Activity for crossdjgppv1
-
Jason Stevens created a blog post
I used Sydney (Microsoft/OpenAI) for Fixed Multiply
-
used ai to get fixed multiply
-
crossdjgppv1 released /GNU Source/gcc-1.38.tar.bz2
-
-
-
Doom 3 BFG doesn't have the titlepic resource in doom2 for some reason. So just load the menu picture in it's place.
-
Super bare minimum for Chex Quest 1. All it does is open the wad, no verbage or level cap put in place.
-
drawspan for Watcom needs TASM. Which I don't have.
-
On the Atari ST, binary mode is 4. Don't ask me why.
-
un-merge the fstream wad code. oops.
-
SO.. apparently the endian stuff doesn't work righ ton 68k.
-
m68k GCC 2.5.8 has some error with am_map. I'll have to investigate some if it's something that is even used.
-
Restore the Watcom math from Frank Sappone. I tested with Watcom 10.00 and -timedemo demo1/2/3 for Doom v1.9's commercial WAD. Yay.
-
Update the DJGPP v1 Makefile to seperate out the assembly. This works on E1M4 for all versions. gcc 1.27 needs to use the assembly fixed math.
-
Make the i386 assembly conditional in the makefile.
-
added absolute fix from prboom so i386 assembly works using gcc 1.x and 2.1
-
-
-
-
-
-
-
-
-
-
-
-
Neozeed committed [8ccf75]
Changes to compile under MinGW TDM gcc 5.1
-
Neozeed committed [c8dc21]
GNU G++ 1.39.1 source (uses GCC 1.39 for base)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
since that o2c utility isn't free I did more digging and found some assembly by Lee Killough. Although it uses a further modified engine, only drawspan is currently working, but it changed performance from 2691 ticks to 1622 ticks! The missing component is _dc_texheight from R_DrawColumns.
-
Add in the assembly version of R_DrawColumn and R_DrawSpan. Dropped timing from 2691 ticks to 1273! Now GCC is within 90 ticks of Open Watcom v2! Currently uses the non-free o2c OBJ 2 COFF converter.
-
Change from register to stack convention... Maybe split the makefiles?
-
make watcom builid again
-
too many initializers for the name table.... oops!
-
minor fixes for using ancient GCC, and the nullified sound code.
-
Add in the allegro lib directly. Built with the DJGPP 2.05 runtime. Modified the makefile to call into the directory. Tested with GCC 2.7/3.4/4.0
-
Added inline assembly for GCC, and assembly for Watcom. Tried to jump the stack to 1MB. Also included the bits from gcc2721 libgcc as fragments seem needed between various gcc versions for unknown reasons. Oh yeah and the music now plays!
-
added in the allegro keyboard and more hooks. slides in the makefiles and all that jazz. I'm not sure about m_swap suddenly being signed now....... :| 1000% untested and unbuildable for the x68000
-
Added sound/music via Allegro 3.12 It's highly unstable, but for the most part it works. It really doesn't like GCC > 2.something I'm not sure of just how old the cutoff it is expecting. This is for DJGPP v2.01 and the digitial audio components were taken from DOSDoom 0.2, and the music from DOSDoom 0.6 I converted the timer over to the allegro timer to make the MIDI component happy, although the digitized sounds didn't care at all. I think I have to use the allegro keyboard, and dump a...
-
Added in GCC 1.27 aka first i386 version of GCC
-
rename gcc_v1 to gcc_v139 now that we have gcc-1.27
-
More changes to get the pre-processor & gcc driver compiling under Windows. It built DooM, so I think that's a good test.
-
-
Added the support files for the x68000. I've verified it compiles, but haven't tested it yet.
-
Changes to ar to have it compile on MinGW
-
Using the assembly version of FixedMul & FixedDiv for the i386. Tested with DOSBox
-
Other fixes for GCC 1.27 & GO32. verfified to run on DOSBox
-
Changes to allow DooM to compile using GCC 1.27 - circa 1988. div.s was taken from 386BSD
-
Modifications to build on MinGW / TDM GCC 5.1 targeting BSD a.out / DJGPP v1
-
Initial import from GCC 1.27
-
Added some experiments in reading the music file externally rather than in the data file.
-
Updates updates...
-
Update Bison to use the current working directory for the gramar files. It's cheating but that's fine by me.
-
Changes from DJGPP + changes I made to compile with MinGW / GCC 5.10
-
Initial import of GCC 2.1
-
Build project & files for Visual Studio 2015.
-
Catacomb 3D data files from http://cd.textfiles.com/softkeyep/PROGRAMS/DOSGAMES/ARCADE/CAT3D/ Also a build of SDL 1.25 done with Visual C++ 2015.
-
Changes to remove the C++ style variables & struct declarations. I also changed the file modes to rb for MS-DOS/Win32. Also changed some long long stuff to int, as the original source was 16bit, I highly doubt it needs 64bit integers. It compiles and runs tested on GO32 in DOSBox, and the MinGW builds are still working with emulated Adlib!
-
Initial import from NotStiller https://github.com/NotStiller/Catacomb3D
-
Lame hack to let cccp.c always accept C++ style of comments, as I'm tired of doing the hard re-formatting of course from C++ style to C
-
Addded the Linux clock to the OS/2 port, and the keyboard monitor I had used in QuakeWorld/2 .. Although I need to remember how to set the extended key status. It's enough for letters and the escape key so you can quit using the menu now.
-
Add some slightly better demo skip logic. v1.1 has version 1 & 2 demos. v1.2 has 2 & 3. version 1&2 crash this engine hard.
-
Better fix for WIMINUS, now the Doom v1.1 registered WAD can change levels.
-
Let the menus work for the DooM v1.1 commercial wad. There is no nightmare mode, so detect it.
-
Minor changes to make older WADS directly playable. v1.0/v1.1 won't work from the menu as there is no nightmare mode...
-
The STTMINUS thing was added no doubt pretty late.. So I put in a little logic to detect if it's there, and now I can load older wad's.
-
update path, & verify that doom works on GCC 1.39
-
Add in some basic support for Watcom. It compiles and runs with 8.5 *HOWEVER* fixed point math is broken. There is no long long type. But newer versions of Watcom work OK. Also for watcom r_data has a bool type in a struct that watcom takes as one byte, but GCC has as a word (4bytes).
-
let the demos run, to get timings from OS/2 under QEMU (about 486DX4 100 speeds for me)
-
Super preliminary support for EMX OS/2 (needs GCC 2). It's enough to get video working.
-
Add in some minor changes to compile on Watcom v8.5 (there is no long long type, so the pure C version of the fixed point math doesn't work)
-
Add in changes to build for EMX. Tested with GCC 1.40 targeting MS-DOS
-
-
-
-
-
-
-
First release!
-
Modifications to build on MinGW32. I wanted a version of GNUmake that didn't have external DLL dependancies.
-
Initial import of GNU Make 3.74 from Fri May 19 16:36:32 1995
-
Merged in changes for DJGPP, plus modifications to compile on MinGW32. Moved many files to the 'old' directory to prevent a conflict in the includes which resulted in a failed build.
-
Inital import of GNU GCC 1.39 from Wed Jan 16 15:34:38 1991
-
Modifications from DJGPP and updates to compile for MinGW32. Added a GNU doprint implimentation.
-
Initial import of GNU GAS from Fri Jan 4 12:48:22 EST 1991
1 >
