There are 72 selections this week
Changed: merge ViewProfileSyncController and ViewProfilesController into ViewProfileController
ViewProfilesController used KXMLGUIClient::plugActionList(...), but that is flawed, as these plugged
actions get lost on any change/update of the XMLGUI clients
So using a KActionMenu (and thus a single controller) to have the full control
about the contained actions in the submenu.
Don't randomly remove the reviewed patches
Only remove them if they are a temporary file created for a command.
This class should be probably splitted in two because the command case
is clearly a specific case of the file, at least in this implementation.
Aggregated KSUSerDB to KStarsData.
Fixed initialization queries.
Implemented adding of observer.
Saving and loading worksheets should work now.
Fix saving/loading of graph structure additional properties.
Add serialization hook for data structures extra properties.
Add unit test to check serialization of graph properties.
fix crash when trying to zoom to discrete zoom level for texture-less maps after having used a texture-based map
After having used e.g. OpenStreetMap, the TextureLayer isn't cleared completely. As a result, it will return seemingly valid values in some cases, but will crash in others. In this case, the TextureLayer used to return a valid zoom level such that MarbleMap was fooled to ask it for the next "sharp" radius, leading to the crash.
A more reliable way is to check whether the TextureLayer participates in rendering. If not, it shouldn't be asked for the next "sharp" radius.
This patch is a quick fix wich comes with unit tests to prevent future regressions. In the future, the TextureLayer should be properly cleared or even recreated/deleted whenever the map theme changes.
centralize interaction notification handling
that way we don't have to handle it all over the place (which we didn't,
which made the chooser dysfunctional).
kinotify: Replace manual directory iteration code with QDirIterator
The existing code has been causing crashes since KDE 4.5, and is very
hard to maintain since no one really understands it properly.
Funnily enough, this version detects a lot of directories that the old
code did not. Weird.
make PlasmoidProtocol::loadFromConfig behave properly in all situations
this is done by greatly simplifying PlasmoidProtocol::loadFromConfig and putting a check
in PlasmoidProtocol::addApplet to see if the host already contains the applet. this is needed
anyways since addApplet can be (and is) called from the manager.
this also adds a check such that if for whatever reason more than one instance of the same
applet was present in this system tray (something which the current design expressly prohibits
though i do wonder if it doesn't make sense to have more than one, e.g., weather widget at a time)
then delete the stray config group, preventing endless #s of configs for the same applet
piling up in the config should anything have gone wrong in the past
nowplaying: prevent the controls from leaving the widget
The minimum width was previously such that the controls could extend
outside the right edge of the widget.
Don't hang on very long queries.
If listing a token takes an excessive time, give the user the ability
to continue or abort the query.
Note that if we had threading support, we might have a more elegant
try to get the date updating properly
the updateDate and paintIcon code all work properly, which leaves us with
something odd with the qtimer .. i've probably done more than necessary, but...
this does have one deliterious effect: it will wake up once per hour instead of trying
to only wake up once per day. so be it.
Fix Bug 301957 - Mail composer ignores identity field when opening saved
Fix Bug 301663 - after creating a imap account (using assistent) the
folder list is still empty
Fix Bug 300868 - No feedback when wrong POP password is entered.
Now we can see it in notification.
Support runaround distance individually on all 4 sides.
Treating them all as one gave layout issues
Related ui is updated too
Anchoring paragraph area refer to only the area, and doesn't change with indentation or dropcaps
Prevent crash after 87dd80da75f5b2425ff8f04954a87650cb7be790
If inserting a textshape, the qtextdocument is now modified which made a rootarea to be created
We need to explicitly set the root area to 0 whenever setting the document. Even if it's the same
Fix issue with text list indenting and numbering
With this patch increasing the list level by pressing tab or the "Increase Indent" button the display level is also increased by one.
Fix saving of style child elements
When saving style child elements the child elements where always saved into the default property. However this is not correct as the output of Jos tests shows
[01:53:45]: kofficetests-odp (17s)
[01:53:45]: [kofficetests-odp] odf/odp/114096.odp
INVALIDSTYLESXML: (unknown file):267: error: element "style:tab-stops" not allowed here; expected the element end-tag or element "style:background-image", "style:columns" or "text:list-style"
INVALIDCONTENTXML: (unknown file):1061: error: element "style:tab-stops" not allowed here; expected the element end-tag or element "style:background-image", "style:columns" or "text:list-style"
(unknown file):1081: error: element "style:tab-stops" not allowed here; expected the element end-tag or element "style:background-image", "style:columns" or "text:list-style"
The problem is that the stye:tab-stops are added to the graphic properties but should be added to the paragraph properties.
This patch adds the possibility to define the property the child elements should be added too.
MediaDeviceCache: remove polling, solid events should suffice
This fixes a bug where Amarok (very probably needlessly) polls solid
for all devices every single second (!!!) just to detect whether some
unmounted paths become mounted or vice versa. This should not be needed
at all, solid should notify us about everything.
