platform/upstream/evolution-data-server.git
12 years agoe_async_closure_callback: Do not test 'object' parameter validity
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.

12 years agoBug #685090 - GOA EWS module doesn't save all required values
Milan Crha [Fri, 2 Nov 2012 09:39:26 +0000 (10:39 +0100)]
Bug #685090 - GOA EWS module doesn't save all required values

12 years agoCollection's child source Enabled should match collections Enabled
Milan Crha [Fri, 2 Nov 2012 08:28:55 +0000 (09:28 +0100)]
Collection's child source Enabled should match collections Enabled

12 years agoRevert "Revert "Bug #678400 - ECollectionBackend populates disabled accounts""
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.

12 years agoAdd AllowAuthPromptAll() method to SourceManager interface.
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.

12 years agoRemove version number from ${privincludedir}.
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.

12 years agoRemove erroneous version number from ${privlibdir}.
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.

12 years agoRemove version number from ${privdatadir}.
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.

12 years agoBug #681962 - Identify server certificates by hostname and fingerprint
Milan Crha [Tue, 30 Oct 2012 17:59:40 +0000 (18:59 +0100)]
Bug #681962 - Identify server certificates by hostname and fingerprint

12 years agoBug #685034 - [imapx] Read status of email is not persisted correctly
Milan Crha [Tue, 30 Oct 2012 17:48:30 +0000 (18:48 +0100)]
Bug #685034 - [imapx] Read status of email is not persisted correctly

12 years agoBug #687100 - Crash under cal_backend_contacts_remove_book_record()
Milan Crha [Mon, 29 Oct 2012 16:05:22 +0000 (17:05 +0100)]
Bug #687100 - Crash under cal_backend_contacts_remove_book_record()

12 years agoBug #686924 - Doesn't parse Gerrit encoded From address correctly
Milan Crha [Mon, 29 Oct 2012 14:12:02 +0000 (15:12 +0100)]
Bug #686924 - Doesn't parse Gerrit encoded From address correctly

12 years agoimapx_command_start: Improve handling of write-to-stream error.
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.

12 years agoUse code coverage macros from gnome-common.
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

12 years agoBug 685986 - ESourceRegistry: Wait for signals after creating sources
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.

12 years agoAdd a test case for bug 685986.
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.

12 years agoPrototype a D-Bus unit test harness.
Matthew Barnes [Fri, 26 Oct 2012 20:30:20 +0000 (16:30 -0400)]
Prototype a D-Bus unit test harness.

Add functions for programmatically overriding XDG base directories
and launching a private D-Bus session bus using GTestDBus.

This requires Evolution-Data-Server to be installed, since the test
harness references D-Bus service files by the installation prefix.

12 years agoBump minimum GLib requirement to 2.34.
Matthew Barnes [Fri, 26 Oct 2012 20:27:36 +0000 (16:27 -0400)]
Bump minimum GLib requirement to 2.34.

Need to utilize the GTestDBus API, which is new in 2.34.

12 years agoUpdated Indonesian translation
Andika Triwidada [Sat, 27 Oct 2012 08:33:20 +0000 (15:33 +0700)]
Updated Indonesian translation

12 years agoBug 686719 - Missing linker flags in Camel tests
JB Lallement [Fri, 26 Oct 2012 15:26:13 +0000 (11:26 -0400)]
Bug 686719 - Missing linker flags in Camel tests

12 years agoUpdated Slovenian translation
Matej Urbančič [Thu, 25 Oct 2012 17:52:58 +0000 (19:52 +0200)]
Updated Slovenian translation

12 years agoPrefer e_source_registry_check_enabled().
Matthew Barnes [Thu, 25 Oct 2012 14:59:59 +0000 (10:59 -0400)]
Prefer e_source_registry_check_enabled().

Prefer e_source_registry_check_enabled() over e_source_get_enabled()
for most instances where we want to exclude disabled data sources.

12 years agoAdd e_source_registry_check_enabled().
Matthew Barnes [Thu, 25 Oct 2012 14:38:43 +0000 (10:38 -0400)]
Add e_source_registry_check_enabled().

Checks the enabled state of a given #ESource and of its ancestors to
determine if the #ESource should be displayed or acted on.

