Milan Crha [Thu, 22 Nov 2012 16:13:49 +0000 (17:13 +0100)]
Bug #688795 - Crash under e_cal_backend_contacts_start_view()
Milan Crha [Thu, 22 Nov 2012 15:15:55 +0000 (16:15 +0100)]
Bug #202576 - A way to refresh IMAP folder list
Nilamdyuti Goswami [Thu, 22 Nov 2012 14:17:43 +0000 (19:47 +0530)]
Assamese translation updated
Daniel Mustieles [Thu, 22 Nov 2012 10:30:02 +0000 (11:30 +0100)]
Updated Spanish translation
Tristan Van Berkom [Tue, 20 Nov 2012 06:40:59 +0000 (15:40 +0900)]
EBookBackendSexp: Fixed to use the new e_util_utf8_normalize() function
Tristan Van Berkom [Mon, 8 Oct 2012 07:43:06 +0000 (16:43 +0900)]
Added test case to create a book with custom summary fields and fetch a contact.
This test case configures some summary fields and indexes and then performs
some queries and asserts the results. Notably the configured summary uses multi valued
fields E_CONTACT_TEL and E_CONTACT_EMAIL.
Some extra vcards are added to the test data directory in order to have some
mixed up emails and phone numbers to test with
Tristan Van Berkom [Wed, 7 Nov 2012 08:31:10 +0000 (17:31 +0900)]
EBookBackendFile: Use the new ESourceBackendSummarySetup to configure the summary
Tristan Van Berkom [Wed, 7 Nov 2012 08:19:57 +0000 (17:19 +0900)]
EBookBackendSqliteDB: Added e_book_backend_sqlitedb_new_full()
This is a major refactor of EBookBackendSqliteDB and adds a few
features. The new_full() variant for creating the SQLiteDB makes
the summary fields and indexes configurable.
String types, Boolean Types and Attribute List types of EContactFields
are all supported by the summary.
The summary fields can all have Prefix or Suffix indexes.
In order to actually hit the created indexes, all field data
is normalized with g_utf8_casefold(), for Suffix indexes,
additional columns are maintained with the data in reverse
(this allows us to run prefix searches in reverse which use
the indexes when asked to perform a suffix search)
Furthermore the patch also deprecates:
e_book_backend_sqlitedb_is_summary_query () and
e_book_backend_sqlitedb_is_summary_fields ()
in favour of some new apis which take the given EBookBackendSqliteDB in context.
Note also that since attribute list summary fields are supported,
we've dropped the custom hackery which checks related email_1, email_2, email_3
and email_4 fields... instead the E_CONTACT_EMAIL field should simply
be used to store however many emails are assigned to a given contact.
Tristan Van Berkom [Mon, 8 Oct 2012 05:18:10 +0000 (14:18 +0900)]
Added ESourceBackendSummarySetup extension
Firstly this extension allows configuration which EContactFields
should be indexed as "summary fields".
Secondly, the configured summary fields can have indexes, those indexes
can be of type PREFIX or SUFFIX
Tristan Van Berkom [Mon, 19 Nov 2012 11:19:43 +0000 (20:19 +0900)]
Added e_util_utf8_normalize()
A utility function to lowercase/remove accents from a utf8 string
Tristan Van Berkom [Mon, 15 Oct 2012 06:45:35 +0000 (15:45 +0900)]
Added e_enum_from_string() and e_enum_to_string() utilities in e-data-server-utiils
Milan Crha [Wed, 21 Nov 2012 18:56:44 +0000 (19:56 +0100)]
Bug #680201 - Backup restore doesn't migrate accounts from GConf
Milan Crha [Wed, 21 Nov 2012 16:55:15 +0000 (17:55 +0100)]
Bug #687670 - Signing with both pgp and S/MIME causes a broken signature
Milan Crha [Wed, 21 Nov 2012 11:34:29 +0000 (12:34 +0100)]
Bug #677378 - Crash in imapx_query_auth_types_sync()
Milan Crha [Tue, 20 Nov 2012 19:08:15 +0000 (20:08 +0100)]
Bug #688479 - Empty name selector dialog on open
Paul Menzel [Tue, 20 Nov 2012 17:59:05 +0000 (18:59 +0100)]
Bug 688639 – SMTP: Correctly deal with 535 authentication error
Entering a wrong password, Evolution would fail to send messages using SMTP with the following error.
Bad authentication response from server.
