summaryrefslogtreecommitdiff
path: root/xc/lib
AgeCommit message (Collapse)AuthorFilesLines
2004-03-26Make the handling of drawableType consistent with recent changes inidr1-2/+2
glxext.c. Consider two GLcontextModes as "the same" if they have a non-empty drawableType union. When converting a GLXvisualConfig from the server to a GLcontextModes, set both GLX_WINDOW_BIT and GLX_PIXMAP_BIT. This matches the expected behavior. This also means that DRI drivers do not have to set GLX_PIXMAP_BIT in the GLcontextModes they return from their __driCreateNewScreen function. This makes sense: the drivers don't really support direct-rendering to pixmaps!
2004-03-21move i810 to new interfaceairlied1-1/+1
2004-03-21Added s_auxbuffer.[cho] to Imakefile.inc.fxkuehl1-0/+7
2004-03-13Fixes a bad interaction between the new libGL and drivers built in theidr1-14/+41
trunk that haven't been converted to the new interface or drivers built before the driinterface merge (i.e., the ones that ship with every XFree86 4.x.x release). Basically, FillInVisuals was excluding modes whose drawableType didn't match GLX_WINDOW_BIT exactly. This means that a mode would get tossed if drawableType was (GLX_WINDOW_BIT|GLX_PIXMAP_BIT), which is what every server-side driver sets! The proper test is to just make sure that GLX_WINDOW_BIT is set. Reported by: Numerous folks on dri-devel and dri-users.
2004-03-05don't use bitfields.alanh1-4/+4
2004-03-03Enable use of the new interface in the Radeon driver.idr1-1/+1
2004-03-03Merged driinterface-0-0-3driinterface-0-0-3-20040303-mergeidr14-588/+1439
2004-02-26Removed savagevb.[ch].mach64-0-0-7-branchfxkuehl1-7/+0
2004-02-24a little morealanh1-2/+2
2004-02-24get rid of BuildXIElib and BuildPEXlibalanh1-8/+0
XFree86 bug: Reported by: Submitted by: Reviewed by: Obtained from:
2004-02-22Added Imakefiles to build the Savage DRI driver from Mesa sources.fxkuehl2-0/+188
2004-02-12Fontfile.diffkeithw3-673/+844
XFree86 bug: Reported by: Submitted by: Reviewed by: Obtained from: XFree86.org, via Roland Scheidegger
2004-02-08There's a client-side and a server-side portion to this commit. Bothidr3-3/+35
do essentially the same thing. If a 'glGetString(GL_VERSION)' call returns a string with a version greater than is supported in protocol, wrap that version string with as with 'sprintf(s, "%s (%s)", supported_version, returned_version);'. Currently the top version supported is 1.2. The jury is still out as to whether or not this is the right thing to do. The other option is to just replace the returned version string with the highest supported version string. This would lose any extra informantion in the version string, but may be safer. XFree86 bug: 1147 Reported by: Andreas Stenglein
2004-02-07Add support for real glDrawArrays protocol. There will be problemsidr3-39/+341
with older versions of the server-side code, so an environment variable, LIBGL_NO_DRAWARRAYS, was added to disable the use of the real protocol.
2004-02-07Refactored the body of __glXSendLargeCommand to call a new functionidr2-32/+59
called __glXSendLargeChunk. The purpose of this refactor wasn't to reduce code size, but instead to provide a new utility function to be used instead of __glXSendLargeCommand. There may be cases where using the new function could eliminate the need to allocate extra memory.
2004-02-07Folks, that's what happens when you're working on 16 things and once,idr1-2/+2
and want to commit it incrementally. You have to actually TEST the incremental commits! Bad idr, no biscuit!
2004-02-07Removed the unused define __GL_BOOLEAN_ARRAY. Re-implementedidr2-31/+13
__glXTypeSize as a macro that does a table look-up.
2004-02-06Client-side vertex array state is now tracked as an array of arrayidr4-422/+293
state structures instead of collection of named array state structures. This greatly simplifies the code to emit array elements.
