Generated by StatCvs v0.2-dev
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Summary Period: 2001-02-01 to 2001-01-26 (Commits 3995-4044 of 12744)
Added new avi plugin to autogeneration process.
This will generate both the VC6 project files for the plugin
and the test app (Eric made the app project file autogenerate
if you generate the avi plugin).
This only affects those folks who generate VC6 project files locally...
1 lines of code changed in:
fixed a compile-time error in CSWS
2 lines of code changed in:
- Added GetPrivateObject() to iMaterialWrapper.
- Added QUERY_INTERFACE_SAFE() which will first test if the given
object is NULL. In that case it will return NULL. Otherwise
equivalent to QUERY_INTERFACE().
- Let csSpriteTemplate implement iSprite3DFactoryState. This is in
preparation of the csSprite3D move to a plugin.
192 lines of code changed in:
More attractive window configuration.
5 lines of code changed in:
Resolved several const vs. non-const issues.
15 lines of code changed in:
Fixed a slew of compilation warnings in cswstest.cpp.
40 lines of code changed in:
Fixed several compilation problems in MazeD and CSWS. Also resolved some
const vs. non-const problems in these modules.
6 lines of code changed in:
Update.
121 lines of code changed in:
Fixed video.mak so that the `vid' target only appears in the application
list if the platform is also able to build video-related plugins. Also
fixed formatting of help message.
18 lines of code changed in:
added output of what is currently done when trying to load the video
15 lines of code changed in:
Once again supports conditional setting of OBJ.CSSYS. (This feature was
disabled in the most recent revision.)
2 lines of code changed in:
My last commit.
-*- Now the toolbox toolbar tracks the active editor window (model or world)
and changes accordingly. I have drawn a few more icons for MazeD in GIMP,
and switched the toolbar texture format to truecolor PNG with alpha
channel. Now toolbar icons can have (and some already have)
semi-transparent parts :-)
-*- CSWS buttons with pictures are grayed (pictures are displayed
semi-transparent) when the buttons are disabled. Also they do not emmit
the cscmdRightButtonClick message when they are disabled (in MazeD this
was manifesting as second-level toolbars popping up even from disabled
buttons).
-*- Floating hints will disappear properly even if some window has captured
the focus (such as a dialog window). To do so they grab the mouse.
-*- Rewrote much of the csTreeCtrl (treeview) control. The basical interface
is very similar (except that I've renamed csTreeCtrl into csTreeBox);
most of the changes are related to the internal implementation.
Other features: a completely new look-and-feel; a extended, more flexible
and less cumbersome API (and a simpler implementation).
Also I have changed cswstest so that you can switch miscelaneous tree
control styles at runtime and see them in effect instantly (like for
notebook control).
-*- Made the csLayout class to set the CSS_TRANSPARENT flag (again?) since
otherwise the dialog background "under" the layout isn't drawn, and since
the layout itself does not paint anything, you get an hall-of-mirrors like
effect (when you move the window).
-*- Dialogs with no children will SuggestSize 8x8 rather than 2x2 like it
did before. This allows for better identification of windows with wrongly
set CSWS_TOOLBAR style bit set :-) (toolbar is a dialog too, so if you
wrongly set this bit when creating the window, you'll get just a couple
of thin black lines which are confusing, at least they were for me :-)
-*- Changed a little the behaviour of csComponent::NextChild() and PrevChild():
now it returns the next/previous *visible* and selectable child rather
than just selectable. This has the effect of not showing anymore the hidden
windows when closing a visible window (per Jorrit's request). However, if
there is nothing left than hidden window, a hidden window will be focused
anyway (but not shown).
-*- Changed a little listbox behaviour and look. No lightblue frame on the
focused item anymore; when listbox is active, the focused item is denoted
by a solid background, otherwise it is denoted with a thin frame of the
same color.
-*- Added a method csComponent::OtherToThis (csComponent, x&, y&) which will
convert coordinates from coordinate system of other component to this
component's coordinate system. This is faster than the using the
LocalToGlobal -> GlobalToLocal pair if the `from' component is a
child (or grandchild, or grandgrandchild etc) of `this' component.
