| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Testing Done:
Compiled and ran tests in valgrind.
Reviewed at https://reviews.imfreedom.org/r/2374/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix two leaks in account parser:
```
512 bytes in 4 blocks are definitely lost in loss record 30,342 of 33,863
at 0x484378A: malloc (vg_replace_malloc.c:392)
by 0x484870B: realloc (vg_replace_malloc.c:1451)
by 0x498071F: g_realloc (gmem.c:201)
by 0x499A343: g_string_maybe_expand (gstring.c:92)
by 0x499A3BF: g_string_sized_new (gstring.c:116)
by 0x499A7AA: UnknownInlinedFun (gstring.c:176)
by 0x499A7AA: g_string_new_len (gstring.c:167)
by 0x4D577C6: purple_xmlnode_get_data (xmlnode.c:460)
by 0x4CCD5D8: parse_account (accounts.c:338)
by 0x4CCD98B: load_accounts (accounts.c:449)
by 0x4CCE234: purple_accounts_init (accounts.c:667)
by 0x4CE3AC2: purple_core_init (core.c:174)
by 0x48A380A: pidgin_application_startup (pidginapplication.c:820)
512 bytes in 4 blocks are definitely lost in loss record 30,343 of 33,863
at 0x484378A: malloc (vg_replace_malloc.c:392)
by 0x484870B: realloc (vg_replace_malloc.c:1451)
by 0x498071F: g_realloc (gmem.c:201)
by 0x499A343: g_string_maybe_expand (gstring.c:92)
by 0x499A3BF: g_string_sized_new (gstring.c:116)
by 0x499A7AA: UnknownInlinedFun (gstring.c:176)
by 0x499A7AA: g_string_new_len (gstring.c:167)
by 0x4D577C6: purple_xmlnode_get_data (xmlnode.c:460)
by 0x4CCD610: parse_account (accounts.c:343)
by 0x4CCD98B: load_accounts (accounts.c:449)
by 0x4CCE234: purple_accounts_init (accounts.c:667)
by 0x4CE3AC2: purple_core_init (core.c:174)
by 0x48A380A: pidgin_application_startup (pidginapplication.c:820)
```
and one in the demo protocol:
```
16,008 (56 direct, 15,952 indirect) bytes in 1 blocks are definitely lost in loss record 33,622 of 33,863
at 0x4DDF1EF: g_type_create_instance (gtype.c:1909)
by 0x4DC4C1F: g_object_new_internal (gobject.c:2228)
by 0x4DC6972: g_object_new_valist (gobject.c:2567)
by 0x4DC6FCC: g_object_new (gobject.c:2040)
by 0x57CF444: g_resource_open_stream (gresource.c:764)
by 0x1CAA54A2: purple_demo_contacts_load (purpledemocontacts.c:233)
by 0x1CAA4CD0: purple_demo_connection_connect (purpledemoconnection.c:43)
by 0x4CE1540: purple_connection_connect (connection.c:955)
by 0x4CC7F5E: purple_account_real_connect (account.c:179)
by 0x4CC85B0: no_password_cb (account.c:355)
by 0x49784C7: g_timeout_dispatch (gmain.c:5007)
by 0x4977CBE: UnknownInlinedFun (gmain.c:3444)
by 0x4977CBE: g_main_context_dispatch (gmain.c:4162)
```
Testing Done:
Ran in valgrind and confirmed above leaks were gone.
Reviewed at https://reviews.imfreedom.org/r/2299/
|
|
|
|
|
|
|
|
|
|
|
| |
I can't think of a good reason why we shouldn't expose the error property on
Accounts. But I would like to be able to set the error from the UI when there
is an error in the settings for example.
Testing Done:
Built the docs, ran the tests, used the prune connection action in the demo protocol plugin and verified everything was still working.
Reviewed at https://reviews.imfreedom.org/r/2216/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
PurpleProtocolActions
We also propagate the PurpleProtocolActions::actions-changed signal from the
PurpleProtocolManager::account-actions-changed signal if the protocol implements
the PurpleProtocolActions interface.