The log file of the mail server – Exim 4 in this case – contains the following error.
cram_md5_server authenticator failed for myhost.example.org ([10.10.10.10]) [xx.xxx.xxx.xxx]: 535 Incorrect authentication data (set_id=joey)
Looking at the code after the rewrite in
commit
68c269a66bd484768b9c1517e7dc2dd30f682485
Author: Matthew Barnes <mbarnes@redhat.com>
Date: Sat Oct 15 09:51:35 2011 -0400
smtp: Adapt to Camel's new authentication API.
present since git tag EVOLUTION_DATA_SERVER_3_3_1, accidentally changed the logic not acting correctly, when receiving a 535 authentication error response and therefore was overridden by the next code block and the variable `result` changed to `CAMEL_AUTHENTICATION_ERROR` instead of `CAMEL_AUTHENTICATION_REJECTED`, which resulted in the situation where the password was not reset and therefore the user not asked to reenter it.
Changing the logic, so the variable is not overridden successfully fixes the issue.
Milan Crha [Tue, 20 Nov 2012 15:35:41 +0000 (16:35 +0100)]
Bug #688660 - Crash on webdav contact cache update
Daniel Mustieles [Tue, 20 Nov 2012 14:50:30 +0000 (15:50 +0100)]
Updated Spanish translation
Tristan Van Berkom [Thu, 8 Nov 2012 08:13:07 +0000 (17:13 +0900)]
Fixing e_book_backend_sexp_match_contact() for emails and phone numbers
This patch fixes the match function to match any phone number or email
found in a given vcard, not only the restrictive list of phone number
and email types defined in the EContactField enumeration.
Tristan Van Berkom [Fri, 5 Oct 2012 09:35:05 +0000 (18:35 +0900)]
EContact: Added boxed type E_CONTACT_TYPE_ATTR_LIST and e_contact_field_type()
This boxed type is used in place of pointer types for properties
installed on EContact which are returned lists of strings.
This does not effect the behavior of e_contact_set()/e_contact_get(),
only it allows e_contact_field_type() to return a distinct type
for those fields, allowing automated code to introspect those useful
contact field types.
The e_contact_field_type() convenience function is used to fetch the GType
for a given field when used with EContact (this allows more automated
routines to be written for EContactFields).
Matthew Barnes [Sun, 18 Nov 2012 12:50:49 +0000 (07:50 -0500)]
ECalClientView cleanups.
Let ECalClientView create its own D-Bus proxy object by implementing
GInitableIface and adding "connection" and "object-path" construct-only
GObject properties.
This eliminates e-cal-client-view-private.h and adds the following
public accessor functions:
e_cal_client_view_get_connection()
e_cal_client_view_get_object_path()
Matthew Barnes [Sat, 17 Nov 2012 22:47:00 +0000 (17:47 -0500)]
EBookClientView cleanups.
Let EBookClientView create its own D-Bus proxy object by implementing
GInitableIface and adding "connection" and "object-path" construct-only
GObject properties.
This eliminates e-book-client-view-private.h and adds the following
public accessor functions:
e_book_client_view_get_connection()
e_book_client_view_get_object_path()
Matthew Barnes [Sat, 17 Nov 2012 20:04:49 +0000 (15:04 -0500)]
Fix EBookClient property fetching crack.
Matthew Barnes [Sat, 17 Nov 2012 13:57:38 +0000 (08:57 -0500)]
Post-release version bump.
Matthew Barnes [Sat, 17 Nov 2012 13:31:49 +0000 (08:31 -0500)]
NEWS update for 3.7.2 release.
Matthew Barnes [Sat, 17 Nov 2012 13:33:51 +0000 (08:33 -0500)]
Missed a couple files when removing EBookBackendVCF.
Matthew Barnes [Sat, 17 Nov 2012 12:54:37 +0000 (07:54 -0500)]
Remove EBookBackendVCF.
Strolling through the commit history of this backend, I had to dig all
the way back to 2005 before I started seeing actual development work.
Since then the commits have all been code cleanups and adapting it to
backend API changes.
There is no way to create a VCF address book from Evolution, which means
this entire backend is dead code and we've been wasting our time with it.
Dump it.
Matthew Barnes [Fri, 16 Nov 2012 19:16:45 +0000 (14:16 -0500)]
Deprecate BOOK_BACKEND_PROPERTY_SUPPORTED_AUTH_METHODS.
This property is not used anywhere.
Remove all references to it from built-in backends.
