2016-06-21 Carlos Garnacho <carlosg@gnome.org>
Release 1.9.0
libtracker-data: Ensure variables are defined in aggregate expressions
Look those up and error out if they are undefined.
https://bugzilla.gnome.org/show_bug.cgi?id=759361
fts: Do not apply stop-words when tokenizing query search terms
FTS5 notifies of the purpose in tokenization in their xTokenize vfunc,
check for the FTS5_TOKENIZE_QUERY flag indicating that this is the
tokenization of the search terms in a query and do not apply the stop
words list in this case.
One example where this is potentially harmful are "search as you type"
UIs. eg. typing the word "ontology" would have you type the word "onto",
which is an ignored word. Only after typing the next character you'd get
matches, which seems irregular behavior.
2016-06-16 Rafael Fontenelle <rafaelff@gnome.org>
Updated Brazilian Portuguese translation
2016-06-14 Sam Thursfield <ssssam@gmail.com>
configure.ac: Remove some unused variables
2016-06-14 Sam Thursfield <sam@afuera.me.uk>
functional-tests: Increase resource wait timeout to 30 seconds
Tests in the 300-miner-basic-ops function tests were failing for me,
which turned out to be just because the extractor was slower than the
timeout in certain cases. It would be good to work out why it's slow,
but let's start by fixing the random test failures.
2016-06-13 Sam Thursfield <sam@afuera.me.uk>
functional-tests: Fix crash in tests with multiple testsuites
For some reason the GLib main loop calls a timeout even after we've
removed it, in some cases, leading to errors like this:
[tracker-store] ready.
test_journal_01 (__main__.JournalReplayTest) ...
Traceback (most recent call last):
File "/home/shared/gnome/src/tracker/tests/functional-tests/common/utils/helpers.py", line 110, in _process_watch_cb
status = self.process.poll ()
AttributeError: 'NoneType' object has no attribute 'poll'
All this does is workaround that problem.
tracker-store: Avoid duplicate timeout removal
I kept seeing warnings like this in the 200-backup-restore
functional-test:
test_backup_01 (__main__.BackupRestoreTest) ...
(tracker-store:8104): GLib-CRITICAL **: Source ID 25 was not found when attempting to remove it
test_backup_02 (__main__.BackupRestoreTest) ...
(tracker-store:8104): GLib-CRITICAL **: Source ID 25 was not found when attempting to remove it
functional-tests: Decode query results as UTF-8 in the correct way
Previously the code would try to cast a 'str' instance to 'unicode',
which fails because it assumes the 'str' instance is valid ASCII (i.e.
all characters are in the range 0-127).
This fixes 2 test case failures that looked like this:
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 1: ordinal not in range(128)
functional-tests: Add missing GLib import
Fixes an ImportError.
2016-06-12 Dušan Kazik <prescott66@gmail.com>
Updated Slovak translation
2016-06-11 Sam Thursfield <sam@afuera.me.uk>
Tweak documentation of tracker_decorator_info_get_task()
Mention what the task's data will be. Readers might not realise that
the documentation of tracker_decorator_info_get_sparql() holds that
information (I didn't).
Fix warning introduced in 0a92485009f08cb2872325d9c6627343134fbcc9
2016-06-09 Sam Thursfield <sam@afuera.me.uk>
utils: Add create-tree-from-real-data script
This script grabs a mix files from your system and creates a link farm,
with the aim of creating a tree of Tracker test data that exercises each
of the extract modules.
tests: Give tracker-test.c a slightly better name
Silence libmediaart log messages
This was removed in commit 2454185dd8a8459c45e4d32e0c3602f14647bbdd, but
it is actually still needed to hide log messages from libmediaart when
running `tracker-extract --verbosity=0`.
The mistake of reusing the same variable to store both log handler IDs
is now fixed.
tracker-extract: Improve console output for when extraction fails
Old output:
tracker_mimetype_info_get_module: assertion 'info != NULL' failed
No metadata or extractor modules found to handle this file
New output:
file:///uri/of/file: No metadata or extractor modules found to handle this file
The message is now also printed to stderr rather than stdout, so that
stdout contains only the metadata itself in whatever output format the
caller chose.
cli: Fix message when `tracker index` is called with no mode arg
If you run `tracker index` with no arguments you get the help output.
