1. 24 Sep, 2020 1 commit
  2. 23 Sep, 2020 6 commits
  3. 22 Sep, 2020 4 commits
  4. 21 Sep, 2020 3 commits
  5. 18 Sep, 2020 2 commits
  6. 15 Sep, 2020 2 commits
    • Ryan Gonzalez's avatar
      trash portal: Handle portal failures · 88032cb5
      Ryan Gonzalez authored
      Previously, only whether or not the D-Bus call itself succeeded was
      checked, regardless of the result code from the actual trash operation.
    • Simon McVittie's avatar
      glocalfile: Never require G_LOCAL_FILE_STAT_FIELD_ATIME · 642baa50
      Simon McVittie authored
      Some filesystems don't have meaningful access times under at least some
      circumstances (see #2189, #2205). In this situation the traditional stat()
      and related kernel interfaces have to put something meaningless in the
      st_atime field, and have no way to signal that it is meaningless.
      However, statx() does have a way to signal that the atime is meaningless:
      if the filesystem doesn't provide a useful access time, it will unset
      the STATX_ATIME bit (as well as filling in the same meaningless value
      for the stx_atime field that stat() would have used, for compatibility).
      We don't actually *need* the atime, so never include it in the required
      mask. This was already done for one code path in commit 6fc143bb
      "gio: Allow no atime from statx" to fix #2189, but other callers were
      left unchanged in that commit, and receive the same change here.
      It is not actually guaranteed that *any* of the flags in the
      returned stx_mask will be set (the only guarantee is that items in
      STATX_BASIC_STATS have at least a harmless compatibility value, even if
      their corresponding flag is cleared), so it might be better to follow
      this up by removing the concept of the required mask entirely. However,
      as of Linux 5.8 it looks as though STATX_ATIME is the only flag in
      STATX_BASIC_STATS that might be cleared in practice, so this simpler
      change fixes the immediate regression.
      Resolves: #2205Signed-off-by: Simon McVittie's avatarSimon McVittie <smcv@collabora.com>
  7. 14 Sep, 2020 5 commits
  8. 13 Sep, 2020 1 commit
  9. 11 Sep, 2020 2 commits
  10. 10 Sep, 2020 7 commits
  11. 09 Sep, 2020 1 commit
  12. 08 Sep, 2020 4 commits
  13. 07 Sep, 2020 2 commits