(seems half of the references were one way, the other half were
the other way). Also made irix link to $(OCGDIR)/intern/*SoundSystem/*
instead of $(OCGDIR)/gameengine/*SoundSystem/*
I need to get openal working on my machine before I can test it so
if it doesn't work feel free to fix it. Hopefully this will be
the majority of the stuff though.
Kent
SoundSystem has been moved from source/gameengine to intern. This was needed
because functionality from SoundSystem was needed by
source/blender/src/editsound.c.
* Removed the option for the openal check in configure.ac. It's needed now by
SoundSystem
* Removed the functionality for checking if 'noaudio' was provied on the
commandline. Now audio is default on.
* this apple method to do it generates unwanted events, so it could not
be used for the menu mouse-move trick. this is #ifdeffed
* we should re-evaluate if this mousemove should remain in blender...
there are better ways (scrolling contents of pup)
* plus: removed warnings from drawview.c
the text below as comment:
adding a glFinish() here is to prevent Geforce in 'full scene antialias' mode
from antialising the Blender window. Officially a swapbuffers does a glFinish
itself, so this feels really like a hack... but it won't harm. (ton)
Removes ssl from blender.
makes the following directorys not needed anymore:
blender/intern/keymaker
blender/source/blender/src/pub
blender/source/blender/encrypt
blender/source/blender/decrypt
blender/source/blender/sign
blender/source/blender/verify
It works with The Nan Makefiles and autoconf, could whoever is working
on the other build systems update them so we can nuke those directorys?
They won't do anything but I figured I'd leave them in for a week or so
to make things easyer on people.
Kent
For some real idiot reason, osx only has the left-right arrow cursor builtin, and not an up-down arrowed one. The little hand cursor seems to be used in other apps, so let's stick with that!
-Ton-
Florian Eggenburger).
Full instructions are in doc/README.windows-gcc.
Main differences from Florian's patch:
- the 'lib' dir should now be the same level as the 'blender' dir (rather
than being a subdir of 'blender'). This is consistent with the other
platforms that bf-blender supports (tuhopuu will also adopt this convention
hopefully soon).
- the script 'free_windows-env.mk' is no longer needed ... see the
docs about how this is overcome (again, tuhopuu will hopefully
also follow this route soon).
- the dlltool dir has it's own Makefile that builds all of the
needed stub libraries from the dll's in cvs.
I wasn't sure maybe the two license files should be put in the root dir
instead....
I also reformated the bf-members.txt file a little to nuke extra spaces and
newlines (as well as updated my entry to include my name instead of just
sirdude... It looks like there are some special chars in it that someone
should fixup, not sure what they are suppose to be though)
Kent
directories exist. They don't exist for a standard Python 2.3 installation
but they aren't needed either... this is really a hack, I wish someone
who understood the freeze process better could look at this.
- fix make/shell buglet to make use of --no-print-directory for cleaner logs
- Conform to the nan_subdirs.mk mechanism of building
intern/Makefile
- include nan_subdirs.mk again
This enables people to run make clean and all those other targets from the top
level dir again, 'make source' has been lost for now
I added a check if the output directory (../frozen) exists and will create
the directory if it does not exist.
This time I tested it also with the NaN make environment :)
Michel
is updated.
* modulefinder creates a file src.list which contains all the *.py files
needed for the python suport in Blender.
* Makefile.am is able to depend on the files listed in src.list.
* makemakefile_ac.py is reverted to the previous version. The old version
contained my first attempt to get the make freeze trick working. It was
wrong.
Michel
build systems (auto, msvc, PB) do not seem to have this problem.
- Change intern/python/freeze/Makefile to freeze outside of the source/ tree
- Change source/blender/bpython/Makefile to not build the frozen stuff anymore
- Change source/blender/bpython/intern/Makefile to use CURRENT_PYTHON_API
instead of FUTURE_PYTHON_API, damage control
I'm guessing its safe.
Basically it fixes a warning about typedef usage.
Reading the docs on gcc-3.2 it says that classes that use a typedef
need to have the keyword "typename" infront of them according to the c++
standard. Currently it only produces a warning but it may be an
error later, so I figured why not fix it.
Kent
*Fixed boolean buglet in the bsp lib. I have not committed any libraries so you'll
have to do this to see the changes in blender.
*Got the bsp test program working again with the new GHOST createWindow
interface.
*(Visual C++) Modfied some of the dsp in the intern library to NOT prompt when
overwirting old libs and header files. This is the /Y option after XCOPY
Cheers
Laurence.
didn't exist instead of installing it as it should.
- @../tools/cpifdiff.sh blenkey.h $(LCGDIR)/$(LIBNAME)/include/blenkey.h
+ @../tools/cpifdiff.sh blenkey.h $(LCGDIR)/$(LIBNAME)/include/
sgefant
clicking on it. The mouse up was already generated. They
are now balanced by mouse downs. The behaviour of the
OSX target now matches that of the WIN32 target.
