1. 23 Feb, 2019 1 commit
  2. 19 Feb, 2019 1 commit
  3. 06 Nov, 2018 1 commit
  4. 11 Sep, 2018 1 commit
  5. 02 Jun, 2018 1 commit
  6. 24 Dec, 2017 1 commit
  7. 09 Jul, 2017 2 commits
  8. 25 Feb, 2017 2 commits
  9. 04 Jan, 2017 1 commit
  10. 06 May, 2016 1 commit
  11. 29 Apr, 2016 1 commit
  12. 06 Feb, 2015 1 commit
  13. 26 Sep, 2014 3 commits
    • Jim Nelson's avatar
      821c4b5a
    • Jim Nelson's avatar
      Reduce debug output of source pixbuf cache · 1c66a979
      Jim Nelson authored
      Previous commit introduced more debug logging than is really
      necessary.
      1c66a979
    • Jim Nelson's avatar
      More conservative source pixbuf caching: Bug #715198 · aad834d2
      Jim Nelson authored
      Prior source pixbuf (untransformed, full-sized decoded images) cache
      was per-photo, holding onto the source pixbuf for 180s.  For casual
      browsing it wasn't onerous, but if the photos were loaded quickly
      in succession (either with the photo viewer or, worse, via the
      programmatic thumbnail regenerator) memory usage would climb into the
      gigabytes.
      
      This uses a more conservative source pixbuf cache, holding only a few
      at a time and releasing them much more quickly.  The PixbufCache
      class does a better job of long-term pixbuf caching by holding onto
      scaled images rather than full-sized.
      aad834d2
  14. 09 Sep, 2014 1 commit
  15. 19 Aug, 2014 1 commit
    • Jim Nelson's avatar
      Don't crash on imported photo: Bug #734986 · 1491418f
      Jim Nelson authored
      The user's photo was identified as JFIF but is an incomplete file.
      The JFIF sniffer reported it could not identify the file, which is
      somewhat incorrect.  The RAW sniffer then tried to identify the file
      and crashed inside of libraw.
      
      Now when a sniffer can appropriately detect the file is corrupt it
      will be reported as such.  All GdkSniffers will do this if they (a)
      read the entire file and (b) no image was prepared.  Because they all
      use pre-condition checking to sanity check that the file is their
      format (via magic header bytes), this works.
      1491418f
  16. 05 Feb, 2014 1 commit
  17. 29 Oct, 2013 1 commit
  18. 19 Sep, 2013 1 commit
  19. 25 Apr, 2013 1 commit
  20. 16 Apr, 2013 1 commit
  21. 15 Apr, 2013 2 commits
  22. 26 Mar, 2013 1 commit
  23. 21 Mar, 2013 1 commit
  24. 27 Feb, 2013 1 commit
  25. 26 Feb, 2013 1 commit
    • Clint Rogers's avatar
      Instead of rotating the pixels and dimensions on export, if we're being asked... · 2c2bdfd9
      Clint Rogers authored
      Instead of rotating the pixels and dimensions on export, if we're being asked to write a JPEG, we write the pixels normally and let the orientation field do the rotation for us, fixes #1938.
      
      NOTE: this does _not_ fix the case where reverting after an external editor has set the _modified.jpg to top-left; this was broken out into its own bug.
      2c2bdfd9
  26. 21 Feb, 2013 1 commit
  27. 20 Feb, 2013 1 commit
  28. 08 Feb, 2013 1 commit
  29. 01 Feb, 2013 1 commit
  30. 28 Jan, 2013 1 commit
  31. 24 Jan, 2013 1 commit
  32. 23 Jan, 2013 1 commit
  33. 22 Jan, 2013 1 commit
  34. 21 Jan, 2013 1 commit
  35. 04 Jan, 2013 1 commit