| *** jrt is now known as Guest86183 | 02:05 | |
| Nico-old-defunct | How do I figure out on what file the tracker meta data extraction process chokes? | 10:13 |
|---|---|---|
| Thaodan | there is a command line tool for tracker that could show it | 11:25 |
| Thaodan | tracker status? | 11:26 |
| Nico-old-defunct | I played around with that a lot, but I only figured out how to show the status and restart the deamon, if it crashes, not what files it is extracting metadata from | 11:26 |
| Thaodan | dbus could help I think | 11:37 |
| Thaodan | use gdbus | 11:37 |
| Nico-old-defunct | I'll try, I think I saw the dbus name somewhere | 11:41 |
| inz | You could also try running tracker-extract from command line with env var G_MESSAGES_DEBUG=all | 11:45 |
| Nico-old-defunct | That sound easier, I'll try that, thanks! | 11:45 |
| Nico-old-defunct | Hm, maybe it is supposed to exit, since it returns 0 | 11:49 |
| Nico-old-defunct | But my indexing is still stuck at 75% | 11:49 |
| inz | Ah, apparently it's changed since last I've debugged it, sorry for the bad clue. | 11:52 |
| Nico-old-defunct | Nah, it helped me alread, now I at least found some dbus names :D | 11:53 |
| Nico-old-defunct | Monitoring the Miner process, it goes to here and never seems to make progress after that: /org/freedesktop/Tracker1/Miner/Files: org.freedesktop.Tracker1.Miner.Progress ('Processing…', 0.75, 17) | 11:57 |
| Nico-old-defunct | Not sure what other interfaces I could monitor? | 11:57 |
| Nico-old-defunct | Okay, when it fails, I get 2 errors: | 12:15 |
| Nico-old-defunct | (tracker-miner-fs:28882): Tracker-CRITICAL **: 14:14:17.472: Could not execute sparql: GDBus.Error:org.freedesktop.Tracker1.SparqlError.Internal: Not a ISO 8601 date string. Allowed form is [-]CCYY-MM-DDThh:mm:ss[Z|(+|-)hh:mm] | 12:15 |
| Nico-old-defunct | (tracker-miner-fs:28882): GLib-GIO-CRITICAL **: 14:14:17.473: g_file_hash: assertion 'G_IS_FILE (file)' failed | 12:15 |
| Nico-old-defunct | Any idea, how to track down, where that is coming from? | 12:15 |
| Nico-old-defunct | Hm, I'll try to strace it | 12:18 |
| Nico-old-defunct | Okay, I got the failing sparql query | 12:38 |
| Nico-old-defunct | So, which Jolla person do I bother, if I know where my issue is, but don't know where the code is? | 12:46 |
| Nico-old-defunct | Basically this: https://forum.sailfishos.org/t/index-for-media-files-documents-not-updating-xperia10-ii-4-1-0-23/6328/10?u=kuroneko | 12:47 |
| inz | Wow, that is quite a ways up in the future xD | 12:48 |
| inz | The analysis on the forum post (and the error message) is wrong, milliseconds are not the problem | 12:50 |
| Nico-old-defunct | Oh, lol, I didn't even see that xD | 12:50 |
| inz | The timestamp 292276973 years in the future is the problem | 12:50 |
| *** ersatzmaus is now known as fledermaus | 12:51 | |
| Nico-old-defunct | Yep | 12:51 |
| Nico-old-defunct | I just looked at the format and saw that milliseconds aren't allowed | 12:51 |
| Nico-old-defunct | So it could also choke on that, but the timestamp is just bad in general :D | 12:52 |
| inz | Millis are in the code | 12:53 |
| inz | But year is limited to 4 digits | 12:53 |
| Nico-old-defunct | Mhm | 12:54 |
| Nico-old-defunct | Well, I added some context | 12:54 |
| Nico-old-defunct | modification timestamp from a stat call is actually the dawn of time, 1970 | 12:54 |
| Nico-old-defunct | (context to the forum post) | 12:55 |
| inz | The error is from https://git.sailfishos.org/mirror/tracker-miners/blob/master/src/libtracker-miners-common/tracker-date-time.c#L72 | 12:55 |
| Nico-old-defunct | Also, where does the following format allow milliseconds? [-]CCYY-MM-DDThh:mm:ss[Z|(+|-)hh:mm] | 12:55 |
| inz | As I said, the error is wrong | 12:56 |
| Nico-old-defunct | Ah | 12:56 |
| Nico-old-defunct | Now I get it, my bad | 12:56 |
| inz | On that line the (\\.[0-9]+)? handles millis | 12:56 |
| Nico-old-defunct | Yep | 12:56 |
| inz | INT64_MAX/365.25/24/60/60-30 is 292271023015 | 12:57 |
| inz | Which sounds pretty close to what the forum post has | 12:58 |
| Nico-old-defunct | Yeah | 12:58 |
| Nico-old-defunct | I think there is just no date, since stat return 1970 | 12:58 |
| Nico-old-defunct | The failing INSERT DATA call is here: https://git.sailfishos.org/mirror/tracker-miners/blob/master/src/tracker-extract/tracker-extract-decorator.c#L183 | 13:02 |
| Nico-old-defunct | (It matches up with the line breaks correctly) | 13:02 |
| Nico-old-defunct | I think | 13:02 |
| Nico-old-defunct | Actually, maybe not :D | 13:03 |
| Nico-old-defunct | So if the file has no modification date, it gets set to 0: https://git.sailfishos.org/mirror/tracker-miners/blob/master/src/miners/fs/tracker-miner-files.c#L2085 | 13:11 |
| Nico-old-defunct | Which probably causes issues with the formatting function? | 13:12 |
| Nico-old-defunct | Hm, that's a relatively new change? https://git.sailfishos.org/mirror/tracker-miners/commit/c6beebb1c55d5f9582b1e332683e787ca96ea349 | 13:14 |
| Nico-old-defunct | Now that I think about it, if that function returns a date equivalent to -1 on error, that would be fine on 32bit, but it would break on 64bit. Since any 32bit year should still only have 4 digits. | 13:18 |
| rinigus | who I am supposed to ping if I want to remove an app from Jolla Store? "OSM Scout Server Module: Route" is not used anymore for a while and should be removed | 14:01 |
| rinigus | ... let's see if it works - asked to remove the app as a part of "new" submission | 14:02 |
| *** frinring_ is now known as frinring | 14:18 | |
| Mister_Magister | how do i sync target with qtcreator? | 15:13 |
| Mister_Magister | ahh just package-install even tho package was installed synces them | 15:14 |
| Mister_Magister | but it would be nice to have separate command | 15:14 |
| bionade24 | TEst | 20:35 |
| Nico-old-defunct | Test failed | 20:36 |
| *** nyov is now known as Guest60655 | 22:48 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!