1. 19 Nov, 2020 1 commit
  2. 12 Nov, 2020 1 commit
  3. 20 Oct, 2020 1 commit
  4. 14 Oct, 2020 1 commit
  5. 28 Jan, 2020 1 commit
  6. 16 Jan, 2020 2 commits
  7. 15 Jan, 2020 1 commit
    • krnlyng's avatar
      [glib2] detect removable storage properly. JB#48442 · 09e19642
      krnlyng authored
      Additionally to using mtab instead of fstab also check whether the mount
      is in /run/media/$USER, in order to detect user removable storage
      properly.
      
      This means tracker will also detect usb-otg devices and encrypted sd cards.
      09e19642
  8. 23 Jan, 2019 1 commit
  9. 14 Jan, 2019 1 commit
  10. 12 Jul, 2018 2 commits
  11. 09 Jul, 2018 1 commit
  12. 11 May, 2018 1 commit
    • Pekka Vuorela's avatar
      [glib2] Update to 2.56.1. Contributes to JB#41420 · 748f954b
      Pekka Vuorela authored
      - drop syslog patch, glib supports structured logging now
      - mtab patch could have been applied, but the rationale for inclusion
      in the first place is not available with this git repo. Just omitted in
      hope of getting rid of it. Upstream by default already moved to libmount,
      but that seems to have problems  with Tracker. Enabling commit
      (496cd8aac6402) does state that it returns slightly different results,
      namely with bind mounts.
      748f954b
  13. 14 Jul, 2016 2 commits
  14. 30 May, 2016 2 commits
  15. 26 May, 2016 1 commit
  16. 23 May, 2016 1 commit
  17. 31 Jan, 2016 1 commit
    • stskeeps's avatar
      Merge branch 'leak' into 'master' · ddd86bad
      stskeeps authored
      Plug memory leak
      
      Bumped into it while debugging mms-engine:
      ```
      ==14981== 72 bytes in 3 blocks are definitely lost in loss record 2,926 of 3,087
      ==14981==    at 0x483F380: malloc (vg_replace_malloc.c:296)
      ==14981==    by 0x515E083: g_malloc (gmem.c:104)
      ==14981==    by 0x51749B5: g_slice_alloc (gslice.c:1016)
      ==14981==    by 0x5191555: g_variant_alloc (gvariant-core.c:478)
      ==14981==    by 0x5191555: g_variant_new_from_children (gvariant-core.c:567)
      ==14981==    by 0x518F009: g_variant_builder_end (gvariant.c:3654)
      ==14981==    by 0x519083B: g_variant_valist_new (gvariant.c:4624)
      ==14981==    by 0x51909B7: g_variant_new_va (gvariant.c:4778)
      ==14981==    by 0x5190AD3: g_variant_new (gvariant.c:4717)
      ==14981==    by 0x111B3: org_nemomobile_mms_engine_complete_push (org.nemomobile.MmsEngine.c:2389)
      ==14981==    by 0x1491B: mms_engine_handle_push_notify (mms_engine.c:332)
      ==14981==    by 0x14A07: mms_engine_handle_push (mms_engine.c:359)
      ==14981==    by 0x5A4DA5B: ffi_call_VFP (in /usr/lib/libffi.so.5.0.10)
      ```
      The leak was introduced in commit aec92880
      
      See merge request !2
      ddd86bad
  18. 29 Jan, 2016 1 commit
  19. 04 Dec, 2015 1 commit
    • stskeeps's avatar
      Merge branch 'no-reply' into 'master' · 9f89b56b
      stskeeps authored
      Silence "Rejected send message" D-Bus errors
      
      This doesn't cover all the cases. For example, it doesn't cover `org.freedesktop.DBus.Introspectable.Introspect` calls that don't expect a reply; this is possible even though it doesn't make any practical sense. But it does cover the most common use case, i.e. a normal D-Bus call that has no reply.
      
      See merge request !1
      9f89b56b
  20. 03 Dec, 2015 1 commit
  21. 15 Sep, 2015 1 commit
  22. 14 Sep, 2015 1 commit
  23. 18 Jun, 2014 2 commits
    • iamer's avatar
      Merge pull request #1 from special/master · 8594609d
      iamer authored
      [packaging] Re-enable 'debug' checks in release builds
      8594609d
    • John Brooks's avatar
      [packaging] Re-enable 'debug' checks in release builds · a00f3b57
      John Brooks authored
      The original intent of using --enable-debug=no on release builds was to
      improve stability by removing overly-aggressive asserts and checks that
      could bring down important processes. This was misguided: I can't point
      to one crash that happens because of an assert/check that wouldn't
      otherwise crash, except when G_DEBUG is set in the environment.
      
      Also, many applications implicitly depend on the functionality removed
      by G_DISABLE_CHECKS to operate safely. Having that flag in our builds
      turns warnings into crashes and could harm stability.
      
      The only valid argument in favor of --enable-debug=no is performance,
      but the benefit there is unclear and we don't use glib in
      performance-sensitive paths. Other distributions seem to agree, and
      don't set this option.
      
      Note that the default is 'minimum', which includes casts, checks, and
      asserts. --enable-debug=yes does much more, but it's generally only
      enabled for unstable glib versions.
      a00f3b57
  24. 06 Jun, 2014 1 commit
  25. 11 May, 2014 1 commit
  26. 18 Mar, 2014 3 commits
  27. 09 Mar, 2014 1 commit
  28. 07 Feb, 2014 2 commits
  29. 06 Feb, 2014 3 commits