[FFmpeg-devel] [PATCH 3/4] fftools/opt_common: stop accessing a private field

[Anton Khirnov]

Quoting James Almer (2024-10-14 22:48:12)
> On 10/14/2024 1:27 PM, Alexander Strasser via ffmpeg-devel wrote:
> > On 2024-10-14 17:52 +0200, Michael Niedermayer wrote:
> >> On Mon, Oct 14, 2024 at 01:36:46PM +0200, Anton Khirnov wrote:
> >>> ---
> >>>   fftools/opt_common.c | 4 +---
> >>>   1 file changed, 1 insertion(+), 3 deletions(-)
> >>>
> >>> diff --git a/fftools/opt_common.c b/fftools/opt_common.c
> >>> index 021ed75272..34da2cee7d 100644
> >>> --- a/fftools/opt_common.c
> >>> +++ b/fftools/opt_common.c
> >>> @@ -808,7 +808,6 @@ int show_filters(void *optctx, const char *opt, const char *arg)
> >>>       printf("Filters:\n"
> >>>              "  T.. = Timeline support\n"
> >>>              "  .S. = Slice threading\n"
> >>> -           "  ..C = Command support\n"
> >>>              "  A = Audio input/output\n"
> >>>              "  V = Video input/output\n"
> >>>              "  N = Dynamic number and/or type of input/output\n"
> >>> @@ -833,10 +832,9 @@ int show_filters(void *optctx, const char *opt, const char *arg)
> >>>                                     ( i && (filter->flags & AVFILTER_FLAG_DYNAMIC_OUTPUTS))) ? 'N' : '|';
> >>>           }
> >>>           *descr_cur = 0;
> >>> -        printf(" %c%c%c %-17s %-10s %s\n",
> >>> +        printf(" %c%c %-17s %-10s %s\n",
> >>>                  filter->flags & AVFILTER_FLAG_SUPPORT_TIMELINE ? 'T' : '.',
> >>>                  filter->flags & AVFILTER_FLAG_SLICE_THREADS    ? 'S' : '.',
> >>> -               filter->process_command                        ? 'C' : '.',
> >>>                  filter->name, descr, filter->description);
> >>>       }
> >>>   #else
> >>
> >> The commit message is not describing this change accurately
> >>
> >> Its not just "not accessing a priavte field", it removes information
> >> from the printed filter list
> > 
> > Learning question: How can we see this is a private field?
> > 
> > This information was useful. How can we bring it back if we decide it's
> > not OK too test if this field is NULL?
> 
> A new flag AVFILTER_FLAG_SUPPORT_COMMANDS that the user can check to 
> ensure a call to avfilter_process_command() will not return ENOSYS could 
> be added. And of course, to print the C in here.

avfilter_process_command() may or may not return ENOSYS whether that
flag is set or not, so I don't see what exactly would it be useful for.
Same holds for printing the capability in fftools - just what is the
user expected to do with that information?

-- 
Anton Khirnov