Issue 68

Fixed some problems, I found that the VectorMap is not being painted as I expected. Needs a rewrite. And I also found that there's a problem which leaves painting lines behind i made some test and i can say that it's not releated with FlatScanlineTextureMapper, most likely the problem comes from texcolorizer I'll solve this issue in the next commit.

- Fix crash described by BUG# 130104.

Init icon sizes even if the cache itself is disabled. Fixes icons not being loaded when cache is disabled for some reason.

Fix Reminders for imported birthdays from KAdressbook don't work as expected

Thanks for the patch Tobias.

Update the way the type of the account is stored.
Now, we use enums instead of strings.
Add the names of the types to KAccount. They can now be translated.
Also, provide an update script for easy config migration.

Backport commits 688956 and 689037 from enterprise branch (crash fixes for folder moving).

I'm aware that this breaks the message freeze, but an untranslated error message seems to be a better option than a potentially dangerous crash.

Make sure the blanking mode combo is populated.
Thanks to Jiří Jiri Palecek for the patch.

Reverted the backport introduced in 1.0.2 which made K3b not reload the medium for verification and the writing of the next session. Sadly many systems fail to properly detect the newly written tracks.

So far I was not able to determine the reason. It does not seem to be the kernel since I also got a report from a system with a kernel more recent than mine. And here, as always, everything works like a charm.

If anyone has an idea....

My last commit reverted to the old behaviour of reloading the medium before verification. I forgot to mention the bugs that are fixed by this.

"Clear project" now only removed the added data but leaves the settings of the project as they are.

Code cleanup...lots.
Use of icons instead of text to signify encrypted messages.
Slight change in architecture by requiring to-self encryption (so we can decrypt our own messages)

This code doesn't really work, since the PGP block that is sent is a too-large message.
Hopefully this can be worked through.

If there are more than one password dialogs, have one propagate input to the others.
This doesn't fix the backwards message display order.

Use correct port for dcc chat.

Patch by Raphael Kubo da Costa.

* Fix for the info update in the details panel when more then one transfer
* Continually update the transfer time

Patch by Raphael Kubo da Costa

I reduced the size of the graphic elements by 50% so that the pop-up message (KGamePopupItem) is now in comparison 50% bigger and always good readable.

I had also to removed the icon from the pop-up message because it was not looking good, because the size of the scene (QGraphicsScene) usually differs from the size of the view (QGraphicsView) and the icon is a QPixmap (not scalable). And changing this behaviour would be a huge rework and would also have some drawbacks... so I don't want to change this. I think it's looking nice now. :)

Anyway, this should solve the last point of the reported bug. Thanks!

better fuzzy/untr indication (statusbar and optional leds)
actions to go to next/prev fuzzyOrUntr

* first (non-working) draft of mirrors which provides a user interface for browsing the Smoke library.

Lots of changes here:
- Removed the background parser member from Language and created one static instance per application.
- Added functionality to parse all project files on project load.
- Moved some code from Langauge that made more sense in LanguageController.

implement support for scheduling local non-compile jobs.

For example: you can add a symlink named "meinproc4" pointing to icecc to your path, and it will automatically schedule the job and then invoke meinproc4 on it locally.

the advantage is that you can use cmake -j50 without having to worry that there are going to be 50 meinproc jobs killing your machine.

same could be done for moc, btw.

---Start of Support for Extended Entity Relationship Modelling --
Added support for making "categories" and "category widgets".
Widgets can only be drawn currently ( and moved around ).
Lots more to come :)

* Made some more or less final changes to smoke.h
* Code generation works well, however it could be beautified a bit more.
Looks more like a big hack than proper code.
Generation for the type list could also be improved, it still has too many redundant types in it.
* The code generation for smoke modules with parents currently relies on a text file generated together with each module. Use a smoke introspection tool like the one proposed by Thomas Moenicke instead.

do an asynchronous connection to the scheduler:
- helps when the scheduler is down, as the daemon is otherwise locked up then and can't schedule local jobs
- helps when the scheduler is flaky and blocked itself on network (like in the suse network).

* First steps to implement Richard's proposal of a global "class name => smoke module" map.

Annotate the grammar rules with information for building the AST.Use a custom debugvisitor which will print out more information than the generated one.

Add a view manager dialog with no model or real UI yet. Also add an about dialog (needs a handler for the URLs)

Added support for associations from categories to entities and vice versa.
So, essentially, one can now almost draw diagrams for EER models as given in this book
<a href="http://www.amazon.com/Fundamentals-Database-Systems-Fourth-Elmasri/dp/0321122267">http://www.amazon.com/Fundamentals-Database-Systems-Fourth-Elmasri/dp/0321122267</a>;

Still TODO:
* change properties of categories ( currently I've set "disjoint specialisation" as default)
* xmi representation of all the foobar I did :P
* and SQL code generation for EER too probably. This should be interesting.

