~vcs-imports/tracker/+git/tracker:wip/ernestask/43

Last commit made on 2018-10-06
Get this branch:
git clone -b wip/ernestask/43 https://git.launchpad.net/~vcs-imports/tracker/+git/tracker

Branch merges

Branch information

Name:
wip/ernestask/43
Repository:
lp:~vcs-imports/tracker/+git/tracker

Recent commits

1d2da6b... by Display Name

libtracker-sparql: Check for null pointers in finalize()

tracker_notifier_finalize() may accidentally stumble upon a null pointer
or two, throwing warnings and making Nautilus CI pipelines fail.

Related: https://gitlab.gnome.org/GNOME/tracker/issues/43

982fe13... by Carlos Garnacho

libtracker-direct: Honor tracker_sparql_connection_local_new() API docs

At some point we started warning on NULL journal/ontology arguments,
even though those are documented as having sensible defaults if NULL
is provided.

b14f45d... by Carlos Garnacho

functional-tests: Remove process watch before killing process

This sometimes causes spurious errors in 12-transactions because
the GSource gets activated somewhere between the process being
restarted, causing failures and logs like:

RuntimeError: tracker-store exited with status: -9

dbff7ef... by Carlos Garnacho

tests: Update tracker-fts-test to using a local TrackerSparqlConnection

Instead of using internals directly. Also make sure the database location
dir is a brand new temporary directory each time.

571c9c3... by Carlos Garnacho

nepomuk: Drop unused ontology file

These NAO ontology extensions were split from the main ontology prior
to 2.0 because they are both non-official and unnecessary for core
ontology concepts.

However, it was apparently never added to build, so this piece of
ontology did effectively disappear. Internally it was just used by the
evolution plugin (which disappeared in 2.0 as well), and since there
has been no outcry it seems mostly unused on the outside. It just doesn't
make sense to mourn for this overly generic ontology.

So drop this file, and just rescue nao:predefined-tag-favorite as
this is still used by apps (it just doesn't happen to be a nao:Tag).

7c0c2e1... by Carlos Garnacho

tests: Do not compare/assert expected errors' text in tracker-sparql-test

While testing for error existence makes sense, expecting specific error
messages makes tests non forward-compatible to possible changes.

ac722f3... by Carlos Garnacho

libtracker-data: Remove get_wal_interface() function

There is no longer need to expose the interface used for deferred WAL
checkpoints, just don't expose this internal API.

202c73c... by Carlos Garnacho

libtracker-data: Move WAL checkpointing to TrackerDBManager

The WAL logging and manual checkpoint handling is defined in the deepmost
layers, it doesn't make sense to handle WAL checkpoints in the highest ones.

Besides, the GThread for the deferred checkpoint is no longer leaked, and
we ensure that deferred and blocking checkpoints are mutually exclusive.

46043ac... by Carlos Garnacho

libtracker-direct: Ignore error on passive WAL checkpoint

The sqlite3_wal_checkpoint_v2() docs don't seem very clear on
whether SQLITE_BUSY may be returned on passive checkpoints (In
fact, I seem to read the opposite).

But that seems to happen in practice. Since a passive checkpoint
is failable in essence, just ignore the error.

df01997... by Carlos Garnacho

Merge branch 'wip/piotrdrag/kill-intltool'