Makes it useable in applications and less confusing. Fixes#6468.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42521 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Add an option to draw the button directly, i.e. when they are clicked.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42493 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Each Window has a reference counted WindowStack class which can be shared between stacked Windows. To keep the Decorator separated from Window there is another tab list in the Decorator now. The index of the stacked Window in the window stack is the same as the index of the tab in the Decorator. Properties like title or window focus are managed on a per tab basis now. This mean when you set the title in the Decorator you also have to specify the tab id which is equal to the window position in the stack.
* When drawing the decorator its important that only the top window is doing the drawing. Also the top window drawing engine should be used.
Actually that is only a problem directly after a window is stacked and the other window has still a none empty dirty region. In this case we clear the dirty region of this window and stop the drawing (the top window will draw everything).
* Track if shifting of a tab is still ongoing, i.e. mouse still down.
* The key event filter called the DesktopListener without holding the window write lock. This probably caused #7801 and #7796.
* Commented out assert's in Window::SetScreen and Window::Screen. Add TODO because I'm not sure about the screen access.
This breaks all existing decorators again, sorry guys! Haven't looked into any other then the default decorator (and the SAT decorator). Will not fix the others in the near future so go for it! Since applications should be able to rely on S&T features the other decorator must be able to handle multiple tabs as well. A simple solution would be to draw all title bars in multiple rows. That probably looks quit poorly. Think the better solution would be to draw a tab interface in the title bar, e.g. like in KDE.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42478 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Remove a superfluously assert which sends the app server into the debugger. More fixes following.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42472 a95241bf-73f2-0310-859d-f6bbb57e9c96
- When using the graphical debugger by default, fall back
to setting up gdb handover if the GUI is unavailable.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42446 a95241bf-73f2-0310-859d-f6bbb57e9c96
teams to the native debugger instead. This assumes the latter is installed
in /boot/system/apps.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42445 a95241bf-73f2-0310-859d-f6bbb57e9c96
preflet. Replacing B_TRANSLATE with B_TRANSLATE_COMMENT to better discern "never" and
"Never".
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42426 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Simplify the algorithm a bit, no functional changes are intended.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42377 a95241bf-73f2-0310-859d-f6bbb57e9c96
causing the message to be sent with a timeout of "true" (getting converted to 1)
instead of the intended 0 meaning no/infinite timeout. This caused the message
sending to be aborted due to the timeout if it was blocking on a full port for
example. Since the return value is never checked noone noticed.
It's possible that this was the cause of some lost input messages (mouse,
keyboard) when the system was under heavy enough load for either the port heap
to be exhausted (unlikely) or the input_server <-> app_server port to run full
(quite possible).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42271 a95241bf-73f2-0310-859d-f6bbb57e9c96
Localize the mail server and preferences. Get rid of MDRLanguage.h.
The Japanese translation is attached to #7689. Please give a notice when it is in the translation system and the ticket can be closed.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42268 a95241bf-73f2-0310-859d-f6bbb57e9c96
* GCC 4.5 complains about switchs with cases not part of the window_look enum (see WindowPrivate.h).
* To keep them private, cast the enum to int before use in switchs.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42250 a95241bf-73f2-0310-859d-f6bbb57e9c96
This changes how Decorators are managed and applied. The app_server
no longer scans and maintains the available ones himself, but is
simply asked to load a Decorator add-on from a provided path.
The Decorator scanning is moved into DecorInfo and DecorInfoUtil,
private classes in the InterfaceKit. The bin command 'setdecor'
uses those.
I cleaned up all the coding style violations that I could find,
removed chunks of code which didn't make sense (if you never put
a NULL pointer into a list, you don't need to check for this and
so on) and also cleaned up other passages for improved clarity
and simplicity.
I also tested the functionality and it works fine. Would even be
Ok to include in Alpha 3, IMHO. Thanks for the patch!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41581 a95241bf-73f2-0310-859d-f6bbb57e9c96
- Adds IPv6 fragmentation support and some partial work to enable configuration of IPv6 in net_server. Not currently in the build.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@41291 a95241bf-73f2-0310-859d-f6bbb57e9c96