This is somewhat similar to widget sensitivity in GTK+.
cf. gtk_widget_get_sensitive() versus gtk_widget_is_sensitive()

12 years agoBug #680497 - POP3 re-adds messages to local Inbox
Milan Crha [Thu, 25 Oct 2012 10:02:13 +0000 (12:02 +0200)]
Bug #680497 - POP3 re-adds messages to local Inbox

12 years agoIMAPX: Explicitly pass GCancellable and GError to callbacks.
Matthew Barnes [Wed, 24 Oct 2012 21:28:25 +0000 (17:28 -0400)]
IMAPX: Explicitly pass GCancellable and GError to callbacks.

Further error handling improvements.  This involves some public API
changes, but the impact is limited to evolution-kolab.

 * CamelIMAPXCommandFunc now takes a GCancellable argument in addition
   to its GError argument.

 * The public GCancellable member of CamelIMAPXJob is now private, and
   the public GError member is removed.

 * The start() method of CamelIMAPXJob now takes a GCancellable and
   GError argument.  The GCancellable pass through this method call
   is always the CamelIMAPXJob's own private GCancellable.

 * A new camel_imapx_job_cancel() function which simply cancels the
   CamelIMAPXJob's internal GCancellable.

Additional changes internal to CamelIMAPXServer are:

 * Adapt operations to the new callback function signatures.

 * imapx_command_queue() now takes a GCancellable and GError argument
   and returns a 'success' boolean.  Other core job and command routing
   functions also now have similar function signatures.

 * Add error handling where needed, so that no failable function call
   goes unchecked.

12 years agoBug #686727 - POP3 provider deadlocks on finalize
Milan Crha [Wed, 24 Oct 2012 16:37:12 +0000 (18:37 +0200)]
Bug #686727 - POP3 provider deadlocks on finalize

12 years agoUpdated German translation
Hendrik Knackstedt [Wed, 24 Oct 2012 08:52:11 +0000 (10:52 +0200)]
Updated German translation

12 years agoCamelIMAPXServer: Handle errors in imapx_continuation().
Matthew Barnes [Tue, 23 Oct 2012 20:07:47 +0000 (16:07 -0400)]
CamelIMAPXServer: Handle errors in imapx_continuation().

There was no error handling at all on the write-to-stream operations.

12 years agoCamelIMAPXServer: Return a boolean from imapx_command_start_next().
Matthew Barnes [Tue, 23 Oct 2012 19:31:56 +0000 (15:31 -0400)]
CamelIMAPXServer: Return a boolean from imapx_command_start_next().

The function is failable and takes a GError, yet returns void.  This
is how GErrors pile up.  Return a boolean and add the necessary error
checks where the function is called.

12 years agoRevert "Bug #678400 - ECollectionBackend populates disabled accounts"
Matthew Barnes [Tue, 23 Oct 2012 18:32:06 +0000 (14:32 -0400)]
Revert "Bug #678400 - ECollectionBackend populates disabled accounts"

This reverts commit 9a7948409852e57721dcb2e9e63ae69d8ac78d40.

Need to find a better solution than corrupting what are supposed to be
separate state flags.

12 years agoBug #678400 - ECollectionBackend populates disabled accounts
Milan Crha [Tue, 23 Oct 2012 18:15:09 +0000 (20:15 +0200)]
Bug #678400 - ECollectionBackend populates disabled accounts

12 years agoCamelIMAPXServer: Improve error handling around stopping IDLE.
Matthew Barnes [Tue, 23 Oct 2012 16:07:50 +0000 (12:07 -0400)]
CamelIMAPXServer: Improve error handling around stopping IDLE.

imapx_stop_idle() returned a boolean but TRUE had an ambiguous
interpretation.  It either meant the DONE command was sent or an I/O
error occurred.  Define a new enum type (CamelIMAPXIdleStopResult) to
resolve the ambiguity.

Also imapx_command_idle_stop() returned a boolean but FALSE had an
ambiguous interpretation.  It either meant an I/O error occurred, or
there was no available CamelIMAPXStream from which to issue the DONE
command.  Resolve the ambiguity by forcing callers to explicitly pass
in a CamelIMAPXStream (and a GCancellable, while we're at it).

