* Re-enable full-width characters detection and display;
* Fix cursor drawing on full-width characters;
* Fix debug dump for multi-byte characters;
* Fix file permissions for debug capture log.
Fixes#6717. Also may improve behaviour related to #6227.
Improve the unicode character processing and classifying routines by
wrapping up the UChar32 procedures from ICU. That fixes functional
regression introduced in hrev38017 and allows to fix East Asian Width
problems int the Temrinal.
Loop backwards if possible, if not, set a variable and use that instead.
There were a couple of instances where the loop style got changed from
for (int32 i = CountItems(); --i >= 0;)
to
for (int32 i = CountItems() - 1; i >= 0; i--) {
but should be functionally equivalent.
On MouseDown draw a diagonal arrow, on MouseUp complete the animation and
expand. If you hold down the button it will stay diagonal until you MouseUp
and either return to normal or animate and expand if over the arrow.
Reformatted ExpandoMenuBar.h and TeamMenuItem.h
Renamed fLastClickItem to fLastClickedItem
Added a DrawExpanderArrow() method
Renamed private InitData() method to _InitData() and moved it to the bottom
* Add ability for fSettings to pass on network
stats
* Show KB Sent / Received for interface
* Drop the wireless / wired tab name.
(we are going to need another tab for wifi)
* Add wifi network name to connection field
if interface is wifi.
... like BeOS R5. I looked in the commmit logs for this one and there wasn't really any
explination for why this got changed, so, I'm changing it back to the way it was in R5 which
is right arrow for unexpanded, down arrow for expanded. Please yell at me if this change
was intentional.
If you have expander turned on with expanded apps and you quickly remove teams with the VDG
you can remove a team not under your mouse pointer, instead you remote the team above. This
is because the window watcher thread hasn't updated yet so the TeamItemAtPoint() method
reads a window menu item instead of the team item. The solution is to lock the window
watcher thread and explicitly remove the window menu items in RemoveTeam().
This bug can be really bad if you accidentially VDG Tracker as your system gets hosed until
you restart Tracker or reboot.
This means the B_COLOR_WHICH_COUNT goes from being a public constant to a
private one. It sill looks like a public constant starting with a B_ though.
I hope that's not a big deal. Too bad we can't get the count of an enum.
This fixes a maintainance problem where you have to update this otherwise
unrelated file to keep it in sync whenever you add a color constant.
I've added a B_COLOR_WHICH_COUNT constant to the color_which enum which should
be updated to point to the newest color constants as new ones are added. I
reworked ServerReadOnlyMemory to use this constant instead of using to the
current largest color constant directly. If you use B_COLOR_WHICH_COUNT to
refer to a color in your code expect to get unpredictable and nonsensical
results. Most likely you'll get an undefined result which will return black
but don't depend on it.
The net effect of this is that ServerReadOnlyMemory doesn't need to be updated
anymore when new color constants are introduced but will continue to produce
correct results.
Eliminate kNumColors constant, replace it with B_COLOR_WHICH_COUNT
This allows you to change the scrollbar thumb color in Appearance preferences.
The default color is 216, 216, 216 so the scroll bar thumb looks the same by
default. Perhaps someday this can be updated to something a bit more colorful.
On some 64 bit architectures program and library images have to be mapped in
the lower 2 GB of the address space (due to instruction pointer relative
addressing). Address specification B_RANDOMIZED_IMAGE_ADDRESS ensures that
created area satisfies that requirement.
Placing commpage and team user data somewhere at the top of the user accessible
virtual address space prevents these areas from conflicting with elf images
that require to be mapped at exact address (in most cases: runtime_loader).