The following EBook functions now simply return an empty list:
e_book_get_supported_auth_methods
e_book_async_get_supported_auth_methods
e_book_get_supported_auth_methods_async
Milan Crha [Fri, 16 Nov 2012 08:59:51 +0000 (09:59 +0100)]
Bug #674454 - Crash in e_book_backend_ldap_authenticate_user()
Milan Crha [Wed, 14 Nov 2012 22:37:24 +0000 (23:37 +0100)]
Replace deprecated Gtk+ symbols (as of Gtk+ 3.6.x)
Milan Crha [Tue, 13 Nov 2012 15:58:49 +0000 (16:58 +0100)]
Bug #688146 - Crash in Contacts calendar backend
Paul Menzel [Tue, 13 Nov 2012 13:02:38 +0000 (14:02 +0100)]
Bug #687634 - Improve camel console warning text
Milan Crha [Tue, 13 Nov 2012 07:37:40 +0000 (08:37 +0100)]
Bug #687865 - Shows old reminders for GOA accounts
Matthew Barnes [Tue, 13 Nov 2012 00:17:29 +0000 (19:17 -0500)]
Revert "EBookBackendGoogle: Make room for other authorizers."
This reverts commit
a5dae1a780f898ab453f574440ca02e108b4ebef.
Actually no, don't want to do that. Misread the code.
Matthew Barnes [Mon, 12 Nov 2012 23:56:10 +0000 (18:56 -0500)]
EBookBackendGoogle: Make room for other authorizers.
Christian Kirbach [Mon, 12 Nov 2012 21:49:34 +0000 (22:49 +0100)]
Update German translation
Matthew Barnes [Mon, 12 Nov 2012 16:44:11 +0000 (11:44 -0500)]
EDataBookFactory: Remove the GOA-specific bits.
Defer to the gnome-online-accounts module. See the previous commit.
Matthew Barnes [Mon, 12 Nov 2012 16:38:45 +0000 (11:38 -0500)]
gnome-online-accounts: Add an EDataFactory extension.
This extension takes over for the GOA bits in EDataBookFactory.
Install this extension to both the addressbook and calendar backend
directories, even though only EBookBackendGoogle uses it at present.
Matthew Barnes [Mon, 12 Nov 2012 14:31:16 +0000 (09:31 -0500)]
EDataFactory: Add a "backend-created" signal.
Allows EDataFactory extensions to catch newly-created backends and
perform additional initialization on it.
The use case I have in mind is to move all the GOA initialization in
EDataBookFactory to the gnome-online-accounts module as an extension.
Note this uses one of the reserved pointer slots in EDataFactoryClass
and therefore does not break ABI.
Matthew Barnes [Mon, 12 Nov 2012 11:54:26 +0000 (06:54 -0500)]
Remove unused OAUTH_CFLAGS / OAUTH_LIBS.
Missed these when I removed the liboauth dependency.
Matthew Barnes [Sat, 10 Nov 2012 02:54:46 +0000 (21:54 -0500)]
Coding style and whitespace cleanup.
Piotr Drąg [Fri, 9 Nov 2012 23:29:49 +0000 (00:29 +0100)]
Updated POTFILES.in
Matthew Barnes [Fri, 9 Nov 2012 16:25:13 +0000 (11:25 -0500)]
EBookBackendSqliteDB cleanups.
Improve error handling and make commits/rollbacks less awkward.
Matthew Barnes [Fri, 9 Nov 2012 12:49:33 +0000 (07:49 -0500)]
Use g_type_ensure() where applicable.
Matthew Barnes [Fri, 9 Nov 2012 11:29:21 +0000 (06:29 -0500)]
CamelIMAPXServer: Work around crash in imapx_exit_idle().
g_cond_clear() doesn't like being given an uninitialized GCond.
The "start_watch_cond" and "start_watch_exit" structures are being
cleared unconditionally, so we should initialize them unconditionally.
Matej Urbančič [Thu, 8 Nov 2012 20:36:46 +0000 (21:36 +0100)]
Updated Slovenian translation
Matthew Barnes [Thu, 8 Nov 2012 13:30:22 +0000 (08:30 -0500)]
Restore e_flag_timed_wait().
Just deprecate the function, don't delete it. GTimeVal isn't going
away any time soon. It's not worth a libedataserver soname bump.
Tristan Van Berkom [Thu, 8 Nov 2012 11:28:06 +0000 (20:28 +0900)]
Disabled debug print in EBookBackendFile
This was somehow missed in my last commit.