12 years agoCamelIMAPXServer: Declare imapx_uidset_* functions as static.
Matthew Barnes [Tue, 23 Oct 2012 15:04:21 +0000 (11:04 -0400)]
CamelIMAPXServer: Declare imapx_uidset_* functions as static.

These functions were never meant to be exported.  The "static" keyword
seems to have been omitted by accident.  Can't find any evidence of them
being called from outside camel-imapx-server.c, so just make them static.

12 years agoCamelIMAPXServer: Clean up static function declarations.
Matthew Barnes [Tue, 23 Oct 2012 14:53:46 +0000 (10:53 -0400)]
CamelIMAPXServer: Clean up static function declarations.

One argument per line.

12 years agoCamelIMAPXServer: Make sure all disabled bits have a comment.
Matthew Barnes [Tue, 23 Oct 2012 14:57:07 +0000 (10:57 -0400)]
CamelIMAPXServer: Make sure all disabled bits have a comment.

Normally I throw away all disabled-with-no-explanation-why code chunks,
but in this case they all seem related, so clarify that in the code.

12 years agoUpdated Hungarian translation
Gabor Kelemen [Tue, 23 Oct 2012 12:50:01 +0000 (14:50 +0200)]
Updated Hungarian translation

12 years agoEAuthenticationSession: Close prompt before storing password.
Matthew Barnes [Mon, 22 Oct 2012 17:20:11 +0000 (13:20 -0400)]
EAuthenticationSession: Close prompt before storing password.

Close our prompt before storing the password in the keyring.
If the keyring is locked, it will need to prompt the user for
a keyring password, but it can't do that if our password prompt
is still open since both prompts are system-modal.  Not sure what
would happen next; probably the store operation would either fail
or deadlock.

12 years agoBug #640594 - Save also UID/REV in WebDAV backend (for CardDAV)
Milan Crha [Mon, 22 Oct 2012 09:24:09 +0000 (11:24 +0200)]
Bug #640594 - Save also UID/REV in WebDAV backend (for CardDAV)

12 years agoPost-release version bump.
Matthew Barnes [Sun, 21 Oct 2012 00:53:15 +0000 (20:53 -0400)]
Post-release version bump.

12 years agoNEWS update for 3.7.1 release.
Matthew Barnes [Sat, 20 Oct 2012 20:45:46 +0000 (16:45 -0400)]
NEWS update for 3.7.1 release.

12 years agoCoding style and whitespace cleanup.
Matthew Barnes [Sat, 20 Oct 2012 21:04:04 +0000 (17:04 -0400)]
Coding style and whitespace cleanup.

12 years agoevolution-source-viewer: Select next source after deleting.
Matthew Barnes [Fri, 19 Oct 2012 18:12:14 +0000 (14:12 -0400)]
evolution-source-viewer: Select next source after deleting.

Deleting a list of sources is less painful when the next source is
automatically selected.

12 years agoFurther test code cleanups.
Matthew Barnes [Fri, 19 Oct 2012 15:49:44 +0000 (11:49 -0400)]
Further test code cleanups.

12 years agoBug 686421 - Restore libebook tests to minimum working condition
Tristan Van Berkom [Fri, 19 Oct 2012 15:14:36 +0000 (11:14 -0400)]
Bug 686421 - Restore libebook tests to minimum working condition

12 years agoMerge cleanups from 'gdbus-codegen' branch.
Matthew Barnes [Thu, 18 Oct 2012 18:56:15 +0000 (14:56 -0400)]
Merge cleanups from 'gdbus-codegen' branch.

Merge cosmetic cleanups from my 'gdbus-codegen' branch so I can more
easily review the important changes.

12 years agoUpdated Spanish translation
Daniel Mustieles [Fri, 19 Oct 2012 08:12:25 +0000 (10:12 +0200)]
Updated Spanish translation

12 years agoUpdated Slovenian translation
Matej Urbančič [Fri, 19 Oct 2012 06:46:53 +0000 (08:46 +0200)]
Updated Slovenian translation

12 years agoconfigure.ac: Revert unnecessary soname bumps.
Matthew Barnes [Thu, 18 Oct 2012 18:16:13 +0000 (14:16 -0400)]
configure.ac: Revert unnecessary soname bumps.

