diff options
| author | Dudemanguy <random342@airmail.cc> | 2023-09-30 08:58:56 -0500 |
|---|---|---|
| committer | Dudemanguy <random342@airmail.cc> | 2023-09-30 09:02:57 -0500 |
| commit | 3a572c7a88329666a803bb8eb4de0a5a19e60cf8 (patch) | |
| tree | 48b8f9f4d218149a67c12f838a8f0989799e338f /sub/dec_sub.c | |
| parent | 84fa7ea411e8e99fced66295c973ecdf94ee07b2 (diff) | |
| download | mpv-3a572c7a88329666a803bb8eb4de0a5a19e60cf8.tar.bz2 mpv-3a572c7a88329666a803bb8eb4de0a5a19e60cf8.tar.xz | |
Revert "demux: improve stream selection state"
The stream selection state wasn't improved. I didn't realize this messed
with caches. All in all, just not a good idea. Back to drawing board I
guess.
This reverts commit f40bbfec4fcd2d9a787a4d98ec7698a646e5607e.
Diffstat (limited to 'sub/dec_sub.c')
| -rw-r--r-- | sub/dec_sub.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/sub/dec_sub.c b/sub/dec_sub.c index 84c6b7bfbe..dc26aa027e 100644 --- a/sub/dec_sub.c +++ b/sub/dec_sub.c @@ -269,7 +269,7 @@ static bool is_new_segment(struct dec_sub *sub, struct demux_packet *p) // Read packets from the demuxer stream passed to sub_create(). Return true if // enough packets were read, false if the player should wait until the demuxer // signals new packets available (and then should retry). -bool sub_read_packets(struct dec_sub *sub, double video_pts) +bool sub_read_packets(struct dec_sub *sub, double video_pts, bool force) { bool r = true; pthread_mutex_lock(&sub->lock); @@ -291,7 +291,7 @@ bool sub_read_packets(struct dec_sub *sub, double video_pts) break; // (Use this mechanism only if sub_delay matters to avoid corner cases.) - double min_pts = sub->opts->sub_delay < 0 ? video_pts : MP_NOPTS_VALUE; + double min_pts = sub->opts->sub_delay < 0 || force ? video_pts : MP_NOPTS_VALUE; struct demux_packet *pkt; int st = demux_read_packet_async_until(sub->sh, min_pts, &pkt); |