletions(-)
>>>>>>
>>>>>> This version looks fine to me.
>>>>>>
>>>>>> Maybe wait a little more for advice about the rename and the whole
>>>>>> thing.
>>>>>>
>>>>>> Thanks.
>>>>>>
>>>>>> -- 
>>>>>>   Nicolas George
>>>>>>
>>>>>> _______________________________________________
>>>>>> ffmpeg-devel mailing list
>>>>>> ffmpeg-devel at ffmpeg.org
>>>>>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>>>>
>>>>>>
>>>>> ping
>>>>
>>>> You can't rename the function. It's already in the 3.3 branch.
>>>> This is one of the reasons why giving people enough time for reviews
>>>> is a must before pushing new public symbols.
>>>
>>> Yet the 3.3 release wasn't tagged yet, so it is not _released_. So is there any practical reason for not renaming it before the release?
>>>
>>> Thanks,
>>> Marton
>>
>> Library versions and feature sets in diverging branches. When you
>> branch you do a feature freeze.
>>
>> Making a mess with this for a single function that should have not
>> been applied in such short notice shouldn't be acceptable.
> 
> I don't think it would cause any problems for the users in practice. But we can also remove the function alltogether from the release, so anybody who want's to use the function (with the new name) will only have to check for a single version number.
> 
> Regards,
> Marton

Fine by me. The thing didn't even get its own bump or APIChanges
now that i look at it, so it might as well not exist.
It would hardly be the worst "breakage" we've gone through.