Let's hope this doesn't cause subtle bugs, but several people already
tested it with success [thanks, Sam]. (probably none of them with older
Play the correct track when inserting tracks into a sorted playlist
When Playlist::Controller::insertOptioned() is told to play the track
that has been inserted, it previously failed to account for the
relocation of the tracks when inserted into a sorted list. The last
track in the list would end up being played, rather than the first
mostly fix streamreader for refactored pipeline
- cleanup & merge stop with unlock
- fix streamer unlocking
- return if the reader got unlocked before the lock was acquired
- always return right away, to reduce waiting time for possibly waiting
- turn on seekability
- make streamreader a pipeline member supposed to be reused after stop
gets thrown away on setting a new source because otherwise libphonon
likes to assert out with the present streamer implementation
- stop reader also on stop()
Fix inconsistent scrobbling with GStreamer backend
With the GStreamer backend, the trackChanged signal seems to be emitted
twice on track change, which in turn sometimes triggered a consistency
check in the scrobbling code and sometimes causes tracks not to be
scrobbled. About 1/5 (variable) tracks would be successfully scrobbled.
This should not affect scrobbling with the VLC backend, which emits
This is a temporary hack until Matěj's statistics rewrite is complete.
Fix pan & zoom effect (paste, edit after reloading project)
Apply a Plasma-like style to the new-tab-page
This commit makes previews on the new-tab-page use pure css instead
of a background-image and replaces the grey background of the
bookmark groups by a fancy one, looking very similiar to Plasma's
It also fixes a few minor optical dissonances on the new-tab-page.
Fix connection list if the plasmoid is launched before the kded module.
Report correct encryption state on FreeBSD
This fixes a problem when KNemo reports a connection as encrypted if a
WEP connection also accepts unencrypted packets. Patch contributed by
Fix crashes in the engine by readding the service provider ID
- Bug was introduced in commit d93b82774d351bb55efdc0abb8f5a42cd60bf154
- Fixed by readding the service provider ID to data sources
of the engine
- It was left TODO in the bigger commit
- Prevent crashes if this happens again, print out a warning message
Sometimes clients did not write into data base when they should
This happened when two clients used the same path for their database,
even if it was on different computers.
Until now, identity of data bases was checked by path and IP address,
but that worked only in my local network.
Now each kajongg data base has a unique ID.
Background: It is legal having two kajongg instances running on one
client - and if that happens, they will use the same database. So only
one of them should save game info in that data base. The server says who
Make it possible to review remote patches
Using the Patch Review UI, make it possible to review a non-local patch.
It's useful so that it makes it far easier to review patches coming from
pastebin or the reviewboard.
Improve the last commits history dialog
Handle the history keeping of the VCS patches from VCSCommitDiffPatchSource.
This way we don't forget to do it somewhere random.
Added animations; plus performance fix
- Eps images are not re-rendered on every goemtry change anymore, but
just when the scale changes. Therefor resizing the window works with
decent speed again.
- There is some code to animate changes in the worksheet. I am not yet
quite sure which changes should be animated so that the experience
is consistent and the user is not annoyed by too many (slow)
animations. But I think that the animations can improve the
experience of the user.
Show a warning if annotations will not be saved to PDF.
The only affected generator is poppler (because it's the only one with
Since version 0.20, Poppler can save annotations to PDF. If the user is
using an older version, this patch warns him that annotations are not
Skeleton for Photivo Integration Plugin
* Based on HelloWorld example
* Included Photivo icon
* Plugin is listed as item in "Tools" menu
Render the checkboard background texture
For this mode the normal shader is used, but different vertices
to make a repeating pattern by setting texCoords which are
larger than 1.0.
TODO: as with solid color, the first selected image has black
background, gwenview expert needed.
OpenGL rendering in RasterImageView
Beginning of OpenGL rendering in gwenview. Current state:
if OpenGL rendering mode is selected the image is rendered
through an OpenGL shader. The scaling of pixmaps is no longer
There are still some areas to improve:
* Better detection of OpenGL availability
* Better detection whether OpenGL is working (creation shaders)
* Some regressions like compare mode not working
* Updating the texture on changes
Implement the battery percent overlay option
This commit fixes the toggling of the config
option and showing/hiding the overay accordingly.
Also, show the overlay on hover when on desktop
even if the option is off.
(This is how the old applet behaved)
Renaming: Don't select unknown extensions
If a file with an unknown extension gets renamed, it is assumed
that the letters after the last point represent the extension and
won't be selected.
Added support for Like, Comment and Application objects
Start to implement import custom header
Implement most of the CPU-based OCIO path
The display filter (only one for now, and I'm not happy about the
api for setting it, it seems to me that the lcms color correction
should also be a "display filter") fitlers the pixels using ocio, which
are filtered using lcms to compensate for the monitor.
The only thing not yet done is correctly setting the ocio display
Even so, it doesn't seem to crash...