Start support for templates: Create and use a template-parameter-type and template-parameter-declaration, create du-contexts for template-parameter-declarations, and import those into template-classes and template-functions.

The result: Template-parameters are not unresolved any more.

xmi representation and other related stuff complete, now you can save and open EER documents.

TODO:
*still can't drag a category from listview to canvas.

Provide some predefined reasonable options for Multiplicities in Association Properties Dialog.

Implement a simpler lexer and according grammar. This will move the resolving logic into a higher lever and referenced variables and executed functions in variable values and arguments have to be resolved there by inspecting the value-QString

Flat implementation for the vector map. Changed the approach seems to work fine. I still need to repeat it and correct the 180 -> 180 issue.

The .svg file. As already said it isn't perfect yet.

- Added a GUI Item for the projection. Disabled for now.

Tried to apply the approach suggested by Tackat, there is one problem inheret to the way we handle the vector map right now:

- As we are not correcting Asia's points it may not get displayed on the screen because the right point (which should be close to the right dateline) is represented in the left side of the map hence the polygon which should let us know if Asia should be calculated and printed doesn't really represent Asia.

- Also, for unknown reasons Antartica is not displayed correctly.

Finally I'm trying hard to come up with another approach for the vectormap as i think I might be taking to hard and complex way. Any comments on this?

Windows functions to get the amount of total and free memory.

Patch by Huhn Put, thanks!

Rewrite threading in CameraController:
- use queued signals
- remove CameraEvent and all the associated serializing
- remove the timer-based thread starting and use a proper command queue
- use blocking queued signals to carry out needed UI operations in sync with the non-UI thread (needs testing)

digikamalbums ioslave:
- remove the old code for file operations that was copied from the file ioslave
- use KIO jobs to do file operations, enter an event loop, forward the result.

This is heavily inspired by ForwardingSlaveBase (read: copied from), but this class calls finished before we are finished and some other details, so we have again code copied from kdelibs (but not that much this time, and no low-level system API)

KDE4 port : use QRubberBand in IconView.
Marcel, please look if all is fine for you...

NOTE: still one file to port (tagspopupmenu.cpp) and digiKam KDE4 port will be complete...

KDE4 port. This is the last one to complete digiKam port under KDE4 !

Note : this port is uncomplete. It just to be able to compile and link digiKam as well, and start testing...

Marcel : QCUstomMenuItem implementation need to be ported to QWidgetAction. Still TODO..

digiKam from trunk (KDE4) : and now, digiKam link fine. Time to test !

Marcel, Link fail because 2 main() are found by linker: one in sqlite2 source code and another one in Qt4

I have just renamed sqlite2::main() to main_main(). I hope than all is fine for you...

To kipi-plugins developpers : It will be time to test libkipi and start to port other kipi-plugins to KDE4 (:=)))

Added a new "Hue, Saturation, Value" effect to the "More effects" menu in Kolourpaint

Now the annotations can be dragged and moved after being placed on a page, even if the page is rotated.

Ctrl+right mouse button is what is needed to do that.

At the moment, only line, stamp and geometric annotations can be moved with mouse dragging.

keybindings are now working.

Started to port slideshow. It's full of FIXME for now...

more progress on qt4/kde4 port

KLDraw goes kconfigxt

Refactor thumbnail generation:
Move code out of ioslave into threadimageio, where most of the special wisdom for loading images is already collected.

All ioslaves are now neat and tidy.

I removed the KDE based loading code using directly KIO::ThumbCreator and the associated hacks, and added a PreviewJob in ThumbnailJob. This should work just as well, using well documented API.

Started to implement file operations. Can just show file properties for now.

Implement an argument-hint widget similar to the completion-widget that is attached to the completion-widget and gets it's data through the completion-models. Extract the expandable-item logic from KateCompletionModel into a more abstract model, and use that as base-class for KateCompletionModel and the new KateArgumentHintModel.

Add the roles into CodeCompletionInterface that will be needed to selectively match completion-items to argument-hints, and the roles needed to determine the argument-hint depth.

Huge commit & code refactoring (again)

- interactive scaling using the bottom right corner
- rotation is cached to a pixmap
- coordinates origin is now at the center of the applet
- various performance improvements & bugfixes

On my system it behaves really fast now :-) I hope I've finished the boring part.

port *ahem* kdebase to the new way of doing global shortcuts

Add sensor for disk request queue depth
Delineate sensors for change with Change and change / s with Rate
Remove some unused cruft

use a real SVG with Oxygen style

The new KIO::Connection stuff I promised on kde-core-devel, including Ralf's patch.
I have not fixed all bugs yet, but it's coming in the next few days.

