7th December 2008by Danny Allen
|Commits||2886 by 245 developers, 6065 lines modified, 1919 new files|
|Bugs Opened||551 in the last 7 days|
|Bugs Closed||769 in the last 7 days|
||Luiz Fernando Ranghetti||
||Karl Ove Hufthammer||
||Albert Astals Cid||
||Aaron J. Seigo||
Internationalization (i18n) Status
There are 61 selections this week
Couple of fixes:
* Close empty extender when clicking in it.
* Saner size policy for Plasma::Meter. (the height should never be able to be smaller then svg's height because labels could otherwise overlap, actually, the whole current approach doesn't consider font size, I think we should think of a better approach for 4.3)
* Use the same spacer as used in the panel for extenders. I think we would want some svg themable shared spacer widget for 4.3, but at least this adds some consistency.
Manager can not be a singleton since some of the protocols may (and do) use bits of libplasma that offer singleton accessors, resulting in random deletion and re-creation-on-exit orderings which causes crashes.
this also prompted:
* limiting severely the usage of namespaces as they were used in a way that made reading the code dreadfully confusing in places (hint: humans are not compilers)
* collapsing the three protocol classes in core into one SystemTray::Protocol class
* genericizing the management of protocols in Manager
* altering the config options controlling which protocols to show from negatives (resulting in !readEntry's in the code) to positives defaulting to true
Don't go crazy if told to reconfigure.
This fixes some very old bugs
Save the upper/lower limit values as doubles, not longs.
Sorry for taking so long to fix this bug - I'm ashamed.
Fixed skype contact actions (call, authorize) in all online status
Fixed geting info abou contact
Fixed destructor for all widgets
Fixed add/edit account widget
Ported skypedetails widget to KDE4
Ported skypecalldialog widget to KDE4
Ported starting scripts befour/after call to KDE4
* Add support for compiling sources at runtime.
Supported languages for now: C#, Boo, Nemerle, VB.NET (this one doesn't quite work yet - the compiler seems to be buggy).
The first non-blank line in the mainscript file needs to be a comment in the form
// language:csharp references:System.Xml sources:otherfile.cs
'meta-comments' for different languages look equivalent.
By default references for qt-dotnet, kde-dotnet and plasma-dll are added.
The default language is C#, the default source file is only 'main'.
So if a script is written in C#, only needs qt, kde and plasma bindings and all the code is contained in 'main', this comment can be omitted.
For each source file a md5 hash is created so it's only recompiled if the source was modified.
Start implementing the reverse of parsing - converting an AST back to c++ code.
The whole AST is covered already, but it's hardly tested yet - for unit tests I need to write code that will check (ast -> code -> ast2) ast == ast2 ... ie tree comparison logic
Learn from Okular:
- make X11 window settings less intrusive (there were problems with notification popups like kopete, Alt+Tab was impossible, crashing in slideshow under GDB behaved almost malignant, portability of Qt::X11 flags unknown)
Use the same flags as Okular
- disable the screensaver using DBus
show a "Copy Here" and "Link Here" action, when a file is dragged within the same folder
Create backups of KJots books on close (max 10 backup versions).
Remove backups of deleted books 7 days later.
That gives you a week to realise you shouldn't have deleted the book and recover it.
don't load the kcm if we just ask for the minimumsize cause it renders the whole logic of the proxy invalid.
this removes e.g. the ~10 seconds delay on calling 'Configure Konqueror'
see also <a href="http://lists.kde.org/?l=kde-core-devel&m=122835969412478&w=2">http://lists.kde.org/?l=kde-core-devel&m=122835969412478&w=2</a>
Alt+O is a KAction for "focus location bar" in konqueror, so it was giving a conflict-dialog when the findbar was shown. Hard to find a free letter in "Options", let's use the 'i' even though it's taken by F&ind already (Alt+i just cycles between both widgets).
All this is because apparently kacceleratormanager doesn't run on the findbar, but if it did, it wouldn't see the Alt+O KAction anyway.
Removed the hintPicture from FilenameLayoutDialog.
I think most of us agree that the hint picture was too empty and to thick. However I feel that by removing it the dialog's workflow structure suffers greatly.
At the moment I see 4 alternatives:
1) Readd the hintPicture and superimpose some hint text.
2) Readd the hintPicture and only center the icon, maybe make it thinner.
3) Leave it as it is, without any hint picture, I'm against this.
4) Add the icon from hintPicture to the FilenameLayoutWidget itself, right before the "Drop tokens here..." text.
Please review this and let me know what you think about it.
Revert "Update containments geometries before changing to zoom out level."
Revert "Remove old hack not needed anymore."
Revert "Fix nasty bug in CV that showed misplaced containments after resizing and changing containments."
This last-minute bugfix introduced regressions.
As it is a large code drop literally HOURS before tagging 2.0 final, this is absolutely the wrong time to do this and so must be reverted.
It randomly breaks the size of the containments so that applets are not always added in the right place, or restored correctly on startup, or other weirdness.
This needs more testing before committing.
Liw-, i understand that this fixes the browser-collapse bug, but we can't deal with a mess of bugs for 2.0 final.
Post-2.0, please recommit and we will work out the weirdness, but for now, this has gotta go.
* Much better, consistent error msgs when mixing different scalar types:
- in matrix-matrix product, static assert on the two scalar types to be the same.
- Similarly in CwiseBinaryOp. POTENTIALLY CONTROVERSIAL: we don't allow anymore binary ops to take two different scalar types. The functors that we defined take two args of the same type anyway; also we still allow the return type to be different.
Again the reason is that different scalar types are incompatible with vectorization.
Better have the user realize explicitly what mixing different numeric types costs him in terms of performance.
See comment in CwiseBinaryOp constructor.
- This allowed to fix a little mistake in test/regression.cpp, mixing float and double
- Remove redundant semicolon (;) after static asserts