[Gtkradiant] GtkRadiant commit Fri Aug 31 11:06:48 CDT 2001
Fri, 31 Aug 2001 11:06:48 -0500
Fri Aug 31 11:06:46 CDT 2001
RCS file: /cvs/GtkRadiant/docs/developer/CHANGES,v
retrieving revision 18.104.22.168
retrieving revision 22.214.171.124
diff -r126.96.36.199 -r188.8.131.52
> map module successfully loaded and saved q3dm1, the saved file was then loaded back into 1.1-TA without problems
> rebuilt and checked on win32
> merged IMap back in trunk, fixed some memory conflicts on win32
> Took me a lot more time than I would have liked to, but there's a script tied to the CVS server now, which will post on this list a diff of the docs/developer/CHANGES file whenever it gets updated. This will probably be very handy for me since I'll only have to put update information in the CHANGES file instead of having to post on the list too.
> The script is likely to be a bit laggy, or miss some features (for instance I'd like to extract the branch name .. anyone know how I can get the branch name (Alpha/IMap/HEAD) from the version number?
> PS: I can email this script to anyone who would like to have a look
> did more work on map module, one big chunk of work left: the core should broadcast interface requests to plugins
> when it doesn't know how to do it by itself..
> new doxygen patch, generates output from core (libs/ include/ and radiant/)
> removed Makefile, use cons damnit!
> automated documentation via doxygen, new scripts and content
> patch for CHAR to xmlChar conversion (xml2 consistency)
> the map module starts to look like something, cleaned up the interface stuff
> started moving the actual code out in the module and removing it from the core
> lots of issues raised on the way, some structures to export, and the macro scheme to access API functions more easily
> it compiles right now, but won't run because it's missing a lot of things .. the process simply happens to be "under way"
> patched more path code, to look for stuff in "bitmaps/" and "modules/" instead of "tools/bitmaps" and "tools/modules"
> modified the Construct files accordingly
> merged in radiant/ishaders.cpp diff into plugins/shaders/shaders.cpp (PreloadShaders)
> merged in radiant/lbmlib.cpp diff into plugins/images/lbmlib.cpp (Sys_FPrintf)
> checked the .rej and patched a few remaining things
> the diffs are space/tab sensisitive, and we used the "beautify source" a bunch of times, so it's a bit fucked now
> next time, generate the diffs not space sensitive..
> map loading is fucked, "textures/" prefix issue?
> merging recent changes from Alpha branch into the Trunk
> this could not be done with a regular cvs merge because we already did a cvs merge of Alpha into trunk some time ago
> manually built a diff between the current Alpha (now tagged Merge-1_1-TA_1-nightly)
> and the Alpha we had right after the former cvs merge: -r Alpha -D 2000-05-28
> binary files ignored in the diff, only going for source stuff
> built with diff -Nru Reference/ Current/
> then patch -p1 < patchfile
> next, started rebuilding:
> big manual updates were in vfs.cpp and texwindow.cpp
> cleaned up some VFS stuff .. it had an absurd QERAppFileSystem / QERPlugFileSystem scheme
> the ISSetup has not been copied over from Alpha
> OK check the Construct files
> OK radiant/ishaders.cpp no longer exists .. apply the patch on the shader module
> OK radiant/lbmlib.cpp no longer exists .. in the image code?
> OK look at the *.rej files
> fixed project dialog to behave right
> proper .def scanning
> fixed shader loading with VFS and mod stuff
> added a local to texwindow.cpp GSList *l_shaderfiles
> holds the names of the active .shader files
> modified q3map to read "fs_basepath" and "fs_game"
> .def files in the media need to move to <fs_game>/scripts/
> rename entities-TA.def to entities-ta.def
> updated cmdlib's Q_Exec to be more widely used through Radiant (during watchbsp.cpp cleanup)
> patching in MarsMattel's code for mod support, and started fixing:
> added m_strFSBasePath m_strFSMain m_strFSGame to g_qeglobals to match Q3's filesystem
> reworked the project file dialog
> changed the way we load and initialize eclass and shaders to work with mod code
> updated VFS initialisation code, cleaner and better console output
> the "game" key in the project file is no longer relevant, only "dir" is
> (if "dir" is not present, then no mod support, vanilla Q3)
> changed the loading of the .def files to scan in scripts/, you might need to move your entites.def to use
> cleanup .. (search where "basepath" is used for instance)
> using fs_game when calling q3map
> fixing project dialog to behave right
> shader loading using VFS functions (seems to work again but I'm not sure)
> win32 ver. might be slightly broken
> .def scanning, don't scan ALL .def
> added version and build info to the log file + current timestamp
> fixed selection of misc_model when viewed as a bounding box
> cons script for q3map building
> added general GtkRadiant versioning (version.h) to q3map
> nightly setup on linux:
> using the right install path (with GtkRadiant's version name)
> cleaned up options to only the stuff relevant to nightly
> fixed Radiant and core binaries path in setup
> added some template processing of setup.xml (similar to what is being done on win32)
> applied ^Fishman's path
> inclusion of version.h and aboutmsg.h moved to qe3.h
> changed base path location process (in most cases it will prompt)
> fixed the path construction to initialize according to the new layout
> - http://zerowing.idsoftware.com/bugzilla/show_bug.cgi?id=115
> fixed wake up crash on linux
> - http://zerowing.idsoftware.com/bugzilla/show_bug.cgi?id=117
> - http://zerowing.idsoftware.com/bugzilla/show_bug.cgi?id=120
> fixed installer bug on win98, was a problem with cygwin config
> - http://zerowing.idsoftware.com/bugzilla/show_bug.cgi?id=64
> cleanup and fixed
> - http://zerowing.idsoftware.com/bugzilla/show_bug.cgi?id=100
> can't locate the "textures: in use" problem, has been fixed already?
> now selecting the right entity in the list, had to go around an inifinite recursion problem
> (i.e. selection message in the entity class list causes UpdateSel recursion)
> - added botclip to missionpack/common.shader
> - backported cons scripts to Alpha branch. Type 'cons' at the head to build
> regular makefiles should soon be outdated..
> - updated the IS script (Gtk changes and and BACK problem)
> - bunch of fixed to make it build on win32 against new STLPort
> using an STLPort configured locally in GtkRadiant (with stl_config.h, new file)
> - fixed GDI leak affecting text widgets, specially the console and the entity inspector
> - more win32 project file cleanup, removing libs/libxml2 from the tree
> - importing back "gtkr_list.h" from 1.2, made Alpha branch STLport compliant too
> - added libxml2 as an external dependency. libxml2 should be installed as a seperate
> directory on win32. dynamic linking now instead of static previously.
> - cleaned up q3map win32 project file, removed opengl dependency
> (the drawflag command line is inoperant now .. I don't think it was used anyway?)
> - checking in Gef's doxygen files