Tristan Van Berkom [Mon, 29 Oct 2012 06:33:09 +0000 (15:33 +0900)]
Local file backend modified to use sqlite db exclusively
Patch includes a migration routine to migrate data from a previously
installed BDB.
Tristan Van Berkom [Mon, 29 Oct 2012 05:40:47 +0000 (14:40 +0900)]
EBookBackendSqliteDB: Added e_book_backend_sqlitedb_get/set_revision()
In preperation for the migration, there needs to be a way to persist the
overall addressbook revision. This patch also increments the internal
SQLite DB format version and adds the new revision column to for main
folders table of an addressbook (and increments the stored DB version)
if the DB version is out of date.
Tristan Van Berkom [Mon, 29 Oct 2012 05:34:55 +0000 (14:34 +0900)]
EBookBackendSqliteDB: Fixed sexp interpretation for 'id' field.
In the sqlite backend, the 'id' is stored as the 'uid' column.
Matthew Barnes [Wed, 7 Nov 2012 13:31:52 +0000 (08:31 -0500)]
Revert "calendar: Drop old ECalSourceTypeEnum from introspected API"
This reverts commit
3183733023f022a89980cbf3f7292bde0bbd9954.
We no longer introspect deprecated API. Remove the cruft.
Matthew Barnes [Wed, 7 Nov 2012 12:53:56 +0000 (07:53 -0500)]
Disable runtime check in source_registry_initable_init().
Rationale explained in the source code.
Dan Williams [Wed, 7 Nov 2012 11:59:51 +0000 (06:59 -0500)]
Bug 687694 - imapx_untagged_vanished() is inefficient
For uid->len = 318320, timing the loop yields:
g_list_append(): 60+ seconds
g_list_prepend(): 0.3 seconds
Milan Crha [Wed, 7 Nov 2012 12:08:51 +0000 (13:08 +0100)]
Bug #687642 - Crash under caldav_synch_slave_loop()
Andreas Kohn [Wed, 7 Nov 2012 11:32:28 +0000 (12:32 +0100)]
Bug #669082 - Workaround crash in imapx_command_copy_messages_step_done()
Giovanni Campagna [Sun, 28 Oct 2012 15:24:46 +0000 (16:24 +0100)]
Google: allow writing X-GOOGLE-SYSTEM-GROUP-IDS
We expose GData system groups as an extended attribute with that
name, but if we found it in the EContact we would turn it into
an extended property. Instead, marshal the system group IDs into
Evolution categories, and then those into GData groups or system
groups.
https://bugzilla.gnome.org/show_bug.cgi?id=687051
Milan Crha [Tue, 6 Nov 2012 17:58:12 +0000 (18:58 +0100)]
Bug #687776 - E-D-S crash in local address book
Milan Crha [Tue, 6 Nov 2012 10:07:07 +0000 (11:07 +0100)]
Introduce --enable-strict configure option
This is currently used to add extra AM_CPPFLAGS to disable
deprecated symbols from GLib.
Kjartan Maraas [Mon, 5 Nov 2012 20:06:43 +0000 (21:06 +0100)]
Updated Norwegian bokmål translation
Milan Crha [Mon, 5 Nov 2012 19:51:41 +0000 (20:51 +0100)]
Use soup_message_headers_get_list() as a replacement of soup_message_headers_get()
CalDAV backend didn't receive proper value for access right, which
caused reject of the login to the server.
Milan Crha [Mon, 5 Nov 2012 19:39:59 +0000 (20:39 +0100)]
Replace deprecated soup_message_headers_get()
Milan Crha [Mon, 5 Nov 2012 19:19:01 +0000 (20:19 +0100)]
Replace deprecated GLib symbols (as of GLib 2.34.x)
Paul Menzel [Sun, 4 Nov 2012 21:43:32 +0000 (22:43 +0100)]
[l10n] Fixed German translation, bug #687562
Matthew Barnes [Sun, 4 Nov 2012 16:41:16 +0000 (11:41 -0500)]
EBookBackendSqliteDB: Add argument guards.
Yuri Myasoedov [Sun, 4 Nov 2012 16:21:41 +0000 (20:21 +0400)]
Updated Russian translation
Matthew Barnes [Fri, 2 Nov 2012 22:01:35 +0000 (18:01 -0400)]
vala: Remove unused metadata.
Matthew Barnes [Fri, 2 Nov 2012 21:49:13 +0000 (17:49 -0400)]
Exclude deprecated API from introspection.
Deprecated APIs were deprecated for a reason. The last thing we
should be doing is enabling their continued use in other languages.
