GitLab

We don't deal with media art in tracker-extract since commit
6a050686 and ac84484d, mostly as it would require
us to weaken the tracker-extract sandbox.

Add testcase for cue sheet extractor

See merge request GNOME/tracker-miners!58

We recently discovered a long standing regression in this feature:
GNOME/tracker-miners#60

As the saying goes: "if it's not tested, it doesn't work"[1].

Depends on GNOME/tracker!85

1. https://yakking.branchable.com/posts/truism-4-if-it-is-not-tested/

tracker-extract-text: Correctly report errors back to the caller

Closes #56

See merge request GNOME/tracker-miners!62

The tracker_extract_get_metadata() function should return FALSE if
an error occurs reading the file. The tracker-extract-text module would
return TRUE in all cases.

This was causing intermittent failures in the functional-tests, as the
following situation could occur:

  1. file2.txt is created
  2. tracker-miner-fs sees file2.txt and processes it
  3. file2.txt is deleted (or moved into an unmonitored directory)
  4. tracker-miner-fs sees the deletion and removes its resource from
     the store
  5. tracker-extract sees the created notification for file2.txt and
     tries to process it
  6. the tracker_extract_get_metadata() incorrectly returns TRUE
     (success), so tracker-extract recreates the deleted resource

This problem was being detected in the functional tests and was
causing intermittent failures.

This hopefully fixes: GNOME/tracker-miners#56

This reverts commit 60145ddd.

Some projects that include Tracker haven't yet updated to Meson 0.50. This
change broke gnome-build-meta for example:
https://gitlab.gnome.org/GNOME/gnome-build-meta/-/jobs/286671

The install: arg isn't actually required as for configure_file() targets
Meson will infer whether an install is needed by the presence of the
install_dir: arg.

This reverts commit ad664049.

I accidentally pushed this to master without review, when I pushed the
trivial Meson change.

The merge request is here:
GNOME/tracker-miners!61

Meson helpfully warns us that this is needed:

    WARNING: Project specifies a minimum meson_version '>=0.47' but uses features which were added in newer versions:
     * 0.50.0: {'install arg in configure_file'}

If tracker-miners is configured with the `-Dtracker_core=subproject` option,
it should use GSettings schemas from the ./subprojects/tracker build
tree. We weren't doing that, which meant that if you ran `dnf remove
tracker` on your test image and then ran the test, you would see
failures like this in the functional-tests:

    (process:8450): GLib-GIO-ERROR **: 09:43:34.625: Settings schema 'org.freedesktop.Tracker.Store' is not installed

This fix is part of GNOME/tracker#24

tracker-extract: Use libz to process ps.gz files

Closes #61

See merge request GNOME/tracker-miners!60

Properly handle (atend) when parsing Postscript file. Also add a test
case.

Process ps.gz files using GZlibDecompressor instead of spawing gunzip. Because
tracker-extract runs the file parsers inside a seccomp sandbox, spawning an
external process during parsing is not a good idea because it leaves us with
little control on which syscalls are used.

Closes: GNOME/tracker-miners#61

Fix memory leaks/corruptions when running tests

Closes #59

See merge request GNOME/tracker-miners!54

tracker-extract: Relate contained audio tracks and container file both ways

Closes #60

See merge request GNOME/tracker-miners!57

This is the most generic class that still triggers GraphUpdated, we need
to delegate on the extractor module to figure out whether the file is a
music piece on its own or a container of music pieces.

In the latter case the role belongs on the contained elements, not on the
file resource itself, so avoid setting it here for all.

If there is a TOC, the nmm:* information belongs in the child nmm:MusicPiece
elements, not on the container file.

The migration to TrackerResource broke the handling of flac/cue, as
tracker_resource_print_*() do expect all related resources to be directly
referenced from the root resource, but the gstreamer extractor just added
backreferences from the nmm:MusicPieces to the parent nfo:FileDataObject.

Add a relation both ways through nie:hasLogicalPart/nie:isLogicalPartOf
so the graph is complete, and the tracks referenced from the root element.

Closes: GNOME/tracker-miners#60

Switch to README.md

See merge request GNOME/tracker-miners!56

Move tests to python3 and reformat with autopep8

See merge request GNOME/tracker-miners!55

Python has a recommended code style, known as PEP-8. The tests now follow this
style.

This commit was created using `autopep8` version 1.4.3, run with this command:

	find . -name "*.py" -exec autopep8 --ignore E501 --in-place {} \;

Closes: GNOME/tracker-miners#59

Closes: GNOME/tracker-miners#59

Closes: GNOME/tracker-miners#59

seccomp_release needs to be called both in case of success and in case of
failure. As per documentation any seccomp filters already loaded into the kernel
are not affected.

Closes: GNOME/tracker-miners#59

Closes: GNOME/tracker-miners#59

Running functional tests in an isolated environment result on gst
trying to build the registry again, and fork for it. This results
in a g_spawn call, and who knows why, in this debug line written to
stdout:

posix_spawn avoided (fd close requested)

This is a g_debug() call from glib which should not end up on screen
with our verbosity settings, so I'll asume someone else is playing
smart and replacing the default log handler with something that
introduces garbage on our stdout.

This notably causes JSON output in video tests to be not quite JSON,
causing test failures.

In ninja dist, build is checked with untouched prefix (which means
/usr/local), most certainly we're not going to find the stop-words
dir there, which triggers errors on tests.

Look up the stop-words dir consistently with the other things we
find through the tracker-sparql-2.0.pc file.

This is not the domain ontologies dir.

functional-tests: Fix failed tests being reported as successful

See merge request GNOME/tracker-miners!46

This code path would fail if two tests tried to create the
directory at the same time. We should ignore EEXIST rather than doing
a check + create sequence that isn't atomic.

This also fixes an issue in the functional-test runner script which caused
failed tests to appear as if they succeeded.

See GNOME/tracker-miners#44

Relevant issues:

  * GNOME/tracker-miners#55
  * GNOME/tracker-miners#56
  * GNOME/tracker-miners#57