Maarten
---------------------------->8----------------------------------------------
g++ -DHAVE_CONFIG_H -I. -I../../../blender/intern/bsp -I../.. -I../../../blender/intern/container -I../../../blender/intern/moto/include -I../../../blender/intern/memutil -I/usr/freeware/include -g -funsigned-char -c ../../../blender/intern/bsp/intern/BSP_CSGMesh.cpp -MT BSP_CSGMesh.lo -MD -MP -MF .deps/BSP_CSGMesh.TPlo -DPIC -o .libs/BSP_CSGMesh.lo
/usr/freeware/lib/gcc-lib/mips-sgi-irix6.5/3.0.4/include/g++/bits/stl_iterator.h: In
constructor `std::__normal_iterator<_Iterator,
_Container>::__normal_iterator(const std::__normal_iterator<_Iter,
_Container>&) [with _Iter = const BSP_MFace*, _Iterator = BSP_MFace*,
_Container = std::vector<BSP_MFace, std::allocator<BSP_MFace> >]':
../../../blender/intern/bsp/intern/BSP_CSGMesh.cpp:270: instantiated from here
/usr/freeware/lib/gcc-lib/mips-sgi-irix6.5/3.0.4/include/g++/bits/stl_iterator.h:474: cannot
convert `const BSP_MFace* const' to `BSP_MFace*' in initialization
*** Error code 1 (bu21)
*** Error code 1 (bu21)
*** Error code 1 (bu21)
*** Error code 1 (bu21)
---------------------------->8----------------------------------------------
The line in question that causes the error is the middle one in this group:
vector<BSP_MFace>::const_iterator f_it_end = FaceSet().end();
vector<BSP_MFace>::const_iterator f_it_begin = FaceSet().begin();
vector<BSP_MFace>::iterator f_it = FaceSet().begin();
Dropping the 'const_' from that middle line enables gcc to compile the
file correctly (this is also consistent with what is going on with other
parts of the file, i.e., stuff that is returned from a begin() method is
declared as vector<BSP_MFace>::iterator instead of
vector<BSP_MFace>::const_iterator.
But I'll be honest: I have no idea what this code does, so if somebody with
better C++ skills wants to check it, then please do.
This change was also tested to compile and run on debian linux/x86
(well, booleans are broken right now, so I wasn't able to do too much
testing).
Chris
I took out the following from the includes in the intern dir that still had
it:
-#ifdef HAVE_CONFIG_H
-#include <config.h>
-#endif
Kent
--
mein@cs.umn.edu
too -- a user reported crashing while using decimation and rolling back
the modifications between rev 1.2 and rev 1.3 for this file fixed it.
Regards,
Chris
(to stop gcc compiler warnings) caused segfaults when
performing intersections (and possibly while doing other
boolean operations). I felt it was best to revert these
changes (I also got rid of the DOS line terminators in
the files). Some more info about the matter can be found
in the first 10 or so posts for Novembor on the old
mailing list:
http://www.soze.com/mailman/private/bf-committers/2002-November/date.html
Chris (Merry Christmas!)
+static void SleepTillEvent(Display *display, GHOST_TInt64 maxSleep) {
maxSleep is checked against -1 so it needs to have a sign.
This could even fix some stability problems since -1 is passed in sometimes
in the code.
Kent
--
mein@cs.umn.edu
IK_QSegment.h: In constructor `IK_QSegment::IK_QSegment(MT_Point3,
MT_Matrix3x3, double, MT_ExpMap)':
IK_QSegment.h:271: warning: member initializers for `MT_Scalar
IK_QSegment::m_length'
IK_QSegment.h:270: warning: and `MT_ExpMap IK_QSegment::m_q'
IK_QSegment.cpp:56: warning: will be re-ordered to match declaration order
IK_QSegment.h: In constructor `IK_QSegment::IK_QSegment()':
IK_QSegment.h:271: warning: member initializers for `MT_Scalar
IK_QSegment::m_length'
IK_QSegment.h:270: warning: and `MT_ExpMap IK_QSegment::m_q'
IK_QSegment.cpp:73: warning: will be re-ordered to match declaration order
Kind of goofy its just the order of which 2 private vars are defined.
Kent
--
mein@cs.umn.edu
Had to add an argument to Create_window... (it was missing the sterio flag)
Also cleaned up some variable overloading warnings...
(i.e. window shadows previous def)
Kent
(I noticed its not completely gone yet from the blender/source dir)
But its a big step in the right direction if it doesn't enable
all of the functionatlity already...
(Using cscope for LICENSE_KEY_VALID still turns up some stuff)
Kent
--
mein@cs.umn.edu
(adding)
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
also the Makefile.in's were from previous patch adding
the system depend stuff to configure.ac
Kent
--
mein@cs.umn.edu
batch file that freezes the Blender Python code.
I had to add a dummy.cpp file to the freeze directory to make MSVC
understand it should call the batch file in a post build step. If someone else
has a better idea, feel free to change it!
Maarten
+ python $(PYFLAGS) freeze.py -d -x os -x pprint -I $(SRCDIR) -o $(TARGETDIR) $(SRCDIR)/VRMLmain.py
Freeze the Blender module
(I got this from some other people on #blendersauce and truthfully I have
no idea if this is correct or not so if someone else could look at it
that would be great)
Kent
--
mein@cs.umn.edu
intern/make/msvc_6_0. Changed paths in all these files to build to
lib/windows and use obj/window/intern for object files and other temporary
stuff.
Added project files for guardedalloc and blenkey (in keymaker directory).
blenkey still assumes openssl being installed in lib/windows.
The only thing not automated is is the frozen Python stuff.
Maarten
This saves other OSX developers the trouble of manually running ranlib.
This is not a good solution (because I don't know the correct one) but it works.
Maarten
I also moved it so that it gets put in:
$(OCGDIR)/blender/bpython/$(DEBUG_DIR)libfrozen.a
and removed the stuff from the readme on how to do it by hand.
(I made one other small change and that was to comment
out the ssr target on solaris and freebsd in source/Makefile
I forgot to commit it yesterday)
Kent
--
mein@cs.umn.edu
removed the action library not used by Blender (actually this was that start of
UNDO in Blender that was never finished)
Maarten (mail@maartengribnau.com)
it wasn't detected before because everyone had a blenkey.h in their dirs also.
also I added an #include "key.h" to key_internal.h so that I didn't
get problems with undefined type byte...
mein@cs.umn.edu