• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

Eyevinn / hi264 / 38416922312
89%

Build:
DEFAULT BRANCH: main
Ran 13 Oct 2026 02:51PM UTC
Jobs 1
Files 53
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

13 Oct 2026 02:50PM UTC coverage: 88.378% (+0.1%) from 88.255%
38416922312

push

github

tobbee
fix: continue a stream with the parameter sets of its last slice

A stream can carry several SPSs and PPSs. One SPS is active per coded video sequence, each picture uses a PPS that refers to it, and the IDR that starts a sequence may switch to another SPS (H.264 7.4.1.2.1). P_Skip continuations instead took whichever parameter sets came last or first: AppendPSkipFrames the last SPS and the last PPS in the stream, PSkipExtender those it was created with, NewPSkipExtenderFromDecConfRec and hi264-mp4-extend the first SPS and first PPS of the avcC, even when that PPS refers to another SPS. LastFrameState checked pic_order_cnt_type of the last SPS. With the wrong SPS, the slice header gets the wrong frame_num and pic_order_cnt_lsb lengths and wrap points, and hi264-mp4-extend encoded its black IDR at the wrong picture size.

PSkipExtender now takes the PPS of each observed slice and the SPS that PPS refers to, and generates with those, including MaxFrameNum and MaxPicOrderCntLsb. NextSlice reports an SPS it cannot generate for. ParameterSets returns the pair, and hi264-mp4-extend uses it for the picture size, pic_timing and the black IDR. AppendPSkipFrames lets the extender learn the parameter sets from the stream. NewPSkipExtender rejects a PPS that refers to another SPS. NewPSkipExtenderFromDecConfRec and hi264-mp4-extend start from the first PPS and the SPS it refers to; the former now rejects a record whose parameter sets do not all parse, since it needs every SPS before it can pair the PPS. LastFrameState and tools/extend_pskip use the SPS of the last slice.

The tests use a stream whose slices use SPS 1 and PPS 1 while SPS 0 and PPS 0 come last, with frame_num wrapping at 256 and 16 respectively. All of them fail on main.

59 of 72 new or added lines in 4 files covered. (81.94%)

2 existing lines in 2 files now uncovered.

8753 of 9904 relevant lines covered (88.38%)

1.02 hits per line

Uncovered Changes

Lines Coverage ∆ File
6
73.89
-0.21% cmd/hi264-mp4-extend/main.go
6
90.91
3.32% pkg/encode/pskip_extender.go
1
84.0
0.0% tools/extend_pskip/main.go

Coverage Regressions

Lines Coverage ∆ File
1
73.89
-0.21% cmd/hi264-mp4-extend/main.go
1
90.91
3.32% pkg/encode/pskip_extender.go
Jobs
ID Job ID Ran Files Coverage
1 38416922312.1 13 Oct 2026 02:51PM UTC 53
88.38
GitHub Action Run
Source Files on build 38416922312
  • Tree
  • List 53
  • Changed 4
  • Source Changed 0
  • Coverage Changed 4
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • d840841c on github
  • Prev Build on main (#37945161666)
  • Next Build on main (#38418580268)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc