- 16 Apr, 2013 1 commit
-
-
Hyejin Kim authored
- check the ringtone type - according to sound type, select the title Change-Id: I3c0962565aff41fa6add5c2ca55bbff42ddf326b
-
- 05 Mar, 2013 1 commit
-
-
Hyoseong Kim authored
Collator.getInstance() returns new Collator object. But, new Collator object is not used in function. Change-Id: Ie6ef4ecb5de8e47ac797aebed58fe5c8a0693cc0
-
- 28 Feb, 2013 1 commit
-
-
Daniel Bonnevier authored
DATETIME was set as primary key but that value is not always unique. This solution is to remove primary key constraint from column "time" to avoid SQLiteConstraintException caused in case same DATETIME value is attempted to be inserted into log table. Change-Id: I06f1cd36a6f4f7becdb444599e4de7866813c9cd
-
- 08 Feb, 2013 1 commit
-
-
Hiroaki Kuriyama authored
When choosing a ringtone for alarm use by "Android System" ringtone picker (RingtonePickerActivity), default ringtone is not played for preview with volume of alarm stream type but with that of normal stream type. It is because stream type is not reflected on mDefaultRingtone instance of RingtonePickerActivity. Change-Id: Idc71c9b620e40e4109604b2337677134bb74a8cd
-
- 07 Feb, 2013 1 commit
-
-
Oscar Rydhé authored
When a client asks the Media Provider to rearrange a playlist the notifyChange callback function is called before the database transaction is ended resulting in that listeners can't get the new data when they get the callback. notifyChange() is called before the database transaction is ended resulting in that listeners can't get the new data when they get the callback. This fix will wait until the database transaction is finished to do the callback. Change-Id: Ifdfbb047b809bd5fba4752bcdd299927f9082eee
-
- 08 Nov, 2012 1 commit
-
-
guoyin.chen authored
MediaProvider will fall back to the image in the same folder of music, if MetadataRetriver failed to extract the AlbumArt image from those mp3 files. MediaProvider will only bypass the albumArt image search if the music located in the root director of the primary external storage. Add this bypass for all external storages. Change-Id: I972586b2f92bd439f51ee932f09e824fc5507236 Signed-off-by:
guoyin.chen <guoyin.chen@freescale.com>
-
- 22 Sep, 2012 1 commit
-
-
Jeff Sharkey authored
Bug: 7205962 Change-Id: Ie0b28501e25758094f14736ac47f6e4dd3a90be5
-
- 19 Sep, 2012 2 commits
-
-
Mike Lockwood authored
Bug: 6949231 Change-Id: Ib301b250c40a6abc3cab17658e19308951b4cd4a Signed-off-by:
Mike Lockwood <lockwood@google.com>
-
Jeff Sharkey authored
Bug: 7174672 Change-Id: Ib44a52ad337ca6df135955e71e6cec3261da780d
-
- 18 Sep, 2012 2 commits
-
-
Jeff Sharkey authored
Watch for BOOT_COMPLETED, and dispatch any current USB state or mounted volumes which may have already occured. This specifically handles starting of secondary users. The current MTP kernel driver at /dev/mtp_usb is exclusive, meaning only one process can have it open. To solve this, the framework cycles the USB host stack when switching users, giving the new user's media process a chance to claim the kernel device. We only start an MtpServer when we're the active user. Bug: 6925114 Change-Id: Idfcda09aed88140bb470a110a9e4434f5b79abdd
-
Mike Lockwood authored
Bug: 6987701 Change-Id: I55d2f95ab63df19f9c9ba39da5215893b8e658e4 Signed-off-by:
Mike Lockwood <lockwood@google.com>
-
- 12 Sep, 2012 2 commits
-
-
Jeff Sharkey authored
Change-Id: I55a93be5e1edd0f00c318d07363242388936aa63
-
Jeff Sharkey authored
When building external storage paths, always include user in path to enable cross-user paths and aid debugging. Bug: 7131382 Change-Id: I85f9eeed418a4fcced6633c579642e947fba65d7
-
- 22 Aug, 2012 1 commit
-
-
Chuanxiao Dong authored