One bug I know is that, when the application disconnects from the ioslave, the slave simply keeps on going (and probably downloads the entire file).

Begun implementation as a QWizard

Work up to date is still valid - just saves some MainWidget uncleanliness

add run command to the desktop context menu

weatherengine: Move more variables to dptr, implemented sourceRemoved() method. set Ions based on KDE's native properties for timezone and measuring units. Remove unused init method.

add parseForecast(), parseShortForecast(), parseForecastTemperatures(), parseUVIndex(). Add uvIndex() to display UV information. Initial capture of 5 day forecast information in debug still.

Astyle code once again. Finish parsing 5 day forecast information. Still not properly displayed in plasmaengineexplorer yet. Rejig some internal API methods to accommodate the 5 day forecasts. Add parseWindForecast(), parsePrecipitationForecast(), parsePrecipTotals() add forecasts() method to store each forcast item.

commit Aaron's case insensitive matching patch.
also included is a regen.sh script that regenerates lex.l and yacc.[c,h] when needed (run by hand after changes to lex.l or yacc.y).

Implement dynamic match-highlighting of argument-hints to completion-items and the opposite way around

Assure that the item size is not exceeded -> the spacing is also respected for texts that cannot be wrapped

A first (working for only me) script applet

Display 5 day forecast information. The format is as follows. For datasources that support a 5 day forecast. They must use 'Short Forecast Day #' for key and then q QString containing the fields: ForecastPeriod|ShortTextOfForecast|HighTemp|LowTemp|PopPrecent.

This goes for any 3rd party datasource that reports basic data back to the weatherengine. For advanced ions/datasources such as the Environment Canada ion, we also have a 'Long Forecast Day #' key.

The applet will be able to parse these keys and display the data appropriately later on

Separate rendering for decoration and window contents, they can have different opacity (opacity only for decoration option).

Add beginnings of kwincompositing config module, meant to replace kwintranslucency one

Add support for a passphrase dialog, the interface to use it is fully asynchronous. Also borrowed some code from kwalletd for reparenting the dialog correctly... doesn't seem to work for me though. Lubos, some help and advices would be welcome here.

Add support for encrypted storage. Request passphrase to the user thanks to the soliduiserver.

Preliminary support for encrypted storage devices... raises interesting usability issues, needs further tuning to get it right. But at least the feature works.

use the underlying default styles to check, whether the current char is valid code or uninteresting part (like comments and strings).

The flow is as follows:
1. get the attribute list of the current highlighting
(it uses the KateRenderer, is there another way to get the list?)
2. get the current attribute
3. use this attribute as index for the attribute list
4. read the property AttributeDefaultStyleIndex and convert it to int
5. now we have the current character's defaultStyle
6. check, whether the defaultStyle is usable
7. if so, count the braces...

Imo, this is much better than the hack in KDE3 (which used the itemData names).
Todo:
* find a good place for a mapping function from attribute<->defaultStyle
* refactor to avoid code duplication ;)
* write a usable indenter (ewh)

Support for yesterday high/low and precip totals. New methods parseYesterdayWeather() and yesterdayWeather() for display to applet. All setData keys are now i18n. Remove some kDebug strings not used anymore.

Work in progress state for replace part of search-and-replace

Initial beginnings of the NOAA ion, so far, it parses the master XML city list and loads the correct city. You can already load multiple US places at once.

icons draw text now, moved corner items to qaction based objects, and support 4 corners. still a bit messy.

I have begun implementing a model-view for the printer selection

Begin work on an accessible color scheme which uses the system-wide colors provided by KColorScheme. Changed the default color scheme to 'Linux' because (a) it was the most popular amongst users in the survey and (b) having used it for a couple of months in KDE 4 it is much more comfortable to use for long periods of use than the glare of a white background. Updated TODOs with a few bugs that need fixing.

Use the standard KDE fixed font (as set in the fonts KDE control module) by default. Aside from correctness this also fixes the font slider in the 'Edit Profile' dialog always being '12' initially instead of the actual point size.

a few visual tweaks: a title, hide/show the edit depending on having a username and password, use the standard background. tried not to do anything toooo drastic so that andred doesn't smack me ;)

Sometimes, one extra command is read from the socket before a suspend takes effect.
This causes a deadlock, because nothing comes from the resumed job and the other one is suspended (FileCopyJob).

Emit a signal to notify that there's data to be read after a resume()

Initial import. Filtering is finished (categories, search, special filters).

Use caching for location URL xml files. add Ion datasource name to identify where the data came from. Finish XML parsing for the NOAA ion and started to add unit conversion support for it. Cleanup code not used.