Testing Done:
Build the docs and ran the unit tests.
I also put a temporary action in the demo protocol plugin and called `purple_protocol_actions_changed` after the demo protocol connected. I will have another pull request that makes this a permenent action, but I want to convert the demo protocol plugin to subclass PurpleConnection before doing so as that will be accounting of everything *much* easier.
Reviewed at https://reviews.imfreedom.org/r/2168/
|
|
|
|
|
|
|
| |
Testing Done:
Compiled and ran the unit tests.
Reviewed at https://reviews.imfreedom.org/r/2133/
|
|
|
|
|
|
|
| |
Testing Done:
Touched all source files and compiled; saw no new warnings.
Reviewed at https://reviews.imfreedom.org/r/2122/
|
|
|
|
|
|
|
|
|
|
|
| |
This API has been needing a rewrite for as long as I've been working on Gaim /
Pidgin and nothing has come of it. Therefore, I'm am removing it for a much
more simplified approach in the future.
Testing Done:
Sent some messages with the demo protocol plugin.
Reviewed at https://reviews.imfreedom.org/r/2065/
|
|
|
|
|
|
|
|
|
| |
The callback expects the account as its callback data, but none was passed.
Testing Done:
Deleted an account and saw no failing `G_IS_OBJECT` critical.
Reviewed at https://reviews.imfreedom.org/r/2041/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
"account-status-changing" signals.
Also sorted the account signal changes in ChangeLog.API.
The account-error-changed signal is easily replaced by the notify signal for the
new property, and account-status_changing wasn't used by anything.
Testing Done:
Used the account actions of the demo protocol plugin to disconnect the account without issue.
Reviewed at https://reviews.imfreedom.org/r/2036/
|
|
|
|
|
|
|
|
|
| |
and remove the old purple signals
Testing Done:
Connected a Bonjour and Demo account without issue.
Reviewed at https://reviews.imfreedom.org/r/2034/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
PurpleAccountManager::account-changed
This new signal supports details and works just like notify does on the account
instances, but it will be emitted for any account that the manager instance
knows about which means we can remove the old purple signals.
Testing Done:
Enabled and disabled accounts in Pidgin via the menus to make sure things were working right, and verified via the console that the accounts were being enabled and disabled.
Ran Finch, but I ran into the infinite loop bug in the buddy list clean when disabling an account which is a known issue.
Reviewed at https://reviews.imfreedom.org/r/2032/
|
|
|
|
|
|
|
|
|
| |
See [this development thread for a future Fedora change](https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/thread/CJXKTLXJUPZ4F2C2VQOTNMEA5JAUPMBD/) or [the proposed change page](https://fedoraproject.org/wiki/Changes/PortingToModernC). These may be made stronger errors in GCC 14.
Testing Done:
Configured with `-Dc_args='-Werror=implicit-int -Werror=implicit-function-declaration -Werror=int-conversion -Werror=strict-prototypes -Werror=old-style-definition'` and compiled.
Reviewed at https://reviews.imfreedom.org/r/1974/
|
|
|
|
|
|
|
|
|
|
|
| |
were deserialized
This is a short term fix until we rewrite the serialization method.
Testing Done:
Set a break point on `pidgin_account_connect` and verified it didn't get call from `purple_accounts_init`.
Reviewed at https://reviews.imfreedom.org/r/1966/
|
|
|
|
|
|
|
| |
Testing Done:
Compiled
Reviewed at https://reviews.imfreedom.org/r/1914/
|
|
|
|
|
|
|
|
|
|
|
|
| |
This will be able to be to set the a user so that an account so that an account
with an optional password will require a password for login.
Testing Done:
Compiled
Bugs closed: PIDGIN-17688
Reviewed at https://reviews.imfreedom.org/r/1876/
|
|
|
|
|
|
|
|
|
|
|
| |
This is to help facilitate the new blist api that's still being designed but we
wanted to use this name for an object. A later review request will rename
contact.[ch] which was avoided for now as it would make the diff harder to read.
