- Jun 03, 2013
-
-
Tristan Van Berkom authored
Cannot feed more than one file to EDS's enum generation rule, or else it results in imbalanced G_BEGIN/END_DECLS declarations.
-
- May 28, 2013
-
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
This patch modifies data/sources/system-address-book.source.in so that it has a customized summary optimized for person names and phone numbers. This customization adds the following fields to the summary with prefix indexes (for optimized SQLite queries): E_CONTACT_FULL_NAME, E_CONTACT_FAMILY_NAME, E_CONTACT_NICKNAME, E_CONTACT_GIVEN_NAME, E_CONTACT_TEL Additionally, the E_CONTACT_TEL field is configured to store the normalized phone number data in SQLite in order to optimize the phone number queries such as E_BOOK_QUERY_EQUALS_PHONE_NUMBER. The code to achieve the same configuration is as follows: e_source_backend_summary_setup_set_summary_fields ( setup, E_CONTACT_FULL_NAME, E_CONTACT_FAMILY_NAME, E_CONTACT_NICKNAME, E_CONTACT_GIVEN_NAME, E_CONTACT_TEL, 0); e_source_backend_summary_setup_set_indexed_fields ( setup, E_CONTACT_FULL_NAME, E_BOOK_INDEX_PREFIX, E_CONTACT_FAMILY_NAME, E_BOOK_INDEX_PREFIX, E_CONTACT_NICKNAME, E_BOOK_INDEX_PREFIX, E_CONTACT_GIVEN_NAME, E_BOOK_INDEX_PREFIX, E_CONTACT_TEL, E_BOOK_INDEX_PREFIX, E_CONTACT_TEL, E_BOOK_INDEX_PHONE, 0);
-
Tristan Van Berkom authored
This is a bit of a messy patch to support the same API backported into 3.6, what we've done is extend the addressboook 'open' D-Bus API to report the locale of the addressbook initially. Using the previously created get_locale() D-Bus API was broken as it caused lockups when an addressbook is opened from a dedicated thread (as test-client-view-operations does). This patch is just a huge workaround to avoid the lockup in test-client-view-operations.
-
Tristan Van Berkom authored
This new test changes the locale dynamically and then checks new sort order of the converted addressbook between every locale change.
-
Tristan Van Berkom authored
Mostly changed to include the locale in the move-by tests and to stop setting the 'EDS_COLLATE' variable to hack the current locale used by the backend.
-
Tristan Van Berkom authored
Also, the 'move-by' testing API now has a locale argument.
-
Tristan Van Berkom authored
Propagate the 'locale' property change notifications when the dbus proxy notifies a locale property change, also load the initial value of the dbus proxy's locale property for the initial value.
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
Locale setting is now only ever guessed when creating a new addressbook before e_book_backend_sqlitedb_set_locale() is called, setting the locale will save the localization setting and it will be reused on subsequent accesses. If and when a locale setting is changed, the contact sort keys are regenerated with a new ECollator for the new locale.
-
Tristan Van Berkom authored
When org.freedesktop.locale1 is available, listen to changes in the LC_COLLATE locale and configure backends with locale changes using e_book_backend_set_locale(), notify property changes via the locale property on the addressbook D-Bus API. Also, load the backend's initially set locale as the locale property value until the org.freedesktop.locale1 D-Bus interface notifies us of a locale change on the system bus.
-
Tristan Van Berkom authored
Use these to dynamically set the locale of a running addressbook, and also to fetch the currently configured locale of an addressbook at startup time
-
Tristan Van Berkom authored
In future versions this is a D-Bus property notification, but those are not supported with the templates in 3.6
- May 27, 2013
-
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
Added migration tests also for fr_CA and de_DE locale, the de_DE locale expects a @collation=phonebook type of sort order but only the test-sqlite-cursor-posix-initial test which creates the shared addressbook defines it, so we also test that the collation rule is properly stored and introspected in the SQLite tables.
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
No point in doing one for every locale, we test a bunch of locales already with the move-by test.
-
Tristan Van Berkom authored
Added new sortable vcards and changed the whole sortable data set. Now there are 4 move-by tests: posix, en_US, fr_CA and de_DE. All tests try the default and phonebook collation rules, the de_DE test data sorts differently under phonebook sort. A table was added to data-test-utils.h for reference, showing the expected order of sorted contacts in given locales.
-
Tristan Van Berkom authored
For now let's keep a single ECollator, it seems that we don't want to configure collation rules but rather have a single collator API which maintains a knowledge base of what tailoring to choose in which locale.
-
Tristan Van Berkom authored
Now check the ESourceBackendSummarySetup extension for custom collations, use a default ECollator for the current locale for every contact field unless a custom collation is specified, if a custom collation is specified for a given field then use a special ECollator to sort that field.
-
Tristan Van Berkom authored
This reverts commit adc333b3.
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
Instead ECollator will try to make a best guess at what type of collation is desired for a given language
-
Tristan Van Berkom authored
The ECollator API wraps up the required ICU APIs into something more convenient for usage in EDS.
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
E_BOOK_QUERY_REGEX_NORMAL is supported in summary queries if the query field is summarized, E_BOOK_QUERY_REGEX_RAW and normalized but non-summarized queries are delegated to EBookBackendSexp for matching.
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
Added E_BOOK_QUERY_REGEX_NORMAL and E_BOOK_QUERY_REGEX_RAW
-
Tristan Van Berkom authored
-
David Woodhouse authored
With large addressbooks, using an inner join which produces hundreds of thousands of duplicate results, and then filtering them back down to what we really wanted with 'DISTINCT' is really inefficient. I can't even tell you *how* inefficient, because it didn't complete a single query in the hours that I left it running. Switch it to a more sensible outer join and it completes in a second or so on my 216000-entry address book.
-
- May 18, 2013
-
-
SyncEvolution Nightly Testing authored
Conflicts: libedataserver/e-gdbus-templates.c Conflicts were due to earlier merge and revert attempts. Now libedataserver/e-gdbus-templates.c is the same as in gnome-3-6.
-
- Apr 26, 2013
-
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored
-
Tristan Van Berkom authored