qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-06 05:01:26 +00:00
Code Issues Projects Releases Packages Wiki Activity
7473 commits 17 branches 20 tags 29 MiB
Commit graph

84 commits

Author SHA1 Message Date
Bill Currie
90de6b0499 Add a movepi instruction to support indirect moves of fixed size. 2011-03-09 10:29:24 +09:00
Bill Currie
18005cc80d Rename move to movei. 2011-03-09 10:28:40 +09:00
Bill Currie
3c9991364b Add a "void return" instruction. 
I got fed up with always having to explicty return something.
 2011-01-13 00:29:56 +09:00
Bill Currie
75ec6bf244 Clean out some unnecessary types from the progs engine and clean up the mess. 
This is a nasty commit, sorry, but 99% of the commit is interdependent.
 2011-01-10 12:25:31 +09:00
Bill Currie
c1b12ed6ec fix a couple of gcc 4.5 issues that Despair pointed out (I don't have it yet) 2010-05-27 12:21:50 +00:00
Bill Currie
00e590f5f5 Really nail down the args param in rua_obj_msg_sendv(). 2010-01-13 06:36:16 +00:00
Bill Currie
5d6977288f implement bounds checking on all opcodes 2010-01-13 06:19:50 +00:00
Bill Currie
ca0bbd909c correctly save and restore the parameters when calling +initialize. fixes the segfault in the qwaq test. 2007-06-09 13:44:06 +00:00
Bill Currie
57bd43fc52 make the watchpoint (optionally) conditional 2007-05-08 02:25:01 +00:00
Bill Currie
fe4a4a9e55 "hardware" (haha) watch points (one!) in progs. only catches changes done by progs, and the expression parser is as flaky as anything, but it's better than nothing :) 2007-05-08 02:04:47 +00:00
Bill Currie
124506fda2 Fix the access alignment bug on 64 bit archs. 2007-04-07 01:41:23 +00:00
Bill Currie
5b761bac83 make progs related code a little more consistent with its int type usage 2007-04-06 00:47:41 +00:00
Bill Currie
99c0954b47 the big dso visibility patch :). Sure, we have to have unique names for static builds, but with controlled visibitly we should get faster program loads (although this isn't C++, so it's not as bad) and complex plugins are cleaner. 2007-03-10 12:00:59 +00:00
Bill Currie
3fb03fc2be hah, should have been using __attribute__((used)) all that time (rather 
than __attribute__((unused))). fixes the missing console in -x11
 2005-08-04 15:27:09 +00:00
Bill Currie
4a21eb02c5 heh, slight oversight when implementing rcall using the param pointer 
trick. ran into the exact same problem as passing addresses of locals to QC
functions. well, actually, that's exactly what it was, but done in the
engine. anyway, wound up potentially overwriting the params passed to the
function. fortunatly, due to how rcall works, only the first two params are
an issue.
 2005-06-14 13:40:34 +00:00
Bill Currie
762aefc9b1 implement Raven's calling convention (rcall: first 2 params in opb and 
opc). should give a general speedup to most progs.
 2005-06-12 09:54:01 +00:00
Bill Currie
f65c25a7c2 oops, signage error 2005-05-01 08:30:30 +00:00
Bill Currie
780fb41858 make a common call function and fix pr_trace such that it autoclears only when the execution session it's set in is exited 
This is an imperfect revision of history.
 2004-11-07 03:00:00 +00:00
Bill Currie
33bc7858f7 avoid using the hunk in the progs engine (except for the override-able load and alloc functions) so tools (qfprogs and qwaq) don't need to create it. 
register only the first instance of a progs provided string

This is an imperfect revision of history.
 2004-11-02 04:59:00 +00:00
Bill Currie
633dada682 quaternion conjugate support (engine side) 2004-04-08 04:57:17 +00:00
Bill Currie
8d6f634f30 engine side quaternion support 2004-04-08 00:56:30 +00:00
Bill Currie
ac67a50a9a handle x % 0 by giving a 0 result 2004-02-20 00:25:08 +00:00
Bill Currie
a821f048a0 change FNC to FN and fnc to fn for function based opcodes 2004-02-11 06:52:19 +00:00
Bill Currie
c3f41e3e69 new opcode: state.f. same as state, but takes a 3rd float operand to 
specify the step for calculating nextthink. accessed using
[frame, think, step] (state is [frame, think])
 2004-02-11 01:43:33 +00:00
Bill Currie
24294c8778 make printing of global contents in PR_PrintStatement optional so 
disassembly output is cleaner
 2004-01-31 04:26:01 +00:00
Bill Currie
595280decb do the builtin function lookup at load time rather than runtime and store 
the function address in the progs function descriptor. this will speed up
calls to builtin functions, especially when ranges get introduced.
 2004-01-05 08:08:46 +00:00
Bill Currie
dfc83c1f5f add PR_PushFrame and PR_PopFrame so temp strings are easy to use for 
parameters to progs functions.