Adding new functions does not break backward-compatibility of the
programming or binary interfaces for these libraries.

libedataserver soname bumps are particularly disruptive to other GNOME
modules, and should be avoided if at all possible.

12 years agoBug #268618 - Add: Save replies in the folder of the message being replied to
Milan Crha [Thu, 18 Oct 2012 17:43:28 +0000 (19:43 +0200)]
Bug #268618 - Add: Save replies in the folder of the message being replied to

12 years agoBug #310978 - USENET always requires authentication
Milan Crha [Thu, 18 Oct 2012 10:48:30 +0000 (12:48 +0200)]
Bug #310978 - USENET always requires authentication

12 years agoUpdated Slovenian translation
Matej Urbančič [Thu, 18 Oct 2012 06:14:09 +0000 (08:14 +0200)]
Updated Slovenian translation

12 years agoBug #201807 - Cannot specify sendmail parameters
Milan Crha [Wed, 17 Oct 2012 17:46:42 +0000 (19:46 +0200)]
Bug #201807 - Cannot specify sendmail parameters

12 years agocamel_junk_filter_classify(): Fix return values.
Matthew Barnes [Wed, 17 Oct 2012 11:59:07 +0000 (07:59 -0400)]
camel_junk_filter_classify(): Fix return values.

12 years agoBug #641825 - Crash in poll_ldap() when finalizing the backend
Milan Crha [Wed, 17 Oct 2012 11:46:28 +0000 (13:46 +0200)]
Bug #641825 - Crash in poll_ldap() when finalizing the backend

12 years agoUpdated Greek translation
Tom Tryfonidis [Wed, 17 Oct 2012 07:50:39 +0000 (10:50 +0300)]
Updated Greek translation

12 years agoUpdated Czech translation
Marek Černocký [Wed, 17 Oct 2012 07:28:39 +0000 (09:28 +0200)]
Updated Czech translation

12 years agoChange camel_junk_filter_classify() signature.
Matthew Barnes [Tue, 16 Oct 2012 14:31:47 +0000 (10:31 -0400)]
Change camel_junk_filter_classify() signature.

Add CAMEL_JUNK_STATUS_ERROR to the CamelJunkStatus enum and have
camel_junk_filter_classify() return the status directly, rather than
forcing callers to check both a status value and a 'success' boolean.

Minor API break, but should only affect Evolution.

12 years agoUpdated Galician translations
Fran Diéguez [Tue, 16 Oct 2012 12:26:05 +0000 (14:26 +0200)]
Updated Galician translations

12 years agoUpdated Slovenian translation
Matej Urbančič [Tue, 16 Oct 2012 06:48:27 +0000 (08:48 +0200)]
Updated Slovenian translation

12 years agoUpdated Slovenian translation
Matej Urbančič [Tue, 16 Oct 2012 06:33:57 +0000 (08:33 +0200)]
Updated Slovenian translation

12 years agoUpdated Indonesian translation
Andika Triwidada [Tue, 16 Oct 2012 00:29:33 +0000 (07:29 +0700)]
Updated Indonesian translation

12 years agoUpdated Indonesian translation
Andika Triwidada [Tue, 16 Oct 2012 00:17:12 +0000 (07:17 +0700)]
Updated Indonesian translation

12 years agoUpdated Polish translation
Piotr Drąg [Mon, 15 Oct 2012 21:36:22 +0000 (23:36 +0200)]
Updated Polish translation

12 years agoBug 684655 - Files in data/sources are not getting translated
Gert Kulyk [Mon, 15 Oct 2012 21:19:13 +0000 (17:19 -0400)]
Bug 684655 - Files in data/sources are not getting translated

12 years agoBug #681314 - Add address dialog should follow composer's view settings
Milan Crha [Mon, 15 Oct 2012 17:50:59 +0000 (19:50 +0200)]
Bug #681314 - Add address dialog should follow composer's view settings

12 years agoFix code examples in ESourceExtension API docs.
Matthew Barnes [Sun, 14 Oct 2012 18:09:13 +0000 (14:09 -0400)]
Fix code examples in ESourceExtension API docs.

The examples did not follow our single-include policy.