However, if you ran `tracker index ~/myfile` before you would see this:
(tracker index:28627): Tracker-WARNING **: (tracker-index.c:369):index_run: code should not be reached
Now you get a more helpful message:
Use `tracker index --file` when giving a specific file or directory
to index. See `tracker help index` for more information.
We could probably just assume --file if no mode is passed, but let's do
that separately. To be honest, `tracker index` should probably only take
a file or a mimetype, and the --backup, --restore and --import modes
should become commands in their own right.
2016-06-06 Daniel Mustieles <daniel.mustieles@gmail.com>
Updated Spanish translation
2016-06-05 Carlos Garnacho <carlosg@gnome.org>
tracker: Don't unref manager in the loop in "tracker index -f" subcommand
Must be unref'ed after the loop. Fixes crash if >1 files are provided.
https://bugzilla.gnome.org/show_bug.cgi?id=767024
2016-06-05 Martin Kampas <martin.kampas@tieto.com>
miner-fs: Fix handling files moved soon after creating
Noticed this when executing functional tests for write-back:
(tracker-miner-fs:21288): Tracker-CRITICAL **: Could not execute sparql:
Subject `(null)' is not in domain `nfo:FileDataObject' of property
`nfo:fileName'
This warning happens in item_move() when the source just didn't have
time to be indexed. One example:
copy ("file.txt", "temp_XYZ.file.txt")
- received G_FILE_MONITOR_EVENT_CREATED ("temp.file.txt")
- received G_FILE_MONITOR_EVENT_CHANGED ("temp.file.txt")
- received G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT ("temp.file.txt")
modify ("temp_XYZ.file.txt")
- received G_FILE_MONITOR_EVENT_CHANGED ("temp.file.txt")
- received G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT ("temp.file.txt")
mv ("temp_XYZ.file.txt", "file.txt")
- received G_FILE_MONITOR_EVENT_MOVED ("temp.file.txt", "file.txt")
- emitted ITEM_MOVED ("temp.file.txt", "file.txt")
It was already handled in item_move() in past, but removed with eef0e7f
(libtracker-miner: Remove useless code) after previously misidentified
as useless in scope of ee58e67 (libtracker-miner: Add compat layer for
tracker_miner_fs_directory_*)
The comment from ee58e67 """FIXME: This situation shouldn't happen from
a TrackerFileNotifier event""" simply cannot be satisfied: no way to get
"temp.file.txt" indexed before ITEM_MOVED is processed - the file
disappears too fast.
https://bugzilla.gnome.org/show_bug.cgi?id=678986
2016-06-05 Carlos Garnacho <carlosg@gnome.org>
libtracker-miner: cater for unbound nfo:belongsToContainer when moving items
Fixes warnings when moving indexing roots around. This query expects this
property to be bound, resulting in no-op if that's not the case (e.g.
indexing roots), later reinsertions of nie:url and other properties with
max cardinality=1 trigger the whole update failure, because those weren't
properly removed.
2016-06-04 Carlos Garnacho <carlosg@gnome.org>
libtracker-miner: Insert into the right graph when moving items
The query to update parent-dependent data was using the source file
urn as a graph urn.
2016-05-31 Carlos Garnacho <carlosg@gnome.org>
libtracker-common: Remove unused variable
libtracker-data: Implement fn:upper-case/UCASE
UCASE() was implemented over a nonexisting function, fix that. Impls
for both libunistring and libicu have been added.
2016-05-30 Carlos Garnacho <carlosg@gnome.org>
tests: Add some BIND() tests
Revert "libtracker-data: Relate Variable objects to their origin Context"
Not needed anymore.
This reverts commit 5c0fe132eb7d0e193ccc669447e015f8f9e8b3d6.
libtracker-data: Refine BIND() variable mapping
The implementation of BIND() introduced an special case in the way we
construct the result columns in selects (Handled in commit 64b54a7784).