-*- Made the toolbars resizeable in MazeD. By default toolbars are now four
buttons wide (rather than two like it was before), to reserve a little
space for coming "world tree" and "material selector" toolbars.
-*- Added a new kind of events: csevMouseClick. This is generated by mouse
event manager (AKA mouse driver) when it detects that the mouse button
has been pressed and then released within "DoubleClickTime" and the
mouse haven't been moved further than "DoubleClickDist". That is, when
you quickly press and release the mouse button twice in sequence, you
will generate the next events:
csevMouseDown
csevMouseUp
csevMouseClick (at the same time with MouseUp)
* This event is emmited only if the mouse haven't moved too far
* between MouseDown and MouseUp and the time between them haven't
* exceeded some threshold value.
csevMouseDown
csevMouseDoubleClick (at the same with MouseDown)
* This event is also generated if the same conditions are true.
csevMouseUp
csevMouseClick (at the same time with MouseUp)
* Yet another mouse click event, if the conditions are met.
You can notice a slight conceptual difference in Click and DoubleClick
because Click is generated after mouse is released, and DoubleClick is
generated when mouse is pressed. Well, this is so for historical reasons -
most GUIs uses mouse this way.
Also note that the behaviour of double clicks event has slightly changed:
now DoubleClick is emmited *in addition* to MouseDown rather than *instead*
like it was before. That is, you will *always* get all the MouseDown and
MouseUp events disregarding whenever click and doubleclick events have
occured; these events are additional and are introduced just to help the
programmer to track these conditions; in general he could do it himself
by using timer, remembering position etc.
This has a slight impact on some seldom-encountered code sequences: if
before to catch absolutely all mousedown events you had to catch both
MouseDown and MouseDoubleClick events, now you have to catch just
MouseDown and don't bother about MouseClick and MouseDoubleClick -
the event you are looking for will be always emmited. In fact, if you
will leave the code as before, you will intercept now two events instead
of one; while in most cases it doesn't matter, in some cases it can.
-*- Added a small context menu to every toolbar in MazeD.
-*- Started implementing world tree toolbar. Now it is one of the standard
toolbars. It displays in a tree fashion the structure of the 3D world.
At the "top" level there can be only sectors (AKA rooms), inside them
you can nest things and sprites (but not rooms).
2023 lines of code changed in:
Fixed after somebody broke it:
- remove ?= from assignment to CSSYS.OBJ since GNU Make 3.77 doesn't expand
the text after such an assignment. That is, "CSSYS.OBJ ?= $(...)" is treated
literally, e.g. the text inside the brackets is not expanded. If someone
really needs ?=, use ifndef instead -- although I don't see any assignments
to CSSYS.OBJ, so I can't understand why somebody have put ?= there.
- Removed "dep: ..." after last line of "clean:" target - it seems like traces
of a CVS conflict introduced by someone in CVS.
2 lines of code changed in:
Altered path to file "video.avi" to the "data"
directory as it was looking for a "this" directory.
1 lines of code changed in:
Update.
16 lines of code changed in:
now includes the classnames for avi and opendivx codec
6 lines of code changed in:
extended the PLUGIN var to build AVI and optionally opendivx
23 lines of code changed in:
removed some debug code and a stupid comment
2 lines of code changed in:
this is a hacked "simple" that features video playing. note that it only plays divx encoded avi's (because this is the only codec thats around right now). get both, the divx library and an example avi, from www.projectmayo.com.
rename the avi to video.avi and place it in CS/data.
462 lines of code changed in:
works for the divx codec
for opengl the pic is flipped vertically , eek ( this is because i abused the csImageArea and filled it myself while it was actually meant to be filled by the renderer. Jorrit talked about introducing a FillPixelArray, which should solve this problem - or the csImageArea can be extended to do that )
306 lines of code changed in:
works for divx now (well, this is the only codec right now)
0 lines of code changed in:
Eric Sunshine made the following changes to jobber.pl:
-*- Now allows user-specification of the temporary working directory
(default is "/tmp").