12 years agoUpdated POTFILES.in
Piotr Drąg [Sun, 14 Oct 2012 15:45:33 +0000 (17:45 +0200)]
Updated POTFILES.in

12 years agoRename 'online-accounts' module to 'gnome-online-accounts'.
Matthew Barnes [Sun, 14 Oct 2012 14:56:40 +0000 (10:56 -0400)]
Rename 'online-accounts' module to 'gnome-online-accounts'.

For better or worse, GNOME Online Accounts isn't the only accounts
service in town anymore.  Use a less generic module name.

12 years agoESourceRegistryServer: File monitor should set flags on new sources.
Matthew Barnes [Sat, 13 Oct 2012 19:29:11 +0000 (15:29 -0400)]
ESourceRegistryServer: File monitor should set flags on new sources.

ESourceRegistryServer forgot to set WRITABLE and REMOVABLE flags on
new ESources created when a GFileMonitor detects a new .source file.
Consequently the new ESource could neither be modified nor removed,
which causes problems for unit tests that use predefined key files.

12 years agoSilence libsecret unstable API warnings.
Matthew Barnes [Sat, 13 Oct 2012 14:21:28 +0000 (10:21 -0400)]
Silence libsecret unstable API warnings.

12 years agoDeprecate e_source_extension_get_source().
Matthew Barnes [Sat, 13 Oct 2012 13:49:00 +0000 (09:49 -0400)]
Deprecate e_source_extension_get_source().

Replace it with a thread-safe version: e_source_extension_ref_source()

Also use GWeakRef within ESourceExtension, which itself is thread-safe.

12 years agoRemove e_data_cal_view_register_gdbus_object().
Matthew Barnes [Fri, 12 Oct 2012 23:15:44 +0000 (19:15 -0400)]
Remove e_data_cal_view_register_gdbus_object().

Similar to EDataCal, the first thing we do after creating a new
EDataCalView is export its D-Bus interface.  This was a separate and
failable operation.  If the export fails, the EDataCalView is useless.
Perfect use case for GInitable.

Now we pass the GDBusConnection and object path directly to
e_data_cal_view_new(), and if exporting fails the function sets a
GError and returns NULL.

This also introduces a couple accessor functions:

    e_data_cal_view_get_connection()
    e_data_cal_view_get_object_path()

12 years agoRemove e_data_book_view_register_gdbus_object().
Matthew Barnes [Fri, 12 Oct 2012 18:39:40 +0000 (14:39 -0400)]
Remove e_data_book_view_register_gdbus_object().

Similar to EDataBook, the first thing we do after creating a new
EDataBookView is export its D-Bus interface.  This was a separate and
failable operation.  If the export fails, the EDataBookView is useless.
Perfect use case for GInitable.

Now we pass the GDBusConnection and object path directly to
e_data_book_view_new(), and if exporting fails the function sets a
GError and returns NULL.

This also introduces a couple accessor functions:

    e_data_book_view_get_connection()
    e_data_book_view_get_object_path()

12 years agoEDataCalView cleanups.
Matthew Barnes [Fri, 12 Oct 2012 18:12:35 +0000 (14:12 -0400)]
EDataCalView cleanups.

Some minor API changes with this:

    e_data_cal_view_get_text() is dropped.  Instead use:

        sexp = e_data_cal_view_get_sexp (view);
        query = e_cal_backend_sexp_text (sexp);

12 years agoEDataBookView cleanups.
Matthew Barnes [Fri, 12 Oct 2012 17:30:06 +0000 (13:30 -0400)]
EDataBookView cleanups.

Some minor API changes with this:

   e_data_book_view_get_card_sexp() is now e_data_book_view_get_sexp().

   e_data_book_view_get_card_query() is dropped.  Instead use:

       sexp = e_data_book_view_get_sexp (view);
       query = e_book_backend_sexp_text (sexp);

12 years agoAdd e_book_backend_sexp_text().
Matthew Barnes [Sat, 13 Oct 2012 11:48:51 +0000 (07:48 -0400)]
Add e_book_backend_sexp_text().

Similar to e_cal_backend_sexp_text(), returns the original textual
expression used to build the EBookBackendSExp object.

12 years agoe-cal.c cleanups.
Matthew Barnes [Fri, 12 Oct 2012 13:16:39 +0000 (09:16 -0400)]
e-cal.c cleanups.