Theming-related changes:
* clear the old theme before importing a new one;
* have the settings really set for each view;
* simplify some import methods;* fix font stretchfactor error.

Added title/category hack support for Wordpress

Added support for archiving handheld records.

* Subclassed Record, because only handheld records can be archived.
* Added support for archiving to RecordConduit
* Added support for archiving to IDMapper and IDMapperXMLSource.
* Added tests for Archiving support.

Theming: start of zip file support.

More work on the commit/rollback algorithm.

* Added support for commit/rollback of updates
* Added a test to DataProxyTest for commit/rollback of updated records.

Fix layout.
To make the buttons appear like in KDE3, I had to port away from KDialog as KDialog does not seem to support spacers between buttons.

Additionally, make the processed message count available again.

A much better algorithm, based upon Kubelka Munk. It actually mixes almost as expected.

New support for the date and time fields through the KoVariable date plugin.
It still has to be improved for correctness with the standard, but it basically works.

Introduce new class; KoInlineNote. This class is used to store data for foot-, and end notes inside the normal text flow.

The text plugin will use this to layout a footnote at the bottom of the shape it appeared in.

Its currently pretty primitive; but I'm pretty happy about the design as this fills a long standing feature request of allowing footnotes to be printed at the bottom of each column, instead of ignoring the columns.

As a side effect; this allows all applications to use footnotes, not just KWord.

* initial work on a model for the bookmarked configuration manager, extended with a model for filters
* use the model in the filter dialog* create a bookmark manager for each filter, it should be override by filters creating their own configuration objects

Make the cloning of a frame be done using a command, so it can be undone.
Also make the newly created shapes be the ones that are selected.
Plus some smaller fixes in debugging.

Now, the ODF compatibility for the date and time fields should be perfect... Should be... This code needs testing.

Add 'greeking'.

When you zoom text out too much all that is left are some blobs of pixels.
In such a case instead of using the slow drawing text; just draw a rect or a line. This speeds up the painting immensely since the side effect of zooming out < 20% is that you see a lot more text on the screen at the same time. This makes that responsive again on all systems.
Note that I calculate the amount of pixels that the text line would take, so this is the most correct manner of doing this I can think of.

Make printing of images use the high-res ones. I actually make it use the krita-shape while printing, but since the Krita shape does not know about KoImageData (yet) the print is empty. So, this is the typical "it gets worse before it gets better".

* add the interface for filtering metadata* don't edit directly the metadata store, which allow to cancel the editing

Rewrite the way date/time fields are handled. Now, most of the work is done by DateVariable...

I'd like comments about the changes in DateVariable, I'm pretty sure things can be improved there...

Beginning for a pageVariable. Currently, only the page-count field is supported.

New class; a basic printing dialog. Well, its a lot more than a dialog (and really a crappy dialog at the moment, not even a canvas button!).

Current features are;
* Multi-threaded printing so while printing the canvas is fully responsive.
* Do the actual painting code in the main-thread to avoid locking problems.
* Nice progress reporting using the KoProgressUpdater.

This multi-threaded printing is only really useful for multi-page documents, I'll adjust KWord to use it and I'm sure that KoPageApp can inherit from this one to implement printing there with ease.

Initial commit for the new and complete Kubelka-Munk algorithm. It uses a pregenerated matrix and the Simplex Method to calculate a realistic reflectance curve for the current colors. Using the generated curves, we calculate K and S values for each wavelenght, and then a linear combination of these values bring us to the final, realistically acceptable final color. It's slow right now because the Simplex Method is used too often. I'm going to implement a Kis(Ko?)Reflectance10ColorSpace to accelerate the process.
One Algorithm to rule them all!

and some new tools again:
- asymptote
- pdflatex/pdftex draftmode available in ptftex 1.40 bundled in texlive 2007
- 8Bit version of bibtex

implemented basic gradient loading and saving (linear and radial gradients for now)

add a ui file for what could be the "basic" program editor

RGB hdr float colour spaces now work with a profile, which means the post-exposure pixels are now colour managed. This only applies to the software renderer at the moment, the shader version is still to come.

Implement object selection (called Picking in OpenGl slang).

This means that we can detect on which object the mouse hovers. Currently the quad will be painted red if selected.

added backend support for the new metadata method in EngineObserver/EngienSubject for the phonon backend

Elding the items in the playlist. They still overlap sometimes and there isn't enough of a margin between the two sides, but its a start.

Each of the text items are now seperate QGraphicsTextItems, to allow more flexibilty in the future (think editing).

actually, i should probably wait on the loading/unload stuff as aseigo said someone was working on it. applet loading works now (but is stateless).

More work on the kioslave and the MTP backend. Hopefully some basic queries will be usable soon.

