On Wed, Apr 29, 2020 at 06:55:33PM +0200, Nicolas George wrote: > lance.lmw...@gmail.com (12020-04-29): > > Thanks, I catch your point now. Most of existing code haven't return ERROR, > > so I > > choose the same way to process it. If you think it's not OK, we'll change > > more > > code I think. > > Maybe the other code needs to be fixed the same way. Maybe the other > code needs not return an error and this one does. Maybe your change is > actually valid. > > You cannot know unless you first understand what the code does and what > it is supposed to do. You cannot program by imitation, it does not work. > > And you have to test your changes: run a ffmpeg command line, make sure > the new code is triggered, and check the output file, compare it with > when the change is not triggered. If they are both correct, good. If one > is corrupted, then you know your change was bogus. And if you did not > test, then your change cannot be accepted.
Sorry, the old code will segment fault but the new code will not if error happened. so I have no idea what's to compare the output as it's error. Also all of the code which are using avio_get_dyn_buf() didn't check the size for error. I had to say the avio_get_dyn_buf() is designed very well. Can we change avio_get_dyn_buf() to return the error directly? also the description of API isn't clear for the error condition. > > Ideally we can trust regular contributors to have tested their changes > and run FATE before submitting patches. > > Regards, > > -- > Nicolas George > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". -- Thanks, Limin Wang _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".