2004-01-31Added bridge functions to convert EXT vertex array pointer calls to GLidr2-0/+85
1.1 calls. Added code to __indirect_glEnable and __indirect_glDisable to call __indirect_glEnableClientState or __indirect_glDisableClientState if 'cap' is one of the vertex arrays. To work around confused / buggy applications, this is done for all vertex arrays, not just the ones supported at the time of the original EXT extension. XFree86 bug: 1131
2004-01-30Fixed a problem matching visuals when GLX_DONT_CARE is used for theidr2-1/+8
GLX_VISUAL_CAVEAT and the only matching visuals have a non-GLX_NONE GLX_VISUAL_CAVEAT. Reported by: Alan Hourihane
2004-01-30silence warningalanh1-1/+1
2004-01-28link driverfuncs.h tooalanh1-0/+1
2004-01-28fix driverfuncs for OSmesaalanh1-5/+6
2004-01-28restore backwards compatibilityalanh1-6/+5
2004-01-26My previous fixes created a binary compatability problem. The issueidr8-280/+399
was that adding extra fields to __GLXvertArrayState and __GLXvertexArraPointerState changed structure offsets in __GLXcontext. This, obviously, caused problems. The old 'state' field n __GLXcontext has been deprecated. State is now tracked via a new structure that is pointed to by the client_state_private field at the end of the __GLXcontext structure. This gives us the freedom to make whatever changes we want to the client-state without breaking DRI drivers. XFree86 bug: 1092, 1092
2004-01-21Added enums for a couple rare SGIX extensions.idr1-3/+17
Also added GL_YCBCR_422_APPLE. I'm not sure it's strictly needed, since the type has to be GL_UNSIGNED_SHORT_8_8_APPLE or GL_UNSIGNED_SHORT_8_8_REV_APPLE. It is included for completeness.
2004-01-21Build and link driverfuncs.c from Mesa. This probably isn't the rightidr1-5/+13
way to do it, but it gets the build going again. :)
2004-01-20A few more vertex array fixes for EXT_secondary_color andidr3-0/+71
EXT_fog_coord that were previously missed. I'm 99% sure that I've got everything now. :) XFree86 bug: 1091
2004-01-20indirect.h now pulls double duty. Depending on howidr3-3561/+708
GENERATE_GLX_PROTOCOL_FUNCTIONS and the glxproto_* macros are set, it will either generate prototypes are generate the functions. Right now only g_render.c should have GENERATE_GLX_PROTOCOL_FUNCTIONS set. This cuts out about 2500 lines of code.
2004-01-19Forgot to commit this.keithw1-12/+5
2004-01-19Generate correct GLX protocol for multitexture with vertex arrays.idr2-7/+23
XFree86 bug: 1092
2004-01-19fixes from XFree86 trunkalanh5-39/+39
2004-01-19Added client-side and server-side GLX protocol support foridr8-33/+433
glPointParameteriv (for ARB_point_sprite, NV_point_sprite, core GL 1.4), secondary-color related functions (for EXT_secondary_color and core GL 1.4), and fog coordinate related functions (for EXT_fog_coord and GL 1.4). The reviewers primarilly reviewed the architectural changes to g_render.c (client-side and server-side versions). These changes were tested with the Mesa demos seccolor and spriteblast. XFree86 bug: 1091 Reviewed by: Brian Paul, Roland Scheidegger, Keith Whitwll
2004-01-16Enable the multisampling extensions by default. If the client-sideidr1-3/+3
driver doesn't support multisampling, then the DDX driver shouldn't export any multisample visuals / fbconfigs. Furthermore, the driver can disable multisampling by not enabling GL_{ARB,SGIS}_multisample. SGIX_visual_select_group is enabled be default since it has no impact on either the client-side driver or the application.
2004-01-16Minor header file clean-up. This files don't need anything fromidr3-3/+3
packsingle.h or packrender.h.
2004-01-14Added *numerous* missing enums. I think I got them all, but havingidr1-26/+46
some more eyes go over it would be helpful.
2004-01-13Added several missing enumerants to __glTexParameterfv_size. Oops.idr1-0/+12
2004-01-13Added indirect support for GL_EXT_multi_draw_elements (which is partidr5-0/+37
of GL 1.4). The GL_SUN_multi_draw_elements extension string is also now exported.
2004-01-06move includes so we don't symlink lots of x86 assembler code.alanh1-4/+4
2004-01-05ss_vb* are no more.keithw1-13/+5
2004-01-05missed some parts of that last commitalanh1-0/+4
2004-01-05add new files to build t_vertex.[ch]alanh1-0/+3
2003-12-28Imakefile fix.alanh1-6/+6
2003-12-10Define new ExpatLibrary macro, which is used in the DRI drivers to link inanholt10-10/+10
expat. Defaults to shared, but can be overridden in host.def, which is done on linux so that binary snapshots will continue to be statically linked. Fixes build on FreeBSD. Approved by: Alan Hourihane <alanh@fairlite.demon.co.uk>
2003-12-10fix the EXPATINCLUDES for FreeBSDalanh1-1/+1
2003-12-09remove old GlxSoProf profiling mechanismkeithw14-215/+20
2003-12-09merge newmesa branch to the trunknewmesa-0-0-1-20031209-mergealanh359-123887/+1567
2003-12-08add expat library to build.alanh3-0/+191
So we can build this ourselves for static inclusion into libGL.
2003-12-08Extend indexed (elt) primitives where possible.keithw1-7/+23
2003-11-28Disable stencil write when stencil test disabled. Should hopefully fixkeithw1-0/+3
some lingering problems on the i845.
generated by cgit v1.2.3 (git 2.39.1) at 2025-08-26 10:56:36 +0000