Update protocol file. Also, find devices with Solid just by typing pmp:/...

WIP: making Magnatune purchasing use the new Meta types. Very much a work in progress! Feels like I have ported this damn code to different frameworks a 100 times by now... no more new frameworks please! :-)

PMP kioslave now displays a list of supported devices, using library-provided names if possible, if pmp:/ is used as the URL. Clicking an available device uses the appropriate URL (encoded with a Solid UDI) to load up the appropriate backend and initialize it.

Completely rewrite what existed so far of the PMP kioslave so that multiple devices should be able to be handled at once within the same kioslave instance.
Not tested yet...too late at night, but it compiles.

Ability to sign messages

Add option to show local cursor. This is useful for connections with a slow link or with a high latency. This option is only supported by VNC.

Load the local dot cursor in RemoteView. This way we can use it later in other protocols also.
Thanks Pino for the input :)

Add cryptography toolbar to chat window

Changes :
- Added ChangeLog
- Added first part of advanced pref page
- Added new feature to select which network interface to use

Optionally ask user for secret key password at startup, also some changes from QByteArray to QString

Changes :
- When there are no views opened, open the default ones
- Load custom URLs if torrent is private (so that trackerlist merges are not lost)
- Make sure the info widget and search plugin are loaded by default
- When trackerlists of two torrents are merged, make sure TrackerView updates itself

Warn user if they are about to use a protocol unfriendly with encryption
Cleanup includes
Delete unneeded files

implement the background
remove frame around menu/toolbar

accessories-dictionary. New application icon. Deleted dictionary.svgz

video-display. Improved contrast for 32x32 and bigger.

preferences-desktop-wallpaper. New application icon.

oxygen-background test program, can be used to fiddle with the background drawing

preferences-desktop-keyboard. New application icon. Small sizes will need to be redone.

The holes are back meaning:
- QlineEdit
- QSpinBox
- QComboBox (editable mode)
- QListView

are all done

new icon

Reimplement part of the TileSet code to be more simple

Partially implement the new scrollbar

* Make a plugin out of the libarchive-using code

Start working on decoders to further analyze syscalls

* Make it possible to create archives again.
Feels *really* weird, but I'll fix the behaviour when refactoring the code so that MainWindow doesn't depend on ArkWidget

* Plugin for reading zip files using KArchive

* KPDF-inspired "open recent" action

New DCOP call: request the amount of seconds since the last activity with your mouse/keyboard.

dcop rsibreak actions idleTime

Added kbd navigation + "autoscroll" config option + "online kbfx news" config option + updated docs

* Add extraction support to the KArchive plugin. Preview now works for tar and zip archives.

* * Start a rework of libkerfuffle to use KJob

Added a configuration option to enable the debugging shortcuts.
They may be useful for level authors, too, or they may be useful for remote debugging.

Added support for appearing in center (fade in) to KGamePopupItem
This mode can be useful for displaying e.g. GameOver messages

Adjusted test accordingly

Show "Win/Lost" messages using KGamePopupItem.
Perhaps it is not the best way... Feedback is appressiated :)

Corrections to display the elements at the right place.
Changed the color of wrong balls in red. (It was green). After thinking about it, I think it's better like this.
The green laser beam is now a bit bigger.
Adjusted the balls with the question mark to look like the new normal balls.

Revived the action "Check positions". (It was called "Done" in the KDE 3 version). Now, "Solve" is for giving up and "Check positions" is for normal game end.
Added a KPushButton for this action in the left part of the game area.

Motivation for this change was the feedback of different players. Thanks!

Make KGameTheme more flexible, inheriting all theme properties by default. This should make it easier to use the class directly for simpler games, without even the need to subclass it.

Added function to get standard level as string.
Nicolas, hope you don't mind.
And I might screwed your tab indentation somewhere 8-)

Add configuration dialog:
* Support theme changing
* Support custom game settings

Support borders: draw corresponding svg elements around the field

Mauricio, there are still some border elements I don't know what to do with :)
Like #border.insideCorner.* and #border.bay.* (in kmines_oxygen.svgz)
Just ignore them?

Added support of drag-and-drop: it's now possible to move balls and markers over the black box! (Added some APIDOX and updated the change log.)

GUI: Using the KGameDifficulty class to manage difficulty levels in games now adds automatically a combobox in the status bar to display the current difficulty level and to let the player change it.

To get this new feature, nothing to change in the code of the games that already use KGameDifficulty. Main idea was to do like in bovo. (Note that bovo don't use KGameDifficulty yet).

For examples, see Kmines, KBlackBox (and soon Kenolaba).

Feedback is as always welcomed.