-*- Added a --version option.
23 lines of code changed in:
eliminated a few cases where options have different names in the walktest
sources and its config file
17 lines of code changed in:
Fixed CS/include/cssys/next/csosdefs.h so that it respects SYSDEF_ACCESS.
2 lines of code changed in:
Fixed a potential memory leak which was introduced in the most recent
revision of glcommon2d.cpp.
7 lines of code changed in:
fixed a compiler warning in the OpenGL canvas driver
6 lines of code changed in:
Removed a duplicate entry.
0 lines of code changed in:
Update.
7 lines of code changed in:
Fixed SED_DEPEND in cs.mak so that it works more portably across different
versions of GNU make. In particular, the old definition failed to work
correctly on MacOS/X Server 1.0 (Rhapsody).
5 lines of code changed in:
Changed the behavior of the -simdepth option in the AppKit-based Apple/NeXT
2D driver. If the user specifies an illegal depth, then a suitable warning
message is issued and the request is ignored. Previously, it would return
a failure status from iGraphics2D::Initialize() in this circumstance.
6 lines of code changed in:
Reverted Philip's change to dnp_tab.cpp and solved the missing malloc.h
problem in a different fashion. The problem with Philip's change was that
his #if test was not well formed, and he also forgot to apply the change to
CS/support/gnu/bison.sim, from which dnp_tab.cpp is built.
7 lines of code changed in:
modified the sound drivers so that they are now opened, closed and updated
via system events
53 lines of code changed in:
Fixed the following problems in csConfigFile which were introduced in the
most recent revision:
-*- Fixed a dangling-pointer bug in csConfigFile::Clear().
-*- Fixed several problem spots in csConfigFile::LoadFromBuffer() where it
could assign the wrong comment as the "end-of-file" comment.
22 lines of code changed in:
Fixed a overfull hbox TeX warning in meshobj/using.txi.
4 lines of code changed in:
update to the config system: comments at the end of a config file are now
preserved and can be modified by the application
51 lines of code changed in:
Added documentation for the Skeletal Animation format, and the INCLUDESPRITE macro
33 lines of code changed in:
Eric Sunshine made the following changes to the project:
-*- Added a makefile (macosx.mak) for MacOS/X. This makefile is for
Apple's new MacOS/X platform using the Cocoa-based Apple/NeXT port.
Since I do not have access to either MacOS/X or recent Macintosh
hardware, this makefile is currently untested, but I believe that the
project should build cleanly.
-*- Added new menu definition to next.cfg for MacOS/X which respects the
new Aqua-style interface. This differs from the existing menu
definition for MacOS/X Server 1.0 (Rhapsody) which is specific to the
Platinum-style interface.
-*- Eliminated reliance upon the Objective-C++ compiler for the Apple/NeXT
port by reorganizing the code and writing plain-C stubs which
interface between the C++ world of Crystal Space and the Objective-C
world of the AppKit frameworks of Cocoa, OpenStep, and NextStep. By
eliminating the need for an Objective-C++ compiler, the project can
now be built on MacOS/X (which does not ship with an Objective-C++
compiler). The port-specific .cpp files now contain only pure C++
code. All of the Objective-C code has been moved to .m files. (The
standard file extension for Objective-C source files is .m.)
-*- Added icursor.h. This is a pure-C header which contains the
csMouseCursorID constants which used to reside in igraph2d.h. Also
added ievdefs.h. This is a pure-C header which contains all of the
event-related constants which used to reside in ievent.h. These files
are pure-C, rather than C++, because they must be included by the
Objective-C source files of the Apple/NeXT port.
-*- Added NeXTDynamicLibrary.cpp, which is the C++ portion of the plugin
loading facility for the Apple/NeXT ports. The existing
NeXTLoadLibrary.m is now pure Objective-C and publishes a plain-C
interface for use by NeXTDynamicLibrary.