This special case however turns out wrong when the resolved variable
is used both in and outside the BIND() statement.
So, admit the nature of this hack and special case it altogether, the
check was too iffy when there is really one situation where this makes
sense.
https://bugzilla.gnome.org/show_bug.cgi?id=766505
2016-05-30 Bastien Nocera <hadess@hadess.net>
tracker-extract: Consider DjVu multipage docs as Ebooks
ddb79e3 and cf7f5df tried to categorise DjVu files as paginated
documents, which would have slotted them next to PDF files in
gnome-documents.
But we'd like them to be near EPubs in gnome-books. As the EBook RDF
type is not too strict in its definition, tagging the DjVu multi-page
files as Ebooks is the easy way to fix it.
https://bugzilla.gnome.org/show_bug.cgi?id=766781
2016-05-30 Cédric Valmary <cvalmary@yahoo.fr>
Updated Occitan translation
2016-05-28 Carlos Garnacho <carlosg@gnome.org>
tests: Fix distcheck
Missing data files in EXTRA_DIST
docs: Do not delete xml directory when redoing docs
We're competing with gtk-doc.make here, which may turn out in
failures when doing mkdir on an already created directory, or
failures in writing xml/gtkdocentities.ent because we deleted
the directory under its feet.
So just do the same, use MKDIR_P to ensure it won't fail if
the directory is already there, and don't bother deleting it
as it's handled in gtk-doc.make. The ontology docs helper app
will rewrite the xml files anyway.
tracker-miner-fs: Ignore .desktop and .directory files
Those are responsibility of tracker-miner-applications.
tracker: Add -f argument to "tracker reset" CLI subcommand
It takes a file to be reset, works recursively for directories.
Immediately after reset, a reindex will be requested, so the data
is promptly indexed again.
tracker-miner-fs: Trigger reindex of files/directories on IndexFiles
If the file was already contained in an indexed directory, the IndexFile
was ignored. Actually cause a reindexing so we can trigger rechecks after
deleting the data.
tracker-miner-fs: Replace direct emission of TrackerIndexingTree signals
We now have tracker_indexing_tree_notify_update().
libtracker-miner: Add tracker_indexing_tree_notify_update()
This function notifies about updates in either an indexing root,
or a child of it. The typical reaction will be reindexing the
element.
This will replace the several ad-hoc emissions of
::directory-updated over the place.
libtracker-miner: Reindex elements on TrackerIndexingTree::child-updated
Directories will be crawled according to the root directory flags, regular
files will be notified upon right away.
libtracker-miner: Add TrackerIndexingTree::child-updated signal
So we can notify on something else than indexing roots. May be
used for the IndexFile request.
2016-05-27 Bastien Nocera <hadess@hadess.net>
tracker-extract: Install DjVu rule
https://bugzilla.gnome.org/show_bug.cgi?id=754479#c5
2016-05-25 Friedrich Herbst <fherbst@opmbx.org>
Updated German translation
2016-05-22 Carlos Garnacho <carlosg@gnome.org>
libtracker-common: Actually ignore XDG dirs set to $HOME
If an XDG dir is set to $HOME, we do ignore it (So it's not inadvertently
indexed recursively). However we just made the expansion to the real path
fail, the alias (e.g. &VIDEOS) was still dealt with as a path, and
attempted to be indexed.
So now we always return TRUE if the alias has a mapping, but only return
the string if the path is to our liking. This makes the alias truly
ignored.
2016-05-21 Carlos Garnacho <carlosg@gnome.org>
tracker-extract-mp3: Use performer as fallback to album artist
Commit 79af6adf6 changed the MP3 extractor to use the TPE2 tag info
to retrieve nmm:albumArtist. It is fine to favor that data, but we
should still resort to TPE1 if we get nothing from there. Otherwise
the album artist is not filled.
2016-05-21 Tiago Santos <tiagofsantos81@sapo.pt>
Updated Portuguese translation
2016-05-16 Daniel Mustieles <daniel.mustieles@gmail.com>
Updated Spanish translation
2016-05-16 Simon McVittie <simon.mcvittie@collabora.co.uk>
Add systemd user services corresponding to D-Bus session services
The one exception is the email miner, which does not actually have
anything to start, so its D-Bus service file won't work anyway.
