After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 715194 - Multi-pass indexing
Multi-pass indexing
Status: RESOLVED DUPLICATE of bug 719802
Product: tracker
Classification: Core
Component: Miners
unspecified
Other Linux
: Normal enhancement
: ---
Assigned To: tracker-general
Depends on:
Blocks:
 
 
Reported: 2013-11-25 21:13 UTC by Xavier Claessens
Modified: 2013-12-03 20:59 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Xavier Claessens 2013-11-25 21:13:29 UTC
AFAIK, currently when tracker's miners crawl the file system, it extract metadata out of files (possibly using gstreamer, etc) which takes time on resource-constraint devices. To get faster result, I suggest first crawling files to record them, and extra metadata later. I don't know yet how possible that is with current tracker architecture.
Comment 1 Xavier Claessens 2013-11-28 22:22:43 UTC
Clarification: It is true that it will make the overall indexing slower, but the use case here is similar than in bug #719306: When I plug an USB key and open the music app, I want my music to be displayed as soon as possible even if it has incomplete metadata. The UI can refresh later when tracker extract more information.

So what I suggest (if I understand correctly the code):

1) add a dbus method SkipMetaDataExtractor(bool) that music app can call when it is displaying music list to user. That tells tracker that it must go as fast as possible to crawl all files, and extract more metadata later. Note that dbus API could be in common with bug #719306.

2) in tracker-miner-files.c, process_file_cb(), skip the tracker_extract_client_get_metadata() call if an app has called SkipMetaDataExtractor. It will probably have to add a "MetaDataExtracted" boolean property on the nie:InformationElement object to false in that case.

3) When tracker is done crawling all files (not sure where that happens in the code) it should query "select %url where {?o nie:url ?url;x:MetaDataExtracted false}" and re-index them.
Comment 2 Xavier Claessens 2013-12-03 20:59:52 UTC
I think this bug has lots in common with bug #719306 , so I opened bug #719802
to handle both cases.

*** This bug has been marked as a duplicate of bug 719802 ***