Henceforth, any symbols tagged with EDS_DISABLED_DEPRECATED will be
automatically excluded from introspection and language bindings.
Matthew Barnes [Fri, 2 Nov 2012 21:47:38 +0000 (17:47 -0400)]
Introspection cleanups.
Matthew Barnes [Fri, 2 Nov 2012 20:57:39 +0000 (16:57 -0400)]
Remove E_BOOK_DISABLE_DEPRECATED and E_CAL_DISABLE_DEPRECATED.
We already use EDS_DISABLE_DEPRECATED, these other definitions only
confuse matters.
Matthew Barnes [Fri, 2 Nov 2012 20:33:03 +0000 (16:33 -0400)]
Suppress g-ir-scanner warnings.
They're far too noisy, and make it difficult to spot new C compiler
warnings, which are the only ones I care about. We can enable them
again when we're serious about cleaning them up. In the meantime,
they're just getting in the way.
Colin Walters [Fri, 2 Nov 2012 18:29:13 +0000 (14:29 -0400)]
calendar: Drop old ECalSourceTypeEnum from introspected API
It clashes when generating the vala bindings, and we're just keeping
it around for C API consumers.
Matthew Barnes [Fri, 2 Nov 2012 17:43:38 +0000 (13:43 -0400)]
e-cal-view.c: Remove unnecessary #includes.
Matthew Barnes [Fri, 2 Nov 2012 17:43:22 +0000 (13:43 -0400)]
e-cal.c: Remove unnecessary #includes.
Matthew Barnes [Fri, 2 Nov 2012 17:42:57 +0000 (13:42 -0400)]
e-book-view.c: Remove unnecessary #includes.
Matthew Barnes [Fri, 2 Nov 2012 17:41:49 +0000 (13:41 -0400)]
e-book.c: Remove unnecessary #includes.
Matthew Barnes [Fri, 2 Nov 2012 17:28:05 +0000 (13:28 -0400)]
Update libedataserverui API documentation.
Matthew Barnes [Fri, 2 Nov 2012 17:15:00 +0000 (13:15 -0400)]
Fix invalid Gtk-Doc syntax in comments.
Not certain this is everything, just what I could find.
Matthew Barnes [Fri, 2 Nov 2012 13:12:38 +0000 (09:12 -0400)]
Reimplement ECal using ECalClient.
This is primarily so I don't have to duplicate work while overhauling
the D-Bus API, but it's good to do anyway; it simplifies the deprecated
logic and helps ensure the two APIs behave consistently.
Matthew Barnes [Fri, 2 Nov 2012 11:41:22 +0000 (07:41 -0400)]
Reimplement ECalView using ECalClientView.
This is primarily so I don't have to duplicate work while overhauling
the D-Bus API, but it's good to do anyway; it simplifies the deprecated
logic and helps ensure the two APIs behave consistently.
Matthew Barnes [Mon, 29 Oct 2012 21:44:32 +0000 (17:44 -0400)]
Reimplement EBook using EBookClient.
This is primarily so I don't have to duplicate work while overhauling
the D-Bus API, but it's good to do anyway: it simplifies the deprecated
logic and helps ensure the two APIs behave consistently.
Matthew Barnes [Tue, 30 Oct 2012 00:24:14 +0000 (20:24 -0400)]
Reimplement EBookView using EBookClientView.
This is primarily so I don't have to duplicate work while overhauling
the D-Bus API, but it's good to do anyway: it simplifies the deprecated
logic and helps ensure the two APIs behave consistently.
Matthew Barnes [Fri, 2 Nov 2012 13:19:16 +0000 (09:19 -0400)]
e-cal-view.h cleanups.
Matthew Barnes [Fri, 2 Nov 2012 13:13:53 +0000 (09:13 -0400)]
e-book-view.h cleanups.
Milan Crha [Fri, 2 Nov 2012 10:09:03 +0000 (11:09 +0100)]
e_async_closure_callback: Do not test 'object' parameter validity
It is unnecessary to test validity of the 'object' parameter,
especially when this callback does nothing with it. The reason
is that the secret_password_store() calls the callback with NULL
'object', which leads to a deadlock on the source registry side.