Restart the services on failure, but do not after SIGKILL,
because that's what `tracker daemon --kill` (formerly
`tracker-control --kill`) does. systemd already treats SIGTERM
as a graceful exit which is not considered to be a failure.
Loosely based on a patch by Alban Crequy, which only applied
this to tracker-store and tracker-miner-fs, and did not have
the special handling for SIGKILL.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=755739
Reviewed-by: Carlos Garnacho
2016-05-14 Carlos Garnacho <carlosg@gnome.org>
tracker-miner-fs: Add MS Office "owner files" to ignored-files
"Owner files" (described in https://support.microsoft.com/en-us/kb/211632)
works as a lock file when opening the document. Those are supposedly
removed when the document is closed, but in practice there's situations
where those are left behind, and are meaningless to tracker.
Those files may be generated for any of the document extensions, so handle
the whole bunch.
https://bugzilla.gnome.org/show_bug.cgi?id=764739
manpages: Drop manpages of configuration settings/files
Gsettings are self-documenting, plus translatable. There is no need to
duplicate this documentation, and even less if it focuses on the
deprecated keyfile format. It is time to hide that under the rug.
tests: Tear down test data in tracker-steroids tests
2016-05-14 Bastien Nocera <hadess@hadess.net>
tracker-extract: Add RDF type for DjVu files
https://bugzilla.gnome.org/show_bug.cgi?id=754479
2016-05-14 Julien Hémono <whackvault@me.com>
IPTC extractors : use nco:contributor for contact metadata
5a27279a7e07a9f1f0c58f663f4c290c397d4aff adds support for contact
IPTC metadata and represents it with a nco:representative predicate.
This contradicts the ontology that stipulates that the subject of
a nco:representative predicate is a nco:Contact. Consequently, updates
from extraction of files containing this IPTC field are rejected by
the store.
This fix uses the nco:contributor predicate instead. Official
documentation at [1] says the field "Identifies the person or
organisation which can provide further background information on
the objectdata." Nco:contributor is therefore too narrow as not every
such person need be a contributor. The adequate predicate doesn't
exist but if it did it would definitely at least be a nao:annotation
or more narrowly a nao:isRelated. But I deem these too broad because
their range isn't a person or an organisation.
https://bugzilla.gnome.org/show_bug.cgi?id=762675
2016-05-10 Cédric Valmary <cvalmary@yahoo.fr>
Updated Occitan translation
2016-05-10 Antoine Jacoutot <ajacoutot@gnome.org>
tracker_file_system_get_remaining_space: use proper units
The fields of type fsblkcnt_t must be reported in units of f_frsize for
statvfs() and f_bsize for statfs().
https://bugzilla.gnome.org/show_bug.cgi?id=766137
2016-05-08 Carlos Garnacho <carlosg@gnome.org>
tracker-extract-gstreamer: Prevent warning if there is no TOC
Warning introduced in commit 99ab02a7. We should only attempt this
TOC when there is one from gstreamer.
2016-05-08 sahaltim9 <fabrice.louveau_ext@softathome.com>
configure: Update configure for sqlite3 support 3.7.15
sqlite3_errstr was added in that version.
https://bugzilla.gnome.org/show_bug.cgi?id=743245
2016-05-08 Carlos Garnacho <carlosg@gnome.org>
libtracker-common: Remove libmediaart log handler code
This is no longer part of tracker, so doesn't need doing. Also
fixes warnings on exit because this unused log handler overwrote
the variable storing the ID for the tracker handlers.
libtracker-miner: Give up on detailed decorator error logging
It was the hope that this error logging would lead to useful
but reports that helped qualitatively improve the extractor
modules soon enough.
But people suck, and instead we got lots of bugs about how
much we spam syslog (as if we used the "warning" logging level
just because) or how we violated their privacy, without further
thinking that there might be reasons for those warnings in the
first place (i.e. bugs that need fixing). Those bug reports
are usually plagued with half-cut,half-edited snippets from
syslog to prove their point, while at the same time invalidating
their usefulness.