12 years agoe-book.c cleanups.
Matthew Barnes [Fri, 12 Oct 2012 13:06:58 +0000 (09:06 -0400)]
e-book.c cleanups.

12 years agoEGDataGoaAuthorizer: OAuth 2.0 fixes.
Matthew Barnes [Fri, 12 Oct 2012 01:36:43 +0000 (21:36 -0400)]
EGDataGoaAuthorizer: OAuth 2.0 fixes.

Changes based on Google's OAuth 2.0 Playground.

Authorization type is "OAuth" instead of "Bearer", and no Base64
encoding of the access token is required.

Still no successful test with GOA 3.7, so somewhat flying blind.

12 years agoonline-accounts: Refresh mail account settings on startup.
Matthew Barnes [Thu, 11 Oct 2012 23:41:14 +0000 (19:41 -0400)]
online-accounts: Refresh mail account settings on startup.

When matching GOA accounts to existing EDS accounts on startup, refresh
both the collection and mail account settings.  This is primarily to aid
the transition from XOAUTH to XOAUTH2, but can serve as a means of other
types of migration in the future.

12 years agoEGDataGoaAuthorizer: Prototype OAuth 2.0 support.
Matthew Barnes [Thu, 11 Oct 2012 18:32:53 +0000 (14:32 -0400)]
EGDataGoaAuthorizer: Prototype OAuth 2.0 support.

Untested, but conforms to the specification as well as I understand it.

12 years agovala: Add libedataserver-1.2 to libebook-1.2.deps file.
Jeremy Whiting [Wed, 10 Oct 2012 22:17:10 +0000 (16:17 -0600)]
vala: Add libedataserver-1.2 to libebook-1.2.deps file.
Fixes part of folks bug https://bugzilla.gnome.org/show_bug.cgi?id=685084
so folks ofono backend will only need --pkg libebook-1.2
not libedataserver-1.2 explicitly.

12 years agoDrop oauth package requirement for GOA support.
Matthew Barnes [Wed, 10 Oct 2012 20:17:41 +0000 (16:17 -0400)]
Drop oauth package requirement for GOA support.

On second thought, Camel's XOAUTH SASL module is just using
g_random_int() to generate the "oauth_nonce" parameter.

Good enough, I suppose.  Do the same for EGDataGoaAuthorizer.

12 years agoEGDataGoaAuthorizer: Use GHmac to sign the OAuth parameters.
Matthew Barnes [Wed, 10 Oct 2012 19:28:38 +0000 (15:28 -0400)]
EGDataGoaAuthorizer: Use GHmac to sign the OAuth parameters.

Use GLib's GHmac API in place of oauth_sign_hmac_sha1().

Unfortunately we still have to keep the liboauth requirement for its
oauth_gen_nonce() function.  Considered just copying it out of oauth.c
but it looks rather messy with OpenSSL and NSS dependencies.

12 years agoconfigure.ac: Tweak GOA-related error messages.
Matthew Barnes [Sun, 7 Oct 2012 13:09:48 +0000 (09:09 -0400)]
configure.ac: Tweak GOA-related error messages.

12 years agoDrop the check for libgdata >= 0.9.1 for GOA support.
Matthew Barnes [Wed, 10 Oct 2012 18:04:32 +0000 (14:04 -0400)]
Drop the check for libgdata >= 0.9.1 for GOA support.

We already require libgdata >= 0.10, so the check is redundant.

12 years agoEliminate the assertion in camel_provider_register().
Matthew Barnes [Wed, 10 Oct 2012 14:08:15 +0000 (10:08 -0400)]
Eliminate the assertion in camel_provider_register().

camel_provider_register() should automatically set up the module and
provider hash tables if needed instead of aborting, but the problem
is the GOnce function that sets up these hash tables also calls
camel_provider_register().

Break this cycle by moving the bulk of camel_provider_register() to a
new internal function (provider_register_internal) and calling it from
the GOnce function (provider_setup).

The public camel_provider_register() function is then simply:

    g_once (&setup_once, provider_setup, NULL);

    provider_register_internal (provider);

12 years agoBug 685713 - ESourceRegistry's manager thread aborts on error
Matthew Barnes [Wed, 10 Oct 2012 12:41:28 +0000 (08:41 -0400)]
Bug 685713 - ESourceRegistry's manager thread aborts on error

If the ESourceRegistry's manager thread fails to create a new
GDBusObjectManagerClient during initialization, propagate the GError
back to the caller of e_source_registry_new() instead of aborting the
whole process with a g_error() call.