Milan Crha [Fri, 2 Nov 2012 09:39:26 +0000 (10:39 +0100)]
Bug #685090 - GOA EWS module doesn't save all required values
Milan Crha [Fri, 2 Nov 2012 08:28:55 +0000 (09:28 +0100)]
Collection's child source Enabled should match collections Enabled
Milan Crha [Fri, 2 Nov 2012 08:24:50 +0000 (09:24 +0100)]
Revert "Revert "Bug #678400 - ECollectionBackend populates disabled accounts""
This reverts commit
1c4fb39c131a03428035a2035fca8ea462fc7714,
because the initial revert was a misunderstanding of the change.
Matthew Barnes [Thu, 1 Nov 2012 17:52:05 +0000 (13:52 -0400)]
Add AllowAuthPromptAll() method to SourceManager interface.
Equivalent to calling AllowAuthPrompt() on each exported data source,
but cuts way down on the D-Bus traffic. Since the registry service is
commonly activated during a desktop session startup, this can actually
clog up the session bus and slow down login.
This leaves the source-specific AllowAuthPrompt() method unused, but
there are still use cases for it which we haven't implemented yet, so
leave it be.
Matthew Barnes [Thu, 1 Nov 2012 13:38:49 +0000 (09:38 -0400)]
Remove version number from ${privincludedir}.
Last man standing, so may as well. Version not needed there either.
Matthew Barnes [Thu, 1 Nov 2012 13:35:00 +0000 (09:35 -0400)]
Remove erroneous version number from ${privlibdir}.
This slipped through because we don't actually reference ${privlibdir}
in our backend/module/provider directory definitions. Fix that too.
Also, some pkg-config files needed to be tweaked as a result of this.
Matthew Barnes [Thu, 1 Nov 2012 11:35:39 +0000 (07:35 -0400)]
Remove version number from ${privdatadir}.
In talking to some people at Red Hat about how to deploy pre-configured
Evolution accounts, one thing I suggested is to provide a package which
installs .source files to one of the system-wide locations:
/usr/share/evolution-data-server-3.8/ro-sources
/usr/share/evolution-data-server-3.8/rw-sources
But that's not going to work in general with the version number in the
path. Upgrading E-D-S to, say, 3.10 would invalidate such a package.
Remove the version number from our ${privdatadir}. It's not needed.
Milan Crha [Tue, 30 Oct 2012 17:59:40 +0000 (18:59 +0100)]
Bug #681962 - Identify server certificates by hostname and fingerprint
Milan Crha [Tue, 30 Oct 2012 17:48:30 +0000 (18:48 +0100)]
Bug #685034 - [imapx] Read status of email is not persisted correctly
Milan Crha [Mon, 29 Oct 2012 16:05:22 +0000 (17:05 +0100)]
Bug #687100 - Crash under cal_backend_contacts_remove_book_record()
Milan Crha [Mon, 29 Oct 2012 14:12:02 +0000 (15:12 +0100)]
Bug #686924 - Doesn't parse Gerrit encoded From address correctly
Matthew Barnes [Sun, 28 Oct 2012 11:38:10 +0000 (07:38 -0400)]
imapx_command_start: Improve handling of write-to-stream error.
For some reason we were calling camel_stream_write_string() without
passing a GError, and on failure setting a very generic error message
and losing potentially useful information about the actual error.
Matthew Barnes [Sat, 27 Oct 2012 16:27:04 +0000 (12:27 -0400)]
Use code coverage macros from gnome-common.
The GNOME_CODE_COVERAGE macro in gnome-common automates the process
nicely. Just run "make check-code-coverage" from the top-level source
directory (or from any of the "tests" subdirectories) and open the nice
coverage analysis report:
evolution-data-server-x.y.z-coverage/index.html
Matthew Barnes [Sat, 27 Oct 2012 02:45:59 +0000 (22:45 -0400)]
Bug 685986 - ESourceRegistry: Wait for signals after creating sources
e_source_registry_create_sources_sync() now waits for "object-added"
signals from its GDBusObjectManagerClient before returning. This allows
the newly-created ESources to be obtained immediately after the function
returns. Prior to this there was a small delay before the newly-created
ESources became available through the ESourceRegistry API.
There is a caveat to this. If any of the scratch ESources reference a
parent UID which is not presently exported by the registry service, the
registry service will refuse to export the newly-created ESource until
its parent becomes available.
In this scenario, to avoid waiting forever for an "object-added" signal
that isn't coming, we place a limit on the wait time (two seconds). If
this time limit expires, we return from the function regardless.
Matthew Barnes [Fri, 26 Oct 2012 20:34:10 +0000 (16:34 -0400)]
Add a test case for bug 685986.
Managed to reproduce the bug using the test case.