| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The media art storage spec
https://wiki.gnome.org/MediaArtStorageSpec#Identifiers specifies the
following possible values for prefix: 'album', 'artist', 'podcast',
'radio', 'track'.
This commit adds support for all of these except 'track', which will
currently raise a g_warning() if used to advise the caller that it is
not supported.
Rather than only supporting a fixed list of identifiers, any unknown
prefix is accepted and handled in the same way as 'artist', 'radio' and
'podcast' media art.
This commit also fixes
https://bugzilla.gnome.org/show_bug.cgi?id=705834 where the path for
'album' media art where artist == NULL was calculated incorrectly.
|
|
|
|
|
| |
The call to media_art_plugin_init() was modified as it had the
wrong arguments.
|
| |
|
| |
|
|
|
|
|
|
| |
This function is similar to the current media_art_process(), but
taking a GFile, mainly added for convenience in applications
already using GIO.
|
|
|
|
|
|
|
| |
This call is similar to media_art_get_path(), but using GFiles
for in/out path parameters. This older function now internally
uses the just added one, so internally operations happen on
GFiles.
|
| |
|
| |
|
| |
|
| |
|
|\
| |
| | |
Add TrackerStorage and fix license file
|
| |
| |
| |
| |
| |
| |
| | |
Autoconf adds COPYING which contains GPLv3, which is not what the
sources are licensed with. Replace it with GPLv2 in a text file
from GNU's website and add COPYING.LESSER for LGPLv2.1, so that the
package would be under LGPLv2.1 license.
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| | |
This will initialize a TrackerStorage on init and unrefs it on shutdown.
Removed warnings for missing TrackerStorage. Uncommented
media_art_copy_to_local which now works again.
|
| |
| |
| |
| | |
These libraries are needed by TrackerStorage.
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
The timeout priority is now lower so ::mount-unmount has a chance to
be processed before the watchdog if the miner's main loop is busy
(say removing monitors). Also, the timeout is longer so we aren't
overzealous on slow unmounts.
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
Fixes NB#248873. If, for whatever reason, an unmount operation fails
or doesn't respond in a timely fashion, TrackerStorage will emit
::mount-point-added again so Tracker keeps monitoring any further
change in there.
|
| | |
|
| |
| |
| |
| | |
Lionel Landwerling <lionel.g.landwerlin@linux.intel.com>
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fixes GB#640845
We don't want to rely on the g_drive_is_media_removable() method
because it does not tell us whether a device can be disconnected from
the system but rather if a device contains a media that might be
extracted from it. In fact, this method maps the removable flag from
the kernel block device subsystem.
If we rely on g_drive_is_media_removable(), most of the USB harddrives
are considered as non removable, and are therefor won't be indexed.
This patch proposes to check whether or not the mount point is part of
the system, and if it's not, we use g_volume_can_mount() method to
check whether the filesystem can be mounted which gives us a better
clue about whether the related device is removable or not (in the way
tracker considers a device from being removable).
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@linux.intel.com>
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
mountpoints' contents
This function will detect content type properly for the different
optical media types, ensuring a NULL content type just means it
contains data. The is_optical check is still done aside, based
on the content type, or the unix device.
Fixes bug #642014 - Tracker cannot index optical data disc's data,
reported by Simon Hong.
|
| | |
|
| |
| |
| |
| | |
Cleaned up several warnings during the build and missing documentation too
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
* TRACKER_STORAGE_TYPE_IS_REMOVABLE
* TRACKER_STORAGE_TYPE_IS_OPTICAL
|
| | |
|
| |
| |
| |
| | |
to do so
|
| |
| |
| |
| |
| | |
- Added information about WHERE we got the code from - GIO
- Added reasoning for why we have the is_multimedia check
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Previously we used content type guessing, this isn't full proof as I
discovered during testing this branch. Now we use the same method GIO
uses by looking at the filesystem type/device/mount point (in that
order).
Seems to work for all the formats tested here for me:
- Compaq flash (EOS Canon card)
- USB stick
- CD (audio)
- CD (game)
- DVD (movie)
- MMC (SD)
- USB hard disk
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| | |
* Using MD5 of the mount name for the UUID of optical media
* Using MD5 of the mount path for the UUID of GMounts without
a corresponding GVolume
|
| |
| |
| |
| | |
(as not all GMounts have a corresponding GVolume)
|
| |
| |
| |
| |
| |
| | |
* If the mount point detected doesn't setup a GVolume (as it
is in fstab), then just force a re-check of the directories
indexed to see if anything changed.
|
| |
| |
| |
| |
| | |
Some major sections like tracker-password-provider, tracker-storage
and tracker-thumbnailer had no section description.
|