Testing Done:
Ran and verified that contacts showed up.
Reviewed at https://reviews.imfreedom.org/r/1770/
|
|
|
|
|
|
|
| |
Testing Done:
Compile only.
Reviewed at https://reviews.imfreedom.org/r/1812/
|
|
|
|
|
|
|
| |
Testing Done:
Compiled, but didn't test on windows. As such I left all the networking stuff untouched as that's where windows will really need it and that can be done in another rr.
Reviewed at https://reviews.imfreedom.org/r/1764/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
replaced it.
Also clear account errors when we attempt a new connection which made clearing
the errors after the account has connected redundant.
Testing Done:
Caused the same account to get disconnected between pidgin 2 and 3 to verify notifications were working as expected.
Bugs closed: PIDGIN-17640
Reviewed at https://reviews.imfreedom.org/r/1521/
|
|
|
|
|
|
|
|
|
| |
account was enabled
Testing Done:
Ran and verified the demo account connected without issue.
Reviewed at https://reviews.imfreedom.org/r/1460/
|
|
|
|
|
|
|
| |
Testing Done:
Enabled and disabled the demo account a bunch of times, including restarts, and the like.
Reviewed at https://reviews.imfreedom.org/r/1458/
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As well as:
* Fix the namespace of the PurpleProxyType members.
* Remove the GNOME proxy detection code.
* Remove the Windows proxy detection code.
Testing Done:
Created an account and messed around with the proxy settings and verified they were saved and restored.
Reviewed at https://reviews.imfreedom.org/r/1368/
|
|
|
|
|
|
|
|
|
| |
Another straight search-and-replace for 'easy' review.
Testing Done:
Compile only.
Reviewed at https://reviews.imfreedom.org/r/1371/
|
|
|
|
|
|
|
| |
Testing Done:
Compiled and ran
Reviewed at https://reviews.imfreedom.org/r/1222/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I suspected this wasn't necessary for awhile, but finally tried it and I can't
seem to find anything that broke when I removed it.
Fixes PIDGIN-17566
Testing Done:
Connected a Bonjour account, set the status to `away` with a message, stopped Pidgin, verified that `status.xml` contained what looked correct, then restarted Pidgin and verified that the status was set to the previous status before exiting.
Added an XMPP account, set the status to `do not disturb` with a message, closed pidgin, verified `status.xml` looked mostly correct, then restarted Pidgin and verified that the statuses were set to the previous statuses for both accounts.
Bugs closed: PIDGIN-17566
Reviewed at https://reviews.imfreedom.org/r/1190/
|
|
|
|
|
|
|
|
|
| |
Testing Done:
* Verified that existing accounts were given ids.
* Verified that ids persisted in accounts.xml.
* Verified that new accounts got ids and were persisted in accounts.xml.
Reviewed at https://reviews.imfreedom.org/r/1161/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This replaces most of the purple_accounts_ API, but not quite everything. The
functions that have been replaced have been marked as deprecated and libpurple
has been updated to use the new API.
Testing Done:
* Connected an XMPP account
* Verified chat and ims worked
* Verified that the saved statuses window didn't crash
* Ran the new unit tests.
Reviewed at https://reviews.imfreedom.org/r/1137/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Both wincred and secretservice return an error if you try to delete a password
that they don't know about. We don't care about this scenario and just want to
make sure the password is gone, so we changed the expected behavior to be as
such and updated wincred and secret service for the change.
I also fixed the NULL deference which is how I stumbled across the issue in the
first place.
Testing Done:
* Verfied the issue was fixed in secretservice.
* Verfied the issue didn't exist in kwallet.
Reviewed at https://reviews.imfreedom.org/r/901/
|
|
|
|
|
|
|
| |
Testing Done:
Full build after a `meson --wipe`
Reviewed at https://reviews.imfreedom.org/r/863/
|
|
|
|
|
|
|
| |
Testing Done:
Compile only
Reviewed at https://reviews.imfreedom.org/r/791/
|
|
|
|
|
|
|
|
|
|
|
|
| |
I've disable the building of the internal keyring for now but left it so it
can be referenced when it is ported to the CredentialProvider API.
