[FFmpeg-devel] [PATCH v4 18/18] tests/fate: Update subtitle tests

[Soft Works]

> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
> Andreas Rheinhardt
> Sent: Saturday, 11 September 2021 12:02
> To: ffmpeg-devel at ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v4 18/18] tests/fate: Update
> subtitle tests
> 
> Soft Works:
> > Signed-off-by: softworkz <softworkz at hotmail.com>
> > ---
> >  tests/ref/fate/filter-overlay-dvdsub-2397 | 181 +++++++++++-------
> ----
> >  tests/ref/fate/sub-dvb                    | 162 ++++++++++--------
> -
> >  tests/ref/fate/sub2video                  | 178 ++++++++----------
> ---
> >  tests/ref/fate/sub2video_basic            |  93 ++++++-----
> >  tests/ref/fate/sub2video_time_limited     |   6 +-
> >  5 files changed, 293 insertions(+), 327 deletions(-)
> >
> 
> This is not how fate-updates are supposed to work: You should update
> the
> fate-references in the commit that makes the adjustment necessary.
> That
> way fate will (or rather: should) always be clean* (which is good for
> bisecting). Furthermore, you should explain why the changes to the
> references are necessary (i.e. what they actually mean). Increased
> locality helps with this, too.
> (This might of course lead to a bigger absolute diff, because you
> might
> have to change lines multiple times. But that is not what really
> matters.)
> 
> - Andreas
> 
> *: In other words, if the mail headers were sent correctly, patchwork
> would show lots of red.

Doesn't patchwork build once per patchset rather than once per commit?

Otherwise then, yes that makes sense, besides from the convention-based 
requirements anyway.

I think I'll also need to split the documentation updates across the 
individual filter commits, right?


My primary intention at this stage regarding fate tests was to 
demonstrate that only those five sub2video tests are affected by the
patchset but nothing else.

I'll fully address your suggestions in a subsequent update.

Thanks,
softworkz