- Changed the position of the ComboBox to select the difficulty level: it's now the last element of the status bar.- Removed the combobox of the difficulty "actions" from the toolbar as we now have a nicer one in the status bar. :)

Support restart on difficulty change.

Nicolas, for some reason KGameDifficulty doesn't emit the
'levelChanged()' signal, when I click "Change difficulty level" button in dialog.
Or is it not supposed to emit it? Well, I'd expect it to emit :)

- Added an information message (KGamePopupItem) at the start of the sandbox mode to shortly explain, what it is.
- Added an alternative information message at the end of a game in the sandbox mode.
- Added some comments for the translators (with i18nc) so that they can pay attention at the width of the text for the messages displayed with KGamePopupItem. - Popup messages are now displayed one second more (5 s).

Now, the sandbox mode do realy work: you can put some balls over red balls with the mouse. (Small bug corrected.)

Make text tool work 100% - port transparent text painting to Qt4. It seems that painting text in Qt::color0 works in Qt4 so no need for the elaborate hacks we had before with contrasting colors. Now it's fast enough so that we don't need a text image cache either.

Later: Unfortunately, I did this wrong and a transparent foreground color "sets" pixels instead of painting them, thus acting differently to transparent background pixels. Will need to revert bits and pieces of this commit and adjust some things.

We rename kpTempImage::mayChangeDocumentMask() to the more accurate paintMayAddMask() and also add one for selections (but this latter change will be reverted due to second paragraph).

PodcastReader is now incremental, already a lot faster than with QDom and I'm not even doing early bailout yet.

There is a weird issue where the XML parser seems to be locked up.

speed up the background - thanks fredrikh start to implement scrollbar

Implement resizing of scene elements, using pixmap cache. This increases performance as there is no transformation in QGV.

new poolview for icecream based on a coolo idea

* creating a branch for smoke2

Re-organize devel docs and add some new docs:

depricated: no longer valid, but of potential historical interest

general_1.x: applies to 1.x, and may apply to 2.x

Kst2Specs: design docs for Kst2

I have added 2 pdfs to Kst2Specs to describe datasource/string changes, and to describe kstplot geometry.

What should be the desired behaviour when one clicks on "Expand All" or "Collapse All" ?
This commit includes the listviewitem as part of the expansion or collapse.
I assumed this is the appropriate thing after I did some *cough* congition *cough* tests on some friends :P

Please revert if this is not the desired behaviour.

Create a branch for an experiment I want to do with the parser without breaking kdevelop4

The buttons to control the program are a little bit like those of Kanagram now. The .svg file isn't perfect yet.

- Redo for the GPS tab (Now called "Current Location")
- Disabling that widget as it's meant to be offered just for a particular use case where Marble is meant to provide a dialog

Sorry Andrew - I needed these changes immediately for a presentation yesterday.

add about data to this generator. also makes a pot file possible again. everyone wins. ;)

Give a "better" subtitle for the Accessibility configuration page so it's more "compliant" (but it was also before, IMHO) to the new HIG.

Riactivate the very very very, and I underline very, experimental newstuff support we had before the KNS1 -> KNS2 switch:
- install a newstuff rc config, pointing to the NS1 stuff on okular.kde.org
- show the download dialog for newstuff, properly (hopefully) initialized
- proper linking for the okularpart

Shift + Space is one screen back, says the HIG.

New Notes Plasmoid by lukrop
Please read README file before hacking!

Fancypants new global shortcuts system codenamed "fruity fruitcake" brought to you by Mercurial (so you didn't see it coming), the KDE eV (even though I probably won't collect the aKademy reimbursement now that I have a job), and a very tired Andreas.

The new system tries hard to avoid conflicts and surprises, e.g. shortcuts changing from program start to program start. It has new shortcut setter semantics that make for an "interesting" API but there seems to be no uninteresting API that's also safe in this case.

The missing documentation and the unimplemented function or two are also interesting. Basic things works, though. There is even a unit test with reasonable coverage.

- rewrite the download dialog to not use KHTML.
Using KHTML is bad here because:
- It is a heavy dependency, which is then introduced for all applications that want to use knewstuff2
- KHTML is still full of KDE3 support and it might take until webkit to get rid of that
- using KHTML is dangerous here because it renders arbitrary HTML, Javascript and plugins, so bad providers could easily remote-execute code on the local machine. In addition none of the rendering code actually took care of properly escaping things.
- KHTML wasn't build to be used in hundreds of little parts

The code was pretty bad, it forgot parent parameters everywhere so it caused hundreds of toplevel windows to pop up temporarily on the desktop when it was launched. it also had a way too complex layout with many unnecessary widgets.

the dialog does not quite look that good anymore. but this can be fixed later

remove khtml libs, since those are not needed anymore.
Removes 28 (!) library dependencies from all applications that use khotnewstuff2. In addition, it helps distributions to get rid of qt3support from the default install

Add licenses, as per libmcrypt license.