Testing Done:
Ran Pidgin and verified the preferences window worked.
Bugs closed: PIDGIN-17486
Reviewed at https://reviews.imfreedom.org/r/624/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Testing Done:
* connected account with wallet locked, verified that we weren't prompted for a password until the wallet was unlocked.
* connected account, didn't save password, made sure it connected and wasn't stored in kwalletmanager5
* connected account, saved password, made sure it connected and verified the password was stored in kwalletmanager5
* reconnected account, made sure the account connected without prompting
* reopened pidgin, made sure the account connected without prompting.
* disconnected pidgin from kwalletmanager5, re-connected account, verified it reconnected via debug and kwalletmanager5
* force closed the wallet in kwalletmanager5, re-connected account, unlocked wallet, verified no password prompts and that the account connected.
* removed the account and verified the password was removed from kwalletmanager5
Bugs closed: PIDGIN-17488
Reviewed at https://reviews.imfreedom.org/r/575/
|
|
|
|
|
|
|
|
|
|
|
| |
This feature has more or less been stalkerware since its inception and the
only useful feature that came of it was offline message emulation which isn't
necessary for most networks nowadays.
Testing Done:
Compiled and ran locally.
Reviewed at https://reviews.imfreedom.org/r/322/
|
|
|
|
|
|
|
| |
Testing Done:
Connected with bonjour (no password), irc (optional password, non set), and xmpp (password required, not saved, entered manually, not saved) and verified that all of them connected properly.
Reviewed at https://reviews.imfreedom.org/r/289/
|
|
|
|
|
|
|
|
|
| |
Remove some migration code for an icq server since we don't ship and icq plugin anymore
Testing Done:
Compiled, verified no new warnings.
Reviewed at https://reviews.imfreedom.org/r/181/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
make ggp_date_from_iso8601 use a GDateTime instead of GTimeVal and have it return a gint64
Wrap a signal callback in a G_CALLBACK like it should be
Remove some unused variables
Only include sysctl.h on platforms where we need it. It's deprecated under gcc now.
Testing Done:
Compiled on freebsd 12.0, netbsd 9.0, openbsd 6.7, mac, and linux without issues.
Reviewed at https://reviews.imfreedom.org/r/69/
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix some issues that were discovered while working on a flatpak
Testing Done:
built a flatpak (that'll be a separate review request) and ran into some issues..
First was bonjour_ft wouldn't build, not sure how/why that isn't happening elsewhere.. but whatever..
Next was the network monitor stuff I moved us to way back in december of 2015 works, but under flatpak the network is initially unavailable but then becomes available. A bunch of our code was __only__ checking network availability at startup and then not listening for changes later. This code is really gross right now, but a lot of it is going to need some major refactoring where as this effort was just to get this going right now.
Reviewed at https://reviews.imfreedom.org/r/66/
|
|
|
|
| |
with the repercussions of that.
|
|
|
|
| |
some pieces to purpleprivate.h
|
|
|
|
| |
purple_status_set_active_with_attrs_list
|
| |
|
| |
|
|
|
|
|
|
|
| |
* `purple_blist_get_buddy_list` moved to `purple_blist_get_default`
* `purple_blist_get_root` copied to `purple_blist_get_default_root`
* `purple_blist_get_root` takes a list argument now
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now that the purple_timeout_* functions internally just use their
GLib equivalents, we can just replace them with the GLib equivalents.
This patch replaces purple_timeout_* functions with
g_timeout_add{_seconds}() and g_source_remove() functions.
There are probably places in this which could use g_timeout_add_full()
or g_idle_add{_full}(), but this was a straight replacement to keep
the functionality identical for now. Further patches can update to use
these instead if desired.
|
|
|
|
|
|
|
| |
It's like "account-status-changed", but emitted before the status change
instead of after.
Instantbird ticket: https://bugzilla.mozilla.org/show_bug.cgi?id=954403
|
| |
|
|\ |
|
| |\ |
|