So, change of approach. Now we just include the URN/URI in the
error warning, and point people to using TRACKER_VERBOSITY=3 in
order to get more information. The debug information will contain
the full unedited sparql, and the relevant info will be surrounded
by dotted lines so it is made clear the information that is useful.
This will require an extra step for the people trying to be helpful,
which hopefully those are willing to take. It will also hopefully
calm the bunch of complains, and will give me no hesitation to close
any further one we get as invalid.
configure: Set better bugzilla link
We can point to the tracker product.
libtracker-data: mark most sqlite functions as SQLITE_DETERMINISTIC
This flag helps the query planner in optimizing functions that are
guaranteed to return the same result given the same input. This
is the case for all functions, but SparqlRand().
2016-05-08 Kevin Haller <kevin.haller@outofbits.com>
tests: Test cases for fn:replace() with regular expression support.
Add five test cases, which shall check the behaviour of the
replace function.
https://bugzilla.gnome.org/show_bug.cgi?id=754961
libtracker-data: Support regular expressions for fn:replace().
Extends the sqlite database by a new function (with the name
SparqlReplace). The function makes use of the g_regex_replace() function
of glib.
To fullfill the XPath 2.0 standard some constraints must be checked for
fn:replace(input, pattern, replacement, flags). The given pattern must
not match a zero-length string. The given replacement string have to use
$ followed by a number for backreferences. If the dollar sign shall be used
"as is", it must be escaped (\$).
For checking and interpreting the given replacement string of fn:replace()
some regular expressions are needed. This expressions are precompiled and
saved in the function_regex hashset of the TrackerDBInterface. The
pre-compilation and initialization of the hashset are done by the
prepare_database() method.
The glib method g_regex_replace() make use of the backslash followed by a
number to inidcate backreferences. So the dollar signs must be interpreted
- the backslashes can be still used for this purpose.
In the sparql expression class the corresponding section is adapted, so
that the new SparqlReplace function is used for fn:replace(..) statements.
https://bugzilla.gnome.org/show_bug.cgi?id=754961
2016-05-08 Felix Riemann <friemann@gnome.org>
tracker-extract-gif: Avoid possible integer overflow
Fix integer overflow when skipping over the decoded image data
of extremely large or specifically prepared images.
https://bugzilla.gnome.org/show_bug.cgi?id=758315
2016-05-07 Gaurav Narula <gnarula94@gmail.com>
extractor: Store albumArtist from TPE2 tag
https://bugzilla.gnome.org/show_bug.cgi?id=764143
2016-05-07 Carlos Garnacho <carlosg@gnome.org>
tracker-extract-gstreamer: Use TOC as "cue sheet" as a last resort
The CUESHEET block in flac files is only meant to allow applications
to figure out CDDB references and whatnot, so it will basically contain
track durations and ISRCs. Other information like title/performer/genre
is not meant to be stored here:
https://xiph.org/flac/faq.html#general__no_cuesheet_tags
The other methods supported here could be considered non-standard,
but will be definitely more complete than this CUESHEET block, so
the lookup order is:
- Embedded cue file as a "CUESHEET=" vorbis comment
- External cue file
- This CUESHEET block
2016-05-07 Sam Thursfield <sam@afuera.me.uk>
tracker-extract: Make `tracker-extract --file` silent when --verbosity=0
libtracker-common: Fix Make rule for enums code
This was depending on some non-existent NOINST_H_FILES variable,
rather than the actual tracker-enums.h file.
2016-05-05 Carlos Garnacho <carlosg@gnome.org>
tracker-extract-mp3: Prevent possible crashes on 0-len comments
Files may contain a COMM frame that turns out to have a 0-len
string, in our calculations the length comes out as -1, which turns
out very wrong.
https://bugzilla.gnome.org/show_bug.cgi?id=764924
libtracker-miner: Do not free the current root data automatically on interruption
Most interruption cases are global to a RootData (eg. unmounts). It however
may turn out with unexpected consequences when requesting crawling of a
subfolder that is being deleted. In that case we'd expect the RootData to
be preserved and crawling to resume on the remaining folders.