Using single storage solution means its primary storage is physical SD card. So, it is possible for user to insert a non-Vfat SD card to make primary storage not mounted. So, before using primary storage in PTP mode, let's have a check first Change-Id: I3932c16399675232d8b2329c8e30cf6030b3586e Author: Chuanxiao Dong <chuanxiao.dong@intel.com> Signed-off-by:
Xiaokang Qin <xiaokang.qin@intel.com> Signed-off-by:
Bruce Beare <bruce.j.beare@intel.com> Signed-off-by:
Jack Ren <jack.ren@intel.com> Author-tracking-BZ: 30725
-
- 20 Aug, 2012 1 commit
-
-
Ruslan Shymkevych authored
In some situations, when not handled exceptions occurred, wakelock was not released. Now wakelock is released always. Change-Id: I356cdfbcbfb9594e4577d8b2c9f54d5560158d48 Signed-off-by:
Andrii Beregovenko <a.beregovenko@ti.com>
-
- 13 Aug, 2012 1 commit
-
-
Marco Nelissen authored
Increment delete/update counters in a few places I missed, and add an extra log for unmounted storage to help diagnose b/6094111 Change-Id: I8118f67bb6114526dfe2839055ff447937266436
-
- 25 Jul, 2012 1 commit
-
-
Marco Nelissen authored
b/6700936 b/6809256 b/6778795 Change-Id: Id6fe3bea44884433f6255199d85fa8a1431e44ea
-
- 23 Jul, 2012 1 commit
-
-
Marco Nelissen authored
This should prevent key uniqueness issues like in b/6700936 Change-Id: Ie0416024f39b2e2c4a84512adede55b4a49755d4
-
- 19 Jun, 2012 1 commit
-
-
Dave Burke authored
Bug: 6613558 Change-Id: I29f5c01b283ab34291bee740926229b61aa56f64
-
- 23 May, 2012 5 commits
-
-
Jeff Sharkey authored
Bug: 6543820 Change-Id: I215921f9bc2196ec16df40bacdc249b5c079254a
-
Fabrice Di Meglio authored
Fix bug #6522190 MountService should respond to configuration changes ("INTERNAL STORAGE" string should be translated dynamically) - use an ID instead of a String for StorageVolume description - use this ID for getting the correct localized version of the description string Change-Id: I7f1b1cf6c953442e3c47e54ed4ffbb4498f4cf78
-
Jeff Sharkey authored
Since external storage is case-insensitive, check if we have an exact match on internal storage before opening directly. When it doesn't match exactly, we fall back to letting the sdcard daemon resolve the path. Bug: 6544070 Change-Id: I6e6094f4b639ac70feae6549648c1e486480ed71
-
Mike Lockwood authored
This will allow recomputing the volume description in case the locale has changed Bug: 6522190 Change-Id: Icc055ae3404841582285c0f2a153d5ee10ebcc6d
-
Jeff Sharkey authored
For the special case where (1) a file lives on external storage, (2) it's being opened for reading, and (3) external storage is emulated, this change rewrites the path at the last moment to open the file directly from internal storage. This avoids the overhead of forcing I/O through the emulated storage layer when we've already asserted the caller has permission. This is only possible because MediaProvider is running with the media_rw GID. Bug: 6539384 Change-Id: Ieda6b30e819340d46466f8db2e34b47bbf06ee08
-
- 22 May, 2012 1 commit
-
-
Marco Nelissen authored
When doing a "like" match on a path, add a second non-like constraint so that sqlite wildcard characters don't match arbitrary other characters. b/6501408 Change-Id: I3912dd70b842ccba2652385a7733965b6691d303
-
- 15 May, 2012 2 commits
-
-
Marco Nelissen authored
Small optimization to bucket field updater: reuse the same ContentValues and not needlessly convert to/from integer. Also add a mechanism to store short messages in the database, in order to get better information from bugreports. b/6467624 Change-Id: I2777d58e5fefe98a0e91794ef25e1be0d2408304
-
Marco Nelissen authored
Restrict updating the "bucket" columns to just images and videos, instead of all files. b/6344917 Change-Id: I93bb236556220a05dffc633219302cda43facbd7
-
- 14 May, 2012 2 commits
-
-
Marco Nelissen authored
The media provider was filling the parent-cache with bogus values during a media scan, causing all files to appear in the MTP root instead of in their proper locations. b/6487442 Change-Id: I3dacd6c69072be13e75ddf59662b16682e806d68