Interesting thing: it seems that if you don't specify the license version explicitly, the code can be distributed under the terms of any version of the LGPL (same goes for the GPL). Including v3.

Finally commit the soliduiserver. It'll be the proxy for several of our components to request input from the user. First one being a resurrected version of the medianotifier dialog... but more generic now.

It'll be used by our plasmoid.

Add 3D Earth Model applet.

Uses Zack's GLApplet, requires shaders and provides a good way to waste processing power ;-)

Earth imagery by NASA.

move the desktop context menu out of corona. this reflects the design better where different views may want different menus; the dekstop menu only being appropriate for, well, the desktop; other apps may be using Corona for which "desktop" is completely irrelevant.

this also moves the desktop related items out of libplasma, making it a bit more light.

functionality has moved into CoronaView, which will soon be renamed

make this work again. =)

* removed the updated signals from IonInterface; we already have that via DataEngine::updated
* respond to source add/removed signals from the Ions (which are now DataEngines themselves)
* call setData in the WeatherEngine when the ion calls updated

essentially, we're chaining engines. drawbacks right now include:

- the sources between all ions must be uniquely named. this means that right now you can't have "Calgary, AB" from two different weather sources. not sure that matters
- with this method we have 2 copies of the data in memory (one in the ion, one in the weather engine) and cycle through the data in WeatherEngine; not the greatest. not sure this matters either, but it is a place for possible future optimization

make it possible to set the theme per-application, allowing krunner to share plasma's theme by default but amarok to have their own Fun(tm) too

you know what, i really don't care about supporting animated svg's in Plasma::Svg as they break the "this is an image" concept that Plasma::Svg is built around and nobody is using them right now anyways

a KConfigXT XML -> KConfigSkeleton class. should have all the most important items.

the qt sax stuff is -really- nice to work with; i wish i could say the same for KConfigSkeleton. i also wish i could find a way to use a templated function instead of all those nasty QLists and newFoo()s, but i got stuck on deletion in the ~Private. oh well, it works =)

* use the case insensitive matching i added to the trader syntax to allow much more natural matching (e.g. it doesn't have to be case perfect anymore).
* launch the selected service
* use a little qaction subclass to store the KService::Ptr to prevent unecessary repeated lookups
* stop pretending we have options when we don't

now it really is better than minicli ;)

A simple solution for alpha channel hacks - ignore them :).

The 'ignore ARGB visuals' option from Kompmgr probably doesn't make much sense, those (usually old) apps can be run with XLIB_SKIP_ARGB_VISUALS=1 set.

Also remove the alpha clear hack used for decorations - I think decorations instead should be fixed not to "unintentionally" have alpha set.

It's really interesting that KWin has had support for an arbitrary root window, but I don't think it really works these days, and it's probably also not useful at all.

Merge the icon and pixmap cache from /branches/soc-iconcache/

First draft of spruced up APIDOX mainpage.

Once this is right, I'll adapt to other modules for their top-level pages, and to the individual component pages.

Comments welcome.

Initial import of Desktop applet

Make compositing config more user friendly by renaming Compositing to Desktop Effects and putting advanced options into separate dialog.

lisp indenter. is a good example of how indentation scripts [sh|c]ould work.

Proxy model for KDirModel that sorts its contents using natural-sort algorithm (aa10 > aa2).

Separated from Dolphin code. All contributors are OK with GPL->LGPL2 license change.

Add a plasma engine that provides data about the global cursor position.

Possibly controversial. Disabled F1 shortcut for handbook in Konsole because it interferes with terminal programs which need this key. Will review following beta feedback. Remove un-used includes and re-enable DBus adaptor for Session.

* coding style guides are not a cute suggestion but a promise to be fulfilled.
* take QAction*s, which belong to whatever throws them at us
* don't polute the public API with unecessary classes
* avoid subclassing Private classes

i'm still not overly happy with this class as it stands right now.

Initial import of the Twitter applet which lets you update your Twitter status (twitter.com)

make the background something that there is a hope of seeing text on; still not final artwork but now perhaps people will actually bother to use it =)

Added an unused but (I feel) complete driver xml parser, general cleanup and fixes

Copying the DBus interface sucks slightly less than kdepim depending on kdetoys.

Disable date-dependent theme tags for now, as discussed with Cornelius

Don't convert the enum values to int, use some Qt magic to get the enum value names instead and store them.

Warning: If you have been using rev. 689148 or higher, you need to manually run kconf_update before starting KMail or you will lose your account settings. See kmail.upd for details.