Additional quick action buttons when mouse is over items, in Stage
Stage slides sorter has +/- buttons that appear when you move the mouse
over an item to allow you select or deselect multiple items without
using the keyboard. This patch add extra actions (remove/duplicate
slide and start slideshow from here) like gwenview, digikam quick actions.
GUI: Hover buttons for quick actions in Stage Slide Sorter
Use an infinite canvas for kopageapps
KoPageApp applications Stage and Flow used to have extra space reserved around the page/slide
so objects could be put there. However that made it really hard for the user to work with.
This patch changes that so there is no space around the page/slide. But instead you can drag
objects outside the page/slide without any limit at all.
Add the keywords "review","keyword","file" to the kbibteximporter
Now the keywords will be imported as:
"review" : same as a note
"keyword": same as "tags or keywords
"file": as local file same as url
Added a method KisCanvas::setMonitorProfile()
This method sets the monitor profile of the canvas and updates
all its caches in a proper thread-safe way.
What is *not* done:
1) Handling of KisCanvas2::Provate::monitorProfile/renderingIntent.
They are currently set in quite a random way in different methods.
There should be some order in it.
2) Handling of the rendering intent in the openGL canvas, although a
variable was added for it.
Changed the Note tab in the sidebar widget
Moved the cited sources box to the Content tab
Introduced a textedit box below the annotation list
This allows to support a workflow which is centered around working with one note file per publication
it is not necessary anymore to open a popup dialog to endit the note directly.
The content of the note will be saved automatically
Support various cases where paragrah borders should not be merged (differnt horiz extent
or flagged to not merge)
Add support for wheel events and try to improve shortcut recognition.
Finally implement formatting plugin
It took me forever becuase of some annoying bug that was really
hard to find. (Turns out I was accidentally trying to match a
regex special character)
It also adds new tests.
Store output separated by projects, add clear output action
- Previously the output dock widget has shown all output from the
currently active project and didn't clear the output when switching
to another project
- Now the output is stored for each project
- The output dock widget shows only output for the active project now
- Can now collect output for multiple projects at the same time
- Add a clear action to the output docks context menu
First implementation of the batch run
# ==[ Blank: Follow the Subject with a blank line, do NOT remove ]=====|
# ==[ Details: Describe what changed and explain why it changed]=======|
Still not functional...
# ==[ Fields: Uncomment and edit where applicable ]====================|
# --[ Close bug in bugs.kde.org as fixed ]-----------------------------|
#BUG: <bug number>
#FIXED-IN: <optional release version>
# --[ Add to release changelog optionally close wish in bugs.kde.org ]-|
#FEATURE: <optional bug number>
#FIXED-IN: <optional release version>
# --[ Copy commit message to a bug or wish in bugs.kde.org ]-----------|
#CCBUG: <bug number>
# --[ Copy commit message to an email address ]------------------------|
# --[ Close a review on git.reviewboard.kde.org as submitted ]---------|
#REVIEW: <review number>
# --[ Notify documentation team of user visible changes ]--------------|
# --[ Notify Commit Digest team of something interesting ]-------------|
Use templates provided by KDevPlatform instead of shipping includes with every template
SqlPlaylist: don't observe track and re-save whole playlist on track change
No other playlist does that and this causes problems when someone has
really big playlists and slow connection to external MySQL db.
Not updating track metadata should not hurt, these are immediately
replaced by real ones by MetaProxy::Track. Moreover for uidUrl updates
there is SqlTrack::updatePlaylistsToDb() (I don't however know whether
it works as expected). Acked by Bart on the bug.
Increase the ResourceManager cache size to 10000
This way one can easily show more than 1000 results in one time.
Ideally, one should manage the cache better, but that will have to be
done in the next release.
No event notifications when Resuming/Suspending indexing of resources
Indexing is a background process that the user shouldn't be aware about.
We should just work, and not annoy the user about stuff like this.
Added a config option in nepomukstrigirc just in case anyone wants it.
Add ShowSuspendResumeEvents = true in [General]
Remove the GraphMaintainer
The GraphMaintainer's job is to remove old empty graph which have
somehow appeared due to buggy code. They do not really harm anyone, but
this approach of removing them on startup really makes virtuoso cpu
usage spike for a very long time.
It would be better to have some kind of "Data Cleaning" application for
Nepomuk which would migrate old data and clean up invalid data like
Create skeleton for ImagesFilter
Should probably save something like this somewhere so I don't make
silly mistakes doing this in futre.
Remove the shared channel announcement in queries again.
I never found a string I liked, I haven't found a satisfactory
way to deal with the race condition during auto-join, and there
were some complaints about the additional highlight vector (also
potentially interacting badly with the actual query messages).
Looks like a better idea might be to show the current shared
channels in the "topic" area of query tabs. This is subtly diff-
erent from having a timestamped record of the shared channels at
query open time, but at least gives the area more to do and avoids
Additionally, removing this again fixes a regression that caused
query tabs not to light up with the message highlight when newly
opened (probably because the command message event overwrote the
highlight event before the event loop cycled enough for priority
sorting to be possible, haven't really worked through it).