-
Mike Kasick authored
Programs may request to media scan files (via MediaScannerConnection::scanFile) before external storage is mounted. If that happens, MediaProvider creates an external-ffffffff.db (volume ID -1) database once external storage is mounted, instead of the one with the expected volume ID. Later, if android.process.media is killed and respawned, the real external database (that is, the one corresponding to the actual external storage volume ID) is attached. At best, that database contains stale records. However, in conjunction with another bug (where the LRU database garbage collector prunes all but one external database), most often a new, empty database is created. This results in the sudden disappearance of all ringtones and media from programs that depend on the media database. Change-Id: If4e23b52cd09a640e26852916eef53710c27d933
-
- 07 May, 2012 1 commit
-
-
Marco Nelissen authored
MediaProvider would not set the scan start time in some cases, which would then make it skip some dumpsys output. This fixes the problem by ensuring that the scan start time is set when attaching to the volume. b/6456260 Change-Id: I99f25ff5f2dd44cf92b08bbdf8fb015cd062e958
-
- 03 May, 2012 1 commit
-
-
Marco Nelissen authored
b/5702487 Change-Id: Ie0392b60a6c258dc378ac5d2bfe73aa7f9b47c4e
-
- 02 May, 2012 1 commit
-
-
Jeff Sharkey authored
Since _DATA can be supplied by external apps, ensure they hold any required permissions before opening. Bug: 6131916 Change-Id: I7dd4e4d1e1a1a2fc690a2d0f5e9f432451e73832
-
- 25 Apr, 2012 1 commit
-
-
Jeff Sharkey authored
To give RINGTONE_PICKER external storage access, move it from system to MediaProvider. Bug: 6346701 Change-Id: Ib5f8e8fa8a962be211bc60c6e09778c7e2b85f2a
-
- 19 Apr, 2012 1 commit
-
-
Marco Nelissen authored
When re-retrieving album art that had previously been deleted, a NullPointerException might occur. b/6359924 Change-Id: Icb1aa1bd52583025893e5b8cc7a73c47a4bfe8f3
-
- 18 Apr, 2012 1 commit
-
-
Marco Nelissen authored
b/6293272 Change-Id: I1b4288b5637452cc19101e78ed32e5c50330afe2
-
- 16 Apr, 2012 1 commit
-
-
Owen Lin authored
Change-Id: I82f90b0762a6b3342e7a84ec40b205e5b9377905 fix: 6326240
-
- 13 Apr, 2012 1 commit
-
-
Marco Nelissen authored
The internal "sd card" storage is now in /storage/sdcard0. However since /mnt/sdcard still exists for compatibility, the media scanner will see a bunch of new files in /storage, in addition to the existing files in /mnt/sdcard, resulting in two entries in the database for every file. This change removes any entries that were already scanned in /storage, then updates the paths for the existing /mnt/sdcard entries to point to /storage instead. This will handle both updates from ICS, as well as fix the database for people that ran a build inbetween the move to /storage and this change. b/6131916 Change-Id: I1c071187651544be0fdb69e1cdb6367b636fc6d5
-
- 06 Apr, 2012 1 commit
-
-
Marco Nelissen authored
If an item was removed from a playlist, the play_order field of the following items would not be adjusted. Because of a bug in the code for reordering items in playlists, this would prevent reordering of items after the deleted item, and could confuse apps that appended more items to the playlist after that, causing them to insert with duplicate play_order values. This change both makes it safe to have gaps in the play_order sequence, and ensures that when deleting items, the following items have their play_order adjusted downward so as to not leave any gaps if there weren't any before. b/6299084 Change-Id: I0709fa7167790a86b966059184e15b97da2ed74f
-
- 05 Apr, 2012 1 commit
-
-
Marco Nelissen authored
b/6293320 Change-Id: I11297ce78ede748b37d7f6c2cec361bc5f837d3b
-
- 04 Apr, 2012 1 commit
-
-
Marco Nelissen authored
b/5820120 Change-Id: Ied554557627dcc29577c316ad30e4f5c3dd8e7f8
-