1. 02 Feb, 2022 2 commits
  2. 28 Jan, 2022 1 commit
  3. 17 Jan, 2022 1 commit
    • Célestin Marot's avatar
      multifilesrc: fix caps leak · f2638a48
      Célestin Marot authored and Tim-Philipp Müller's avatar Tim-Philipp Müller committed
      since `gst_caps_replace()` and `gst_pad_set_caps()` both ref the caps and neither of them takes the ownership of the caps -> it must be unreffed in `gst_multi_file_src_set_property()`
      
      to test the leak (on Unix): `echo coucou > /tmp/file.txt && GST_TRACERS=leaks GST_DEBUG="GST_TRACER:7" gst-launch-1.0 multifilesrc location=/tmp/file.txt caps='txt' ! fakesink`
      
      Part-of: <!1109>
      f2638a48
  4. 20 Nov, 2021 3 commits
  5. 05 Nov, 2021 1 commit
  6. 29 Oct, 2021 2 commits
  7. 08 Sep, 2021 2 commits
  8. 07 Sep, 2021 1 commit
  9. 06 Sep, 2021 5 commits
  10. 02 Sep, 2021 1 commit
  11. 28 Aug, 2021 1 commit
  12. 24 Aug, 2021 2 commits
  13. 23 Aug, 2021 1 commit
  14. 15 Aug, 2021 1 commit
    • Sebastian Dröge's avatar
      souphttpsrc: Always use the content decoder but set `Accept-Encoding:... · f484e87c
      Sebastian Dröge authored and Tim-Philipp Müller's avatar Tim-Philipp Müller committed
      souphttpsrc: Always use the content decoder but set `Accept-Encoding: identity` if no compression should be used
      
      Some servers respond with gzip-encoded responses regardless of whether
      the request allowed it to be used in the response. By always having the
      content decoder enabled, these invalid responses can be decoded
      correctly while for well-behaving servers the `compress` property
      selects between allowing compressed responses or not.
      
      Fixes #833
      
      Part-of: <!1054>
      f484e87c
  15. 08 Jul, 2021 3 commits
  16. 11 Jun, 2021 1 commit
  17. 03 Jun, 2021 1 commit
  18. 02 Jun, 2021 1 commit
    • Tim-Philipp Müller's avatar
      rtpjpegpay: fix image corruption when compiled with MSVC on Windows · 41c100bc
      Tim-Philipp Müller authored
      On Windows with MSVC, jpeg_header_size would end up 2 bytes larger
      than it should be. This then leads to the first 2 bytes of the
      actual jpeg image data to be dropped, because we think those
      belong to the header, which results in an undecodable image when
      reconstructed in the depayloader.
      
      What happens is that when the compiler evaluates
      
        jpeg_header_size = mem.offset + read_u16_and_inc_offset_by_2(&mem);
      
      it actually uses the mem.offset value after it has been increased
      by the function call on the right hand size of the equation.
      
      From section 6.5 of the C99 spec:
      
        3. The grouping of operators and operands is indicated by the syntax [74].
           Except as specified later (for the function-call (), &&, ||, ?:, and
           comma operators), the order of evaluation of subexpressions and the
           order in which side effects take place are both unspecified.
      
      Fixes #889
      
      Part-of: <!1003>
      41c100bc
  19. 01 Jun, 2021 5 commits
  20. 14 May, 2021 3 commits
  21. 30 Apr, 2021 2 commits
    • Jan Schmidt's avatar
      qtmux: Make sure to write 64-bit STCO table when needed. · 218f924b
      Jan Schmidt authored and GStreamer Marge Bot's avatar GStreamer Marge Bot committed
      qtmux attempts to choose between writing a 32-bit stco chunk offset table
      when it can, but switch to a 64-bit co64 table when file offsets go over
      4GB.
      
      This patch fixes a problem where the atom handling code was checking
      mdat-relative offsets instead of the final file offset (computed by
      adding the mdat position plus the mdat-relative offset) - leading to
      problems where files with a size between 4GB and 4GB+offset-of-the-mdat
      would write incorrect STCO tables with some samples having truncated
      32-bit offsets.
      
      Smaller files write STCO correctly, larger files would switch to
      co64 and also output correctly.
      
      Part-of: <!972>
      218f924b
    • Qi Hou's avatar
      v4l2object: Add interlace-mode back to caps for camera · ea1a6c47
      Qi Hou authored and Tim-Philipp Müller's avatar Tim-Philipp Müller committed
      skip_try_fmt_probes is set to TRUE for v4l2src to skip interlace-mode and
      colorimetry when probe caps. gst_v4l2_object_set_format_full() will add
      colorimetry back to caps when iterating over the negotiated caps. There is
      one case that v4l2src is first in preview state then starts recording.
      v4l2src caps will change with an additional interlace-mode structure after
      renegotiation, then v4l2src needs to reset. But this camera driver can't
      orphan buffer pool, it causes require buffer failed as streaming is still
      in active state.
      
      To fix this, also need to add interlace-mode back to caps for camera to
      avoid reset.
      
      Part-of: <!971>
      ea1a6c47