So, only delete the current root data across regular operations as long
as we're not being interrupted, and ensure we do so manually in all the
cases where tracker_crawler_stop() is manually called.
tracker: Make it superscary to reset databases
This is possibly a data loss situation, so warn the user that this is
possibly what will happen.
tracker: Add comment for translators
tracker-preferences: Remove the "System" tab altogether
Resetting data is not something we should offer a big button for,
really.
tracker-preferences: Do not recommend reindex/restarts
This is an absolute antipattern, we can honor most of those settings
at runtime (and as of the last commits, we do).
The remaining configuration options are the ones affecting the tokenizer,
and we just need a tracker-store/session restart, so just suggest that,
instead of restarting daemons.
tracker-miner-fs: Trigger full recheck on enable-monitors changes
The crawling will recursively add/remove the monitors as requested.
libtracker-miner: Check deleted files on TrackerIndexingTree:directory-updated
Add the TRACKER_DIRECTORY_FLAG_CHECK_DELETED flag, so deleted contents are
checked on all folders.
libtracker-miner: Check for content filters being triggered on deletion checks
Store temporarily whether a folder triggered a filter based on its content
so we can perform the deletion of previously indexed files in that folder.
libtracker-miner: Add TRACKER_DIRECTORY_FLAG_CHECK_DELETED flag
This forces a check for deleted contents on all folders, as if
the folder mtime changed. Will be useful to recheck indexed directories'
contents after configuration changes.
libtracker-miner: Check for filtered out files during deletion checks
When checking the deleted contents of a folder, also find out whether
those are currently filtered out. Those must be deleted as well.
libtracker-miner: Add indexing roots to the tree only when those are crawled
We must not insert information for indexing roots prior to their processing.
This could be the case of indexing roots that are descendants of another
recursive root (eg. ~/Music in ~/).
This was inocuous when crawling the indexed folders from scratch, although
could make indexing roots being spuriously deleted/reindexed when folders
are reindexed (eg. because of TrackerIndexingTree::directory-updated signals)
libtracker-miner: Do not recurse up the hierarchy for out-of-tree index requests
If we receive a tracker_miner_fs_check_file() request for a file out of
indexing trees, it'd usually end up recursing until it ran out of parents
(that is, up to file:///). This is quite pointless, if only one file was
requested to be indexed.
libtracker-miner: Avoid storing mtime on when querying the iri ad-hoc
Retrieve only the iri from the query, and store it in the TrackerFileSystem,
the mtime is unrequested, and even misleading in future operations.
2016-05-03 Carlos Garnacho <carlosg@gnome.org>
libtracker-miner: Do not insert RootData for the same file multiple times
If the file is already there, there is no need to crawl/recheck it
several times.
libtracker-miner: Empty query_files cache when crawling the next folder
If, for the usual reasons, a directory is ignored after crawling, we
must remove files accumulated till that point. Those are pointless when
processing the next folder.
libtracker-miner: Ensure stored mtime data is freed after use
This data is not needed after comparing store and filesystem mtimes,
and keeping it stored could mislead later checks happening on that
same folder (say, directories being recursively rechecked because of
configuration changes).
It is safer to just dispose the data after use, the data might need
retrieving again at the time such recheck is needed.
libtracker-miner: Add tracker_file_system_steal_property()
This call retrieves and removes data stored in the TrackerFileSystem,
the caller must free the retrieved data.
libtracker-miner: TrackerIndexingTree should not return shallow roots
Those are created for internal maintenance, and should not leak outside.
2016-04-28 GNOME Translation Robot <gnome-sysadmin@gnome.org>
Updated Portuguese translation
2016-04-25 Timm Bäder <mail@baedert.org>
miner-manager: Plug memory leak
https://bugzilla.gnome.org/show_bug.cgi?id=765172
2016-04-07 Sam Thursfield <sam@afuera.me.uk>
tracker-extract: Fix error handling in tracker-extract-persistance
A couple of GCC warnings highlighted the fact that in case of error,
this code would pass uninitialised string pointers to g_critical().