12 years agoBug 685725 - Handle "muser" attributes in %gconf.xml files
Bastien Nocera [Tue, 9 Oct 2012 01:06:55 +0000 (21:06 -0400)]
Bug 685725 - Handle "muser" attributes in %gconf.xml files

The migration of account data from GConf to key files parses raw
%gconf.xml files.  The parsing code neglected to handle "muser"
attributes in <entry> tags, so that when an "muser" attribute was
encountered, g_markup_collect_attributes() would throw an error an
cause the whole GConf migration phase to be aborted.

12 years agoDon't migrate sources with empty attributes
Dan Vrátil [Mon, 8 Oct 2012 19:48:44 +0000 (21:48 +0200)]
Don't migrate sources with empty attributes

12 years agovala: Don't ignore CalClientSourceType
Colin Walters [Mon, 8 Oct 2012 15:03:07 +0000 (11:03 -0400)]
vala: Don't ignore CalClientSourceType

It's needed by the API after the calendar refactoring.

Signed-off-by: Colin Walters <walters@verbum.org>
12 years agoUpdated Slovak translation
Marián Čavojský [Mon, 8 Oct 2012 14:40:39 +0000 (16:40 +0200)]
Updated Slovak translation

12 years agoRemove e_data_cal_register_gdbus_object().
Matthew Barnes [Sun, 7 Oct 2012 22:23:56 +0000 (18:23 -0400)]
Remove e_data_cal_register_gdbus_object().

The first thing we do after creating a new EDataCal is export its D-Bus
interface.  This was a separate and failable operation.  If the exporting
fails, the EDataCal is useless.  Perfect use case for GInitable.

Now we pass the GDBusConnection and object path directly to
e_data_cal_new(), and if the exporting fails the function sets a GError
and returns NULL.

This also introduces a couple accessor functions:

    e_data_cal_get_connection()
    e_data_cal_get_object_path()

12 years agoDeprecate e_cal_backend_foreach_view().
Matthew Barnes [Sun, 7 Oct 2012 17:34:52 +0000 (13:34 -0400)]
Deprecate e_cal_backend_foreach_view().

Use e_cal_backend_list_views() instead.

12 years agoECalBackend: Prefer e_cal_backend_list_views().
Matthew Barnes [Sun, 7 Oct 2012 16:59:57 +0000 (12:59 -0400)]
ECalBackend: Prefer e_cal_backend_list_views().

Less klunky than e_cal_backend_foreach_view().

12 years agoe_cal_backend_foreach_view: Use e_cal_backend_list_views().
Matthew Barnes [Sun, 7 Oct 2012 16:01:51 +0000 (12:01 -0400)]
e_cal_backend_foreach_view: Use e_cal_backend_list_views().

Holding a locked mutex while invoking callback functions we don't
control raises the potential for deadlock.

12 years agoAdd e_cal_backend_list_views().
Matthew Barnes [Sun, 7 Oct 2012 15:51:11 +0000 (11:51 -0400)]
Add e_cal_backend_list_views().

Alternative to e_cal_backend_foreach_view().

12 years agoECalBackend: Prefer GList over GSList.
Matthew Barnes [Sun, 7 Oct 2012 12:20:10 +0000 (08:20 -0400)]
ECalBackend: Prefer GList over GSList.

12 years agoEGdbusCal: Remove the remove() method.
Matthew Barnes [Sun, 7 Oct 2012 03:48:36 +0000 (23:48 -0400)]
EGdbusCal: Remove the remove() method.

Not used anymore.  Remove the ESource instead.

This also takes out the following backend functions:

    e_cal_backend_remove()
    e_cal_backend_sync_remove()
    e_data_cal_respond_remove()

12 years agoECal: Reimplement "remove" functions.
Matthew Barnes [Sun, 7 Oct 2012 03:40:18 +0000 (23:40 -0400)]
ECal: Reimplement "remove" functions.

Have them just call e_source_remove() or e_source_remove_sync().