-*- Added NeXTConfigFile which is a plain-C wrapper over the C++ Crystal
Space configuration facility.
-*- Added rules and support facilities to the Apple/NeXT makefile
(next.mak) for recognition and compilation of Objective-C (.m) files.
Also added Objective-C support to next2d.mak.
-*- cssys.mak now sets OBJ.CSSYS conditionally; only if it has not already
been set. This allows the platform-specific makefile to override the
list of object files for the system driver. This facility is used by
the Apple/NeXT port to mention Objective-C (.m) object files in
OBJ.CSSYS as well as C and C++ files.
-*- cssys.mak now automatically invokes `vpath' for all C files in the
system driver source directories (in addition to the `vpath' for C++
files which it has always invoked).
-*- NeXTSystemDriver::SystemExtension() understands the following new
messages: "timerfired", "continuelooping", "continuerunning",
"dispatchevent", "hidemouse", "showmouse", and "requestshutdown".
-*- De-coupled the tight binding which existed between the AppKit-based
NeXT 2D canvas and the NeXT system driver's Objective-C delegate.
Some of the functionality in NeXTDelegate was actually specific to the
AppKit-based canvas and may not have been useful or correct for other
drivers, such as the planned OpenGL and CoreGraphics drivers. The
following functionality has been moved from NeXTDelegate into the
canvas plugin:
o Setting of "Paused" status message on window when application is
suspended. The canvas now listens for cscmdFocusChanged events
to implement this behavior.
o Creation and management of the window's tracking-rectangle as
well as the associated behavior.
o Window event mask management.
o Window position adjustment for video optimization.
o Response to user initiated window-close events.
Other functionality which may be useful to any canvas implementation
for Apple/NeXT is published by the system driver via
iSystem::SystemExtension(). The messages which may be useful to 2D
drivers include "dispatchevent", "keydown", "keyup", "mousedown",
"mouseup", "mousemoved", "hidemouse", "showmouse", and
"requestshutdown". Furthermore, the canvas no longer communicates
directly with the system driver's Objective-C proxy object; nor
vice-versa. All platform-specific communication with the system
driver is now accomplished via iSystem::SystemExtension().
-*- The default NeXT 2D canvas plugin no longer includes NeXTDelegate.h
from the CS/libs/cssys/next directory. 2D drivers now communicate
with the platform-specific library strictly via
iSystemDriver::SystemExtension().
-*- The Objective-C application delegate in the Apple/NeXT port now
listens for application-activation and application-deactivation
notifications in order to suspend and resume the engine rather than
relying on window-activation and window-deactivation messages. Not
only is this more sensible, but it also helped to de-couple the
AppKit-based 2D canvas from the system driver.
-*- Fixed a simple logic problem in the AppKit-based Apple/NeXT 2D
graphics driver where it failed to emit a warning message when the
user selected an invalid display depth via the -simdepth option.
-*- The AppKit-based 2D canvas for the Apple/NeXT port no longer relies
upon the WindowServer to report its own preferred bit-depth. It turns
out (so I discovered) that this is an unreliable operation on OpenStep
since the WindowServer will actually report a different value the very
first time the program is run versus subsequent runs. Instead, the
driver now relies upon the preferred depth as reported by the NSWindow
or Window class.
-*- Mouse tracking in the canvas' window is now more accurate when user
switches away from the window or application and then returns. In
particular, the driver now pays attention to the mouse's location at
the time that it once again begins tracking the mouse.
-*- Fixed a problem with the AppKit-based Apple/NeXT 2D driver. If the
mouse was hidden prior to the application being paused, it would not
be re-hidden when the application resumed.
-*- The special CFLAGS needed to build the AppKit-based 2D driver
(next2d.mak) are now only applied when explicitly building that driver
rather than being applied to all modules.
-*- Added a new `inherit' keyword to the list of keywords understood by
NeXTMenu when processing a menu item entry. This keyword allows a
menu item to inherit attributes from some other menu item. These
attributes can then be overridden locally. This facility is useful,
for example, when one menu item definition differs from another only
by its title.
