* added another optimized bitmap drawing routine
for B_OP_ALPHA and BGR[A]32 bitmaps. When actual
blending takes place, it is more than 1.7 times
faster than R5, and about the same speed when
the bitmap is fully opaque.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15761 a95241bf-73f2-0310-859d-f6bbb57e9c96
with quite good results I might add. Drawing B_RGB32
bitmaps is more than 1.2 times faster than on R5, while
B_CMAP8 bitmaps are slightly slower. The optimization
is only for B_OP_COPY and unscaled bitmaps
(B_RGB32 and B_CMAP8). Drawing only parts of the bitmap
is supported. Adding optimization for scaled bitmaps
should be beneficial, since the generic version is 2 two
4 times slower. I think it gets even worse for partial
bitmaps.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15758 a95241bf-73f2-0310-859d-f6bbb57e9c96
with the "EventMask" test app the last time:
* The focus view didn't get any mouse messages anymore if there was a permanent
B_POINTER_EVENTS view.
* B_KEYBOARD_EVENTS now works again.
* B_MOUSE_WHEEL_CHANGED messages no do arrive their targets without prior
keyboard input.
* Added TODO item that other focus messages don't go through to the app without
keyboard input (fix only works for B_MOUSE_WHEEL_CHANGED).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15754 a95241bf-73f2-0310-859d-f6bbb57e9c96
it could crashed the server.
* ViewLayer now deletes the view bitmap on destruction, if any.
* BitmapManager::Delete() now also accepts NULL bitmaps.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15723 a95241bf-73f2-0310-859d-f6bbb57e9c96
or even the resizing mode isn't done yet, though. See TODOs.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15720 a95241bf-73f2-0310-859d-f6bbb57e9c96
as only the BitmapManager class is allowed to call them.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15718 a95241bf-73f2-0310-859d-f6bbb57e9c96
and ResizeBy()
* WindowLayer::SetSizeLimits() needs to be called with the
AllWindows lock held
* I was observing weird behaviour with "unclickable" windows
that I might have fixed by explicitly excluding invisible
windows from Desktop::WindowAt(), there might be something
wrong with the "current" window list though, Axel would know
* finally found the problem with "delayed background clearing"
* enabled delayed background clearing and removed unnecessary
code. It should be more efficient, since it clears larger
areas at once, and it solves the problem of views unable to
draw into regions that are pending for another update - among
other things, updates in resizing windows are more fluent,
especially for B_FULL_UPDATE_ON_RESIZE views. "Cut off" scroll
bars should no longer appear when the view being scrolled takes
too long to redraw.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15714 a95241bf-73f2-0310-859d-f6bbb57e9c96
into a new frame_buffer_support.h
* added blend32 routine for blending a certain color with
a scanline in the frame buffer
* added "solid" versions of B_OP_ALPHA drawing with
B_ALPHA_OVERLAY alha function (blending on top of
a non-transparent background such as the frame buffer)
* implemented an optimized shortcut for alpha blended
FillRect() in Painter
* used the "packed" version of scanlines for shapes with
an outline thicker than 4 pixels (and filled shapes anyways),
this improves drawing speed when there are a few anti-aliased
pixels at the beginning of a scanline, then a solid fill and
some anti-aliased pixels at the end of the scanline. Such as
large letters.
To summarize: The alpha blending in Painter seems to be about
1.45 times faster than on BeOS R5 which benefits drawing large
shapes. For example, drawing a large alpha blended rounded rect
is 1.28 times faster on the Haiku app_server. On the other hand,
B_OP_COPY is quite tough to beat. It is currently 10 times faster
on R5. But a great deal seems to be caused by the Painter
rasterization algorithm itself, since commenting out the actual
drawing doesn't gain any speed.
The other useful experience I collected was that reading and
writing and over the PCI bus in the same loop really hurts
performance. It is actually faster (like 1.8 times!!) to allocate
a second buffer, read from frame buffer into that, doing the
blending at the same time, then writing the buffer back to the
screen.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15698 a95241bf-73f2-0310-859d-f6bbb57e9c96
mappings yet, that's a good way to test this functionality.
* Turns out the directory of the mappings must be known already - they
should be added automatically, but I've added them manually for now
(which is okay for the default system directory).
* Having more than one style with the same family in the mappings didn't
work as planned - it now does.
* On my current system, time spend in the font manager constructor went
down from 1.5 secs to around 12000 usecs (and I have only a moderate
number of fonts installed, or so I thought - looks like the mappings
were a good idea :-)) - and that directly cuts down the boot time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15580 a95241bf-73f2-0310-859d-f6bbb57e9c96
