| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Actually neither glFlush nor glFinish need to be called manually [1].
However, the rendering pipeline is then free to queue up future frames
which can cause a lag which is especially noticable with mouse movement.
To avoid this lag, we call glFinish after SDL_GL_SwapBuffers. This makes
sure processing of the next frame does not start before the current
frame is being displayed.
[1] http://www.opengl.org/wiki/Common_Mistakes#glFinish_and_glFlush
Reviewed-by: Yohann Ferreira
Sounded-fine-to: Erik Schilling
|
|
Reviewed-by: Erik Schilling
|
|
It was also setting the screenshot directory.
Reviewed-by: Erik Schilling
|
|
Reviewed-by: Erik Schilling.
|
|
Reviewed-by: Erik Schilling.
|
|
An optimization in the tile layer rendering code meant for drawing
repeated tiles faster was not taking into account the case where the
tile width does not match the width of the tile grid.
Reviewed-by: Stefan Beller
|
|
Before it was only enabled by default for Mac.
Reviewed-by: Erik Schilling
|
|
Maybe it used to be necessary for these methods to be virtual, but this
is no longer the case. Hence we can avoid wasting CPU ticks searching
through virtual function tables, especially for frequently used methods
like getWidth, getHeight and setAlpha.
MapLayer::getTile was inlined.
Reviewed-by: Erik Schilling
|
|
Instead, support for dynamically adjusting layout was added to the
Container class.
Various other places were also adapted to use the new layout support in
Container.
Reviewed-by: Erik Schilling
|
|
When uploading an SDL surface to an OpenGL texture, it was always making
a copy that had the desired size and pixel format. Now this copy is no
longer being made when the existing surface already has the target size
and pixel format.
In fact most images are already in 32-bit RGBA format after loading them.
Reviewed-by: Erik Schilling
|
|
While simply specifying "4 components" works, it is preferred to specify
the actual image format like GL_RGBA8 in our case.
http://www.opengl.org/wiki/Common_Mistakes#Image_precision
Reviewed-by: Erik Schilling
|
|
If the graphics driver supports this, there is no need to create
textures with power-of-two dimensions. It is then also preferred to use
regular textures than relying on the older GL_ARB_texture_rectangle
extension.
Reviewed-by: Erik Schilling
|
|
* Removed empty and unused function `setTargetPlane`
* Removed duplicated enabling of `GL_VERTEX_ARRAY`
* Removed strange smoothing code from `drawRescaledImage`
Reviewed-by: Erik Schilling
|
|
Textures were not rendering correctly on cards that did not support the
GL_ARB_texture_rectangle extension, since the alternative code path was
passing ints as floats and floats as ints to OpenGL.
Reviewed-by: Erik Schilling
|
|
Reviewed-by: Stefan Beller
|
|
This reverts commit 4eea727b7649726670d8963d11ab4fd429624b3e.
It broke the overrides of the virtual functions and leaves the Graphics
API in inconsistent state with some Image* being const and others not.
|
|
Acked-by: Erik Schilling
|
|
The class DebugSwitches is a derived class from Container, 'which
automatically deletes its child widgets, so there is no need for
such fragile cleanup code.'
acked-by: bjorn
|
|
Reviewed-by: Ablu
|
|
This reverts commit 137ade7226af37d073a5755b90181275664dd65c.
Turns out that when leaving the default values, the textures won't
actually work since they are incomplete unless the mipmaps are generated
(since GL_TEXTURE_MIN_FILTER is GL_NEAREST_MIPMAP_LINEAR by default).
For now we don't want to generate mipmaps anyway since we're going for a
pixely look, so revert the filters back to GL_NEAREST.
The reason this had worked for me is because on my system the texture
type GL_TEXTURE_RECTANGLE_ARB is used instead of GL_TEXTURE_2D and this
texture type does not support mipmaps.
|
|
|
|
Change is tested.
Reviewed-by: bjorn.
|
|
Reviewed-by: Ablu.
|
|
Reviewed-by: bjorn.
|
|
- shows manaserv skills at all
- shows the progress bar for manaserv skills again
- shows the level for this skills (server side patch needed)
TODO: URGENT: Fix attributes and skills getting stored in same list.
Reviewed-by: Bertram.
|
|
|
|
Reviewed-by: bjorn.
|
|
The window menu (the row of buttons at the top right in game) was
displayed 3 pixels left of where it should be, which makes the setup
button jump 3 pixels left when logging in.
Reviewed-by: Ablu
|
|
|
|
Mantis-Issue: 423
Reviewed-by: Thorbjørn Lindeijer
|
|
They are no longer sent as fixed-length string since this doesn't work
when the new protocol debugging mode is used.
Reviewed-by: Ben Longbons
|
|
|
|
This makes the client able to send and receive messages sent in debugging
mode, where the contents of the messages are annotated.
For messages sent from the client the debugging mode is currently always
enabled. Later on this could be an internal option or controlled from the
server side.
Reviewed-by: Erik Schilling
|
|
There wasn't a whole lot gained by sharing a common base class, and it makes
extending the manaserv Message{In,Out} classes with a debugging mode
unnecessarily complicated.
Reviewed-by: Yohann Ferreira
|
|
Also, since GCC 4.7 there is a binary compatibility issue when linking with a
Guichan that was not compiled in C++11 mode. This commit also allows compiling
with GCC 4.7 with C++11 mode turned off.
Reviewed-by: Erik Schilling
|
|
- kick code untested because no kick button exists yet
Reviewed-by: bjorn.
|
|
|
|
Reviewed-by: bjorn.
|
|
Reviewed-by: bjorn.
|
|
|
|
The tmwa/BeingHandler was adjusting the effect id rather than the emote
id, causing it to not find the last emote (and the rest of the emotes
only worked correctly because the effect ids were consecutive in the same
order as the emote ids).
Furthermore, the EmoteShortcutContainer refused to draw the icon for the
last emote due to an off-by-one error in dealing with the 1-adjusted
emote ids used by the EmoteShortcut class.
Also cleaned up some old remains of a player ignore strategy that used
to use the two balloon emotes (this had been their original purpose).
Reviewed-by: Erik Schilling
|
|
The Windows releases have not been properly identifying themselves, and
it seems to be due to a compiler issue, as discovered by Stefan Dombrowski
(@cody). He did a similar change but didn't push it anywhere so I've redone
it.
The line that is now split up was very preprocessor-heavy, which may have
been part of the problem, but we'll probably never know what it was
exactly.
|
|
|
|
Reviewed-by: Thorbjørn Lindeijer
|
|
Also added Xcode project for others to compile for Mac OSX.
Reviewed-by: Thorbjørn Lindeijer
|
|
|
|
This is in order to still support older compilers, in particular GCC 4.2.1,
so that Mana may be compiled for Maemo 5.
Reviewed-by: Yohann Ferreira
|
|
When creating a custom server entry, the default port is used if no
other port is specified.
Reviewed-by: Thorbjørn Lindeijer
Mantis-issue: 464
|
|
Added a black background for maps that are smaller than the screen resolution
and centered them on screen.
Reviewed-by: Thorbjørn Lindeijer
Mantis-issue: 193
|
|
|