2016-04-02 Andrew Stormont <astormont@racktopsystems.com>
Fix build on illumos.
https://bugzilla.gnome.org/show_bug.cgi?id=761177
2016-04-02 Carlos Garnacho <carlosg@gnome.org>
autogen.sh: Remove build time dependency on gnome-common
Follow the recommendations in
https://wiki.gnome.org/Projects/GnomeCommon/Migration
https://bugzilla.gnome.org/show_bug.cgi?id=763119
2016-03-27 Antoine Jacoutot <ajacoutot@gnome.org>
Adapt to new GKqueue monitor naming.
2016-03-27 Carlos Garnacho <carlosg@gnome.org>
configure: Check that sqlite3 has sqlite3_auto_extension() enabled
Only do this if we need to load the FTS5 module, sqlite3 might have
been compiled with SQLITE_OMIT_LOAD_EXTENSION, which will make things
go very wrong (poking NULL vfuncs in a 0'ed out sqlite3_api_routines)
at runtime.
This facility must be enabled if we need to load our FTS module, so
bail out at configure time if it's not there.
2016-03-27 Sam Thursfield <ssssam@gmail.com>
configure.ac: Add note about Automake subdir-objects option
It seems to break the build because of an Automake bug, see:
<http://debbugs.gnu.org/cgi/bugreport.cgi?bug=13928>
Revert "configure: Use subdir-objects Automake option"
This reverts commit 5ae20fb915659163e68f5f18bbd444e8de30d02b.
Lesson learned (again): don't do something just because Automake tells
you to.
Notice the same thing happened back in commit e46bfc1c43 which
was reverted again in b1e3997ce7b894.
configure: Remove AC_PROG_RANLIB, it's not needed
Based on this warning from libtoolize:
libtoolize: `AC_PROG_RANLIB' is rendered obsolete by `LT_INIT'
configure: Use subdir-objects Automake option
This fixes the following warnings:
src/libtracker-miner/Makefile-shared-sources.decl:5: warning: source file '$(top_srcdir)/src/libtracker-miner/tracker-monitor.c' is in a subdirectory,
src/libtracker-miner/Makefile-shared-sources.decl:5: but option 'subdir-objects' is disabled
src/libtracker-miner/Makefile.am:10: 'src/libtracker-miner/Makefile-shared-sources.decl' included from here
automake: warning: possible forward-incompatibility.
automake: At least a source file is in a subdirectory, but the 'subdir-objects'
automake: automake option hasn't been enabled. For now, the corresponding output
automake: object file(s) will be placed in the top-level directory. However,
automake: this behaviour will change in future Automake versions: they will
automake: unconditionally cause object files to be placed in the same subdirectory
automake: of the corresponding sources.
automake: You are advised to start using 'subdir-objects' option throughout your
automake: project, to avoid future incompatibilities.
src/tracker-preferences/Makefile.am:34: warning: source file '$(top_srcdir)/src/miners/fs/tracker-config.c' is in a subdirectory,
src/tracker-preferences/Makefile.am:34: but option 'subdir-objects' is disabled
src/libtracker-miner/Makefile-shared-sources.decl:17: warning: source file '$(top_srcdir)/src/libtracker-miner/tracker-crawler.c' is in a subdirectory,
src/libtracker-miner/Makefile-shared-sources.decl:17: but option 'subdir-objects' is disabled
tests/libtracker-miner/Makefile.am:11: 'src/libtracker-miner/Makefile-shared-sources.decl' included from here
src/libtracker-miner/Makefile-shared-sources.decl:5: warning: source file '$(top_srcdir)/src/libtracker-miner/tracker-monitor.c' is in a subdirectory,
src/libtracker-miner/Makefile-shared-sources.decl:5: but option 'subdir-objects' is disabled
tests/libtracker-miner/Makefile.am:11: 'src/libtracker-miner/Makefile-shared-sources.decl' included from here
autoreconf: automake failed with exit status: 1