[FFmpeg-devel] [PATCH]lavf/subviewerdec: Support higher sub-second precision

[Carl Eugen Hoyos]

Am Sa., 21. März 2020 um 18:35 Uhr schrieb Michael Niedermayer
<michael at niedermayer.cc>:
>
> On Fri, Mar 20, 2020 at 09:05:12AM +0100, Carl Eugen Hoyos wrote:
> > Am Fr., 20. März 2020 um 01:49 Uhr schrieb Michael Niedermayer
> > <michael at niedermayer.cc>:
> > >
> > > On Thu, Mar 19, 2020 at 10:18:33PM +0100, Carl Eugen Hoyos wrote:
> > > > Am Do., 19. März 2020 um 19:32 Uhr schrieb Michael Niedermayer
> > > > <michael at niedermayer.cc>:
> > > > >
> > > > > On Wed, Mar 18, 2020 at 02:16:30AM +0100, Carl Eugen Hoyos wrote:
> > > > > > Hi!
> > > > > >
> > > > > > Attached patch fixes ticket #8575 for me, Google describes such files.
> > > > > >
> > > > > > Please comment, Carl Eugen
> > > > >
> > > > > >  subviewerdec.c |   10 +++++++---
> > > > > >  1 file changed, 7 insertions(+), 3 deletions(-)
> > > > > > 1a6133ed3dbb66945d6bf16e212c0e77e3fd0579  0001-lavf-subviewerdec-Support-higher-sub-second-precisio.patch
> > > > > > From 39d0748782bb3e37fb2f92c679ffa58b239374c7 Mon Sep 17 00:00:00 2001
> > > > > > From: Carl Eugen Hoyos <ceffmpeg at gmail.com>
> > > > > > Date: Wed, 18 Mar 2020 02:11:33 +0100
> > > > > > Subject: [PATCH] lavf/subviewerdec: Support higher sub-second precision.
> > > > > >
> > > > > > Fixes ticket #8575.
> > > > > > ---
> > > > > >  libavformat/subviewerdec.c | 10 +++++++---
> > > > > >  1 file changed, 7 insertions(+), 3 deletions(-)
> > > > > >
> > > > > > diff --git a/libavformat/subviewerdec.c b/libavformat/subviewerdec.c
> > > > > > index 06b827b70f..ed48b3388d 100644
> > > > > > --- a/libavformat/subviewerdec.c
> > > > > > +++ b/libavformat/subviewerdec.c
> > > > > > @@ -56,11 +56,15 @@ static int read_ts(const char *s, int64_t *start, int *duration)
> > > > > >      int64_t end;
> > > > > >      int hh1, mm1, ss1, ms1;
> > > > > >      int hh2, mm2, ss2, ms2;
> > > > > > +    int multiplier = 1;
> > > > > >
> > > > > > +    if (sscanf(s, "%u:%u:%u.%2u,%u:%u:%u.%2u",
> > > > > > +               &hh1, &mm1, &ss1, &ms1, &hh2, &mm2, &ss2, &ms2) == 8)
> > > > > > +        multiplier = 10;
> > > > > >      if (sscanf(s, "%u:%u:%u.%u,%u:%u:%u.%u",
> > > > > >                 &hh1, &mm1, &ss1, &ms1, &hh2, &mm2, &ss2, &ms2) == 8) {
> > > > >
> > > > > which lengths can the ms1 / ms2 part have ?
> > > > > do all of the following work ?
> > > > > x.1
> > > > > x.19
> > > > > x.199
> > > > > x.1999
> > > > >
> > > > > or are some of these guranteed not to occur ?
> > > >
> > > > I have no idea but my suspicion is that whoever created the files
> > > > in question misunderstood the original format.
> > >
> > > then i would suggest that the code either supports all variants of this
> > > or checks if the input is one of the supported ones.
> > > basically checking how long the ms1/ms2 fields are IIUC
> >
> > New patch attached.
> >
> > Thank you, Carl Eugen
>
> >  subviewerdec.c |   16 +++++++++++++---
> >  1 file changed, 13 insertions(+), 3 deletions(-)
> > 09ba97edd94b332e899d1669626229cacab4fbbe  0001-lavf-subviewerdec-Support-higher-sub-second-precisio.patch
> > From 17574e5cb44dead6990051540ada103610ee1e5b Mon Sep 17 00:00:00 2001
> > From: Carl Eugen Hoyos <ceffmpeg at gmail.com>
> > Date: Fri, 20 Mar 2020 09:02:59 +0100
> > Subject: [PATCH] lavf/subviewerdec: Support higher sub-second precision.
> >
> > Fixes ticket #8575.
> > ---
> >  libavformat/subviewerdec.c | 16 +++++++++++++---
> >  1 file changed, 13 insertions(+), 3 deletions(-)
> >
> > diff --git a/libavformat/subviewerdec.c b/libavformat/subviewerdec.c
> > index 06b827b70f..83378eab5f 100644
> > --- a/libavformat/subviewerdec.c
> > +++ b/libavformat/subviewerdec.c
> > @@ -56,11 +56,21 @@ static int read_ts(const char *s, int64_t *start, int *duration)
> >      int64_t end;
> >      int hh1, mm1, ss1, ms1;
> >      int hh2, mm2, ss2, ms2;
> > +    int multiplier = 1;
> >
> > +    if (sscanf(s, "%u:%u:%u.%2u,%u:%u:%u.%2u",
> > +               &hh1, &mm1, &ss1, &ms1, &hh2, &mm2, &ss2, &ms2) == 8) {
> > +        multiplier = 10;
> > +    } else if (sscanf(s, "%u:%u:%u.%1u,%u:%u:%u.%1u",
> > +                      &hh1, &mm1, &ss1, &ms1, &hh2, &mm2, &ss2, &ms2) == 8) {
> > +        multiplier = 100;
> > +    }
> >      if (sscanf(s, "%u:%u:%u.%u,%u:%u:%u.%u",
> >                 &hh1, &mm1, &ss1, &ms1, &hh2, &mm2, &ss2, &ms2) == 8) {
> > -        end    = (hh2*3600LL + mm2*60LL + ss2) * 100LL + ms2;
> > -        *start = (hh1*3600LL + mm1*60LL + ss1) * 100LL + ms1;
>
> > +        ms1 = FFMIN(ms1, 999);
> > +        ms2 = FFMIN(ms2, 999);
>
> iam not sure this is the best solutiuon, no warning, no error just cliping

I believe this is a useful solution for subtitles, especially if no such samples
are known.

> except that, the patch is probably ok

Patch applied.

Thank you, Carl Eugen