Creating branch for KitchenSync port to OpenSync 0.30 API

create a branch for kde3 development. trunk will be ported to kde4

remove kphone stuff we don't need anymore with Decibel

Quick n' dirty port of KitchenSync for the ported version of libqopensync.
It builds - lets ship it!

I introduced another algorithm. It's not yet finished (nor working) because I needed some more data. But probably it will be the one that rules them all. It will support:
- Selection of illuminant (because different light means different colors).
- Selection of precision (basing on the computational power, it will chose less or more "reflectivity wavelenghts" to compute).
- *All* visible colors (thanks to the XYZ colorspace).
It will use Kubelka Munk so we are using together the most powerful academic tools.
Stay tuned...

restore metadata framework (under the strict condition stated on krita's mailing list)

meaning krita can again load exif and iptc tags

Initial import of collaboration project.
Build disabled by default (plugin is not ready for demostration).

Don't show progress window when downloading sites-list

But since the changes to freedb.org, only the default server seems available so maybe it should be removed

<homer>I am so smart! S-M-R-T!</homer>

finally plasma is completely integrated with no patch needed! yay! should work of the box now :)

also, we now use Plasma/Applet as the servicetype as discussed with asiego on panel-devel as well as amarok-devel.

random fixes, there is an annoying crash on exit i'm still working on.

A rather brutal hatchet job to get the services to stop showing duplicate tracks. This leaves the tracks without a genre pointer which is not at all optimal and will have to be fixed somehow. The problem stems from the services allowing an album to have several genres and the meta framework not supporting this

Congratulations Kevin, Solid works :-)

kio_pmp detects a plugged-in portable media player as a libmtp device via Solid, looks up the serial number via Solid (in a HAL-specific manner), and detects whether the device is the correct one, releasing it if not.

Now for the nitty gritty... :-)

REVERT CRAP. (playlist header thingies were never meant to be dragged to the playlist. misunderstanding.)

After three weeks of wondering why not, the playlist now accepts the drop!
The model wasn't correctly setting the tracks to the mime data.

It doesn't actually work, since it appears its trying to do a copy operation, but details details. :)

Credit goes to Mark for figuring out what the problem was. And thanks to Seb for sharing in my frustration.

A new light-colored splash screen.
Emunkki is again looking for feedback to the amarok-promo mailing list.
He plans on adding back the 2.0 when he finds a nice place for it.

Since I'm messing around with JuK's guts, and there's a significant chance that my motivation won't hold out long enough to get things working again, move things over here for now.

The history playlist has always been a bit half-baked. Removing it so that there's one less thing to port to the new playlist stuff.

Render test with fake lighting. If you look at the covers, they get darker the further they're away from the camera.

Such a lighting system would really require per-pixel lighting with pixel shaders, but we do it without shaders by simply coloring the vertexes.

Make the Del key remove selected items from the playlist.

My ulterior motive was to make sure removeRows worked, thinking maybe its why drags within the playlist don't work correctly.

working on snazzing up the last.fm events applet. vboxlayout class needs more work so it collapses properly, but looks waay better than before :)

Maximum limit for smart playlists has been bumped from 1000 to 100000 tracks

use KColorScheme to make the background colors in oxygen style
TODO: don't use KCS for really dim colors, the extra bright it does is good for frames, not so good for the oxygen bg :-)

get user hover/focus colors from KDE; oxygen is now first style in KDE4 to use KConfig :-)

oxygen style now shares code with the win decoration

* Rename arknew to ark :)

Welcome KBlocks! This is my single player Tetris-like game started at last aKademy. It implements most of the technologies used in the kdegames module for KDE4, including user-selectable SVG themes, Oxygen art (by Johann Ollivier Lapeyre), resizable interface, KGamePopupItem, QGraphicsView. It will hopefully provide KDE 4.0 with simple single player Tetris-like fun while we wait for the rewrite of Ksirtet.

Of course, not everything is implemented yet. KScoreDialog is missing, as well as some configuration options and scoring feedback, and better theme alignment. These will come during the next couple of days, but should land by the end of the week. Animations are already implemented, plese direct feedback to me.

Moved KSirtet and the lib libksirtet to playground/games because the game will not make it for KDE 4.0.

However, we'll try to continue to work on it. We hope also to have a good alternative with the new game KBlocks from Mauricio. :)

Approved by Johann Ollivier Lapeyre.

Deleted old code.
Long live new KMines :)

My first redesign attempt. It looks like the most of the graphics is coded into the game :) here is no way to change kube shape or color.

There also is a border abound the cubes and no way to change background
Hope this gets fixed soon.

Removed KBoard from svn.
KBoard (renamed Tagua) can now be found on the git repository:
<a href="http://repo.or.cz/r/tagua.git">http://repo.or.cz/r/tagua.git</a>;