Iterate through CPP elements and use positionView(ListView.Center)
as other approach seems bugguy or slower.
Change-Id: I43879969ccb457166879a156efb482e77ff07d6b
We sort the invites by last interactions. However on swarm, there
is no interaction yet. Add one with the received timestamp given
by the daemon.
(Regression appears because we were showing trust request before
not conversation requests)
Change-Id: Ic7f4d1a5d485325c95a19c2c38ae9c74630829ce
GitLab: #956
Calling a conference and stopping the call will lead the camera
opened. This code seems useless and bugguy.
Change-Id: Iff06b03838a2e13fe378f60f84df96ceb7735f1f
If the avatar for an account is set on a device, it will send
to detected devices on new connections. The client was ignoring
the new avatar if it was not empty, now we just use the last
received avatar
Change-Id: I0b2cde9c193d46014c4d600775743312219b84df
If the timestamp is shown on a contact message, the line was not
taking the full width like other messages.
Change-Id: I94a448acb4f7fd6400365445dd912abc5103ea2a
- isEmojiOnly changed to message containing only an emoji, not an edit or reply.
- adapted size
GitLab: #957
Change-Id: I78388a406a2a0b9c93f1c291850062fd9bf04a54
- Standard time/date is now used
- If today, show local time, otherwise show local date
GitLab: #545
Change-Id: I07f1e706868c725d1c917c473dc0fdbad8d6810f
* extras/packaging/gnu-linux/scripts/release-version.sh: Add script
for deriving (reproducibly) a release version. The format will follow
a 'YYYYMMDD.X' scheme, where 'YYYYMMDD' is the date of the last commit
on the current branch, and 'X' the number of releases tagged that day,
starting from zero. Example: the first stable release with a newest
commit date of 2023-01-02 would have a version of '20230102.0', the
subsequent one '20230102.1', and so on. This is derived by counting
the number of git tags with the prefix 'TYPE/YYYYMMDD', where 'TYPE'
is the release type: either 'stable', 'beta', or 'nightly'.
* extras/packaging/gnu-linux/Jenkinsfile: Call the above new script to
derive the release version to use for tagging.
* extras/packaging/gnu-linux/Makefile: Switch from the previous
'jami_YYYYMMDD.hhmm.COMMIT.tar.gz' tarball filename format to
'jami-YYYYMMDD.X.tar.gz'. The new format is shorter, omits the
unnecessary and confusing commit id, and uses dash instead of
underscore. Also, it contains everything in a top-level directory
with the same name as the tarball (without the .tar.gz extensions),
so that 'jami-YYYYMMDD.X.tar.gz' nicely extracts to 'jami-YYYYMMDD.X'.
With these changes, our release tarballs will follow more closely the
related instructions from the GNU maintainers manual:
https://www.gnu.org/prep/maintain/html_node/Distribution-tar-Files.html
Lastly, this change inadvertently 'fixes' a recent issue where the
'extras/ci/client-qt-gnulinux' subdirectory and its contents would be
moved to top-level in the tarball due to one of the tarball transforms
matching the 'client-qt' in the directory's name, replacing the
'extras/ci/client-qt' prefix with 'client-qt', thereby resulting in
'client-qt-gnulinux' being moved to the tarball's top-level.
* .gitignore: Ignore several packaging-related directories and files.
* extras/packaging/gnu-linux/rules/rpm/jami-daemon.spec:
* extras/packaging/gnu-linux/rules/rpm/jami-libclient.spec:
* extras/packaging/gnu-linux/rules/rpm/jami-qt.spec:
* extras/packaging/gnu-linux/rules/rpm/jami.spec: Update the expected
release tarball filename pattern in the 'Source' field. Also replace
references to previous top-level directory name 'client-qt' with the
new one 'jami-%{version}' where needed.
* extras/packaging/gnu-linux/rules/rpm/jami-libqt.spec: Update the
'Source' field to follow new tarball name consistent with the main
Jami release tarball.
* extras/packaging/gnu-linux/scripts/build-package-rpm.sh: Update the
tarball top-level directory pattern for the new format, now passing
the --wildcards option to GNU Tar. Also use a name for the jami-libqt
source tarball that is more consistent with the main Jami release
tarball.
* extras/packaging/gnu-linux/scripts/build-package-snap.sh: Update the
top-level directory pattern for the new format.
* extras/packaging/gnu-linux/tarballs.manifest: Drop file accidentally
checked into git.
Change-Id: I2695f560c0bc5e61fb99dc0244f8fbb8e7026a62
Because in a swarm you can talk with non contacts, we should at
least retrieve the username when possible.
Change-Id: Ie01cd10cc1231fc4024bd5a08c30c08d69e8e7f0
GitLab: #948
SFPM performs poorly compared to its C++ counterpart in cases where re-filtering often is required. Tests show SFPM taking roughly 10x that of QSFPM in certain cases when the entire conversation needs to be loaded.
Change-Id: I25792f0ab376ae11d42073a548934b743e59fa3e
So that users can know when a message is deleted and can check
previous version like any edited version.
Change-Id: If3c1daeec930bd4b0f359573b971790199397489
Scenario, create a swarm with self.
As a warning, it will show "Can't find contact YOUR_ID this is a
bug, please report" because your account is not in your contacts.
Change-Id: I6f49f8fa77132daf00396b62da9ded470df49f98
If a user join a conference with video enabled, but without any
valid camera and try to un-mute/mute in loop, it will add
new empty medias with incorrect sources causing duplicated participants
in the conference.
Change-Id: Ifa0dc090ca49ae6f9799f1209f3514cf540365da
1. sourceParent and destinationParent were incorrect.
2. Moving always all the message can be incorrect if parentIdx >
msgIdx after a merge
Change-Id: I0ab8eeabb61926d5c0c22d8fed68c60d54178bf6
GitLab: #939
On some cases (like many edition/reply) the atYBeginning is still true
causing the view to not load messages anymore (after retrying it loads
as the beginning)
Change-Id: I9da474224f6fc0a39cf405c8bfe9d5201849ea64
If a window changes its name, we should share the parent process windows.
This can happen with chrome, visual studio code and file explorer for example.
+ Fixes non opening SelectScreen window when Jami language changes.
Change-Id: Ib16992dde08e6a9fa191c9fb1cbc4efd304efe6d
Because docker mount extras/packaging/gnu-linux/packages:/opt/output
the Makefile MUST creates the directory for current user else
the resulting directory will use the Docker user (root), causing
the file to not be removed.
Change-Id: I91dadfa766730d163de4cf091bb1b9b8193f7d24
+ moveMessage was doing useless computation to move messages one
by one even if blocks are supported
+ beginMoveRows/endMoveRows seems to do weird stuff if all the rows
should be moved causing some crashes. beginReset in this case
doesn't trigger this weird behaviour and is more efficient.
Change-Id: Ia4eb3cdbbe74bf9215fb673cb9af571f67225ffe