-*- Added a copious amount of documentation to next.cfg describing all of
the supported configuration options.
-*- Fixed a bug in NeXTMenu on NextStep where it would fail to properly
build submenus on account of a mis-constructed submenu key prefix.
-*- Converted NeXTMenu and NeXTConfigFile to work with new csConfigFile
and related SCF interfaces rather than the deprecated csIniFile.
-*- Fixed vfs.cfg so that it works with MacOS/X Server again following
Andrew's change to the meaning of CS_PLATFORM_NAME.
-*- Fixed the Apple/NeXT system driver and next.cfg so that they once
again work following andrew's change to the meaning of
CS_PLATFORM_NAME. More importantly, re-added the OS_NEXT_DESCRIPTION
macro since it is needed to properly distinguish between platform
options in next.cfg. This is necessary since the menu layout differs
between MacOS/X and MacOS/X Server 1.0 (Rhapsody). However, Andrew's
change made it impossible to make this distinction.
-*- Fixed the csMaterial(csTextureWrapper*) constructor so that it works
correctly for the Apple/NeXT platforms when the engine is used as a
plugin module. The problem was that this constructor was invoking
another constructor in the same class. The result of invoking a
constructor in this situation is undefined at best since the C++
specification leaves the behavior up to the compiler implementation.
In this case, the result was that none of the instance variables,
except `texture', were being initialized. In particular, the lack of
initialization of the SCF goop resulted in IncRef() being called
infinitely. Having fixed this bug, Simplep finally works on
Apple/NeXT.
-*- Fixed a problem where many `friend' functions of ddgMatrix4 and
ddgVector2 were not declared as `inline' in the respective headers.
This resulted in implementations of these functions appearing in every
file which included those headers, which ultimately caused linkage
errors.
-*- Eliminated all remaining imports of the deprecated <malloc.h> from
platform-independent project sources files. (Some platform-specific
files may still include malloc.h, though.)
-*- Eliminated several compilation warnings reported by the OpenStep
compiler throughout the project regarding implicit conversion of void*
to some other type.
-*- Eliminated a compilation warning reported by CodeWarrior in
csEventQueue::Lock(). The compiler did not like the empty statement
following the while() loop.
433 lines of code changed in:
Fixed a bug in csConfigFile where the SetInt(), SetFloat(), and SetBool()
methods were applying `delete[]' to the Data value twice.
3 lines of code changed in:
Fixed a bug in csConfigFile where the SetInt(), SetFloat(), and SetBool()
methods would not actually set the key of a newly created node if the value
being set was 0, 0.0, or false, respectively.
15 lines of code changed in:
Fixed the fix from Philip Wyett to make it slightly more
robust.
7 lines of code changed in:
Update.
31 lines of code changed in:
Minor:
A precision addition to previous fix
and code clean for readability.
1 lines of code changed in:
Added a divide by zero check.
Was causing an runtime error in
"appptlab" under Windows. "appptlab"
now back up and running.
8 lines of code changed in:
- texture size on caching is now only calc'ed when needed
- keycolor setting is now remembered so that its not processed again when texturehandle is created again
24 lines of code changed in:
Fixed a crasher in PTLab where it was trying to access a texture
(cslogo.gif) which it had never loaded.
3 lines of code changed in:
Save & RestoreArea now tries to be good with PixelBytes == 2
40 lines of code changed in:
Took out reference to non-existant texture archive
at http://server.empathy.com/crystal/
0 lines of code changed in:
fixed a bug in the sound loader
1 lines of code changed in:
moved the main system/application config files to the new config system
377 lines of code changed in:
Made the following changes to inifile.cpp:
-*- Eliminated compilation warnings about implicit conversion of void* to
some other type.
-*- GetYesNo() now also recognizes "1" as true in addition to "yes",
"true", and "on".
6 lines of code changed in:
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Generated by StatCvs v0.2-dev