double the progs stack sizes (call depth of 32? eek)

allow one extra call on the progs stack :)

misc minor cleanups
 2004-01-05 07:10:32 +00:00
Bill Currie
bd561fafc3 o all progs strings are now stored in progs memory 
o  temporary strings automaticly get freed when the function they're
    created in (results of str + str and most strings returned from
	builtins). a way to keep temp strings will be provided later
 o	fix up qwaq to test the temp strings
 2004-01-03 08:43:57 +00:00
Bill Currie
cd8035319d PR_GetString should have been returning const char *. now does 2003-11-21 21:34:53 +00:00
Bill Currie
fea382d1d9 fix an accounting error 2003-09-15 21:13:13 +00:00
Bill Currie
96202a5984 bah, forgot to -1 the instruction pointer for jumpb 2003-08-24 07:23:12 +00:00
Bill Currie
afb42f5ed0 fix the unsigned comparison opcodes 2003-08-24 05:53:15 +00:00
Bill Currie
98beb38540 implement jumpb correctly 2003-08-22 19:48:36 +00:00
Bill Currie
fe3b544528 handle "return 0" a bit better. assumes location is is always 0, but it 
always has been (and always /should/ be anyway:)
 2003-08-13 17:27:34 +00:00
Bill Currie
6f519c7e0f don't bother doing the copy if the value to be returned is already in the 
return slot
 2003-08-13 17:07:05 +00:00
Bill Currie
dd9ab1e050 get unsigned mostly working 2003-08-01 21:20:04 +00:00
Jeff Teunissen
c9858f4f65 Preliminary "unsigned" support for qfcc and the gamecode engine. Should still 
be compatible.
 2003-07-26 21:07:51 +00:00
Bill Currie
f7d806e8d5 can now handle > vector sized parameters 2003-04-22 20:11:16 +00:00
Bill Currie
e732cf591d remove all references to OFS_* from the C code. only the macros in progs.h 
use them now
 2003-04-22 18:20:15 +00:00
Bill Currie
eb099ae0d3 make qf gcc 3.3 clean 2003-04-17 00:01:48 +00:00
Bill Currie
761a7546dd re-arrange things so __attribute__ and __builtin_expect are properly 
autoconfiscated so rcsid will continue to work with gcc 3.3
 2003-01-15 15:31:36 +00:00
Bill Currie
71196fedf6 make gcc more anal about prototypes, string constants and function 
visibility (ie, global functions must have a prototype)
 2003-01-06 18:28:13 +00:00
Bill Currie
836c469737 sys.[ch]: 
handle SIGFPE and allow the registered signal handler to do recovery
	rather than bail
progs.h, pr_exec.c, pr_load.c:
	if pr_faultchecks is 1 (0 is default), handle division by 0 gracefully
	by loading the maximum representable number into the answer
	Closes: #58
the rest:
	kill the SIGFPE stuff
 2002-11-13 19:26:44 +00:00
Bill Currie
ec0273e3dc do a pre and post pass for maps/mapname.cfg and use maps_default.cfg if 
maps/mapname.cfg doesn't exist.
 2002-11-08 02:43:04 +00:00
Bill Currie
012cc58eb4 move the debug related stuff from pr_exec.c to pr_debug.c 2002-10-23 20:42:02 +00:00
Bill Currie
80e85226b9 pr_xstatement now always reflects the currently executing statement making 
debugging easier in the event of a segfault
 2002-10-22 15:07:54 +00:00
Bill Currie
6203d450eb resture pr_xstatement on return from PR_ExecuteProgram so recursive calls 
don't mangle things with the new semantics of builtin calls
 2002-10-22 02:51:57 +00:00
Bill Currie
aad04efe25 move the correct amount of bytes (I keep forgetting the * 4 :P) 2002-10-21 16:38:45 +00:00
Bill Currie
be4b8096d4 two new instructions: move and movep: basicly memmove. qfcc support coming 
as soon as I get it working :)
 2002-10-16 06:44:41 +00:00
Bill Currie
1f0b6e84a5 change pr_obj_msgSend and pr_obj_msgSend_super so they either call builtin 
methods directly or setup the progs struct for a normal function call to
call a progs method rather than recursively calling PR_ExecuteProgram. This
will reduce method call overhead (both cpu and stack usage), fix the loss
of tracing when calling a method and makes it possible to break out of the
progs engine simply if threaded progs are ever implemented.
 2002-10-15 18:53:33 +00:00