Commit graph

175 commits

Author SHA1 Message Date
Bill Currie
bc1b483525 Nuke the rcsid stuff.
It's pretty useless in git.
2012-04-22 10:56:32 +09:00
Bill Currie
d910c14935 Clean up some doxygen warnings. 2011-07-10 19:12:07 +09:00
Bill Currie
0f7390dd60 Clean up all the "set but not used" warnings.
gcc on my system is failing to treat this specific warning as an error :/
2011-06-19 10:48:02 +09:00
Bill Currie
b522853765 Ensure the target of def_op relocs gets updated.
These occur only when switch statements use jump tables, thus why the
problem went unnoticed :/
2011-04-08 17:09:40 +09:00
Bill Currie
92ba110a87 Ensure a field def is emitted for @this. 2011-04-03 13:03:48 +09:00
Bill Currie
7f67e6eb17 Refactor the def handling code a little. 2011-04-03 13:03:22 +09:00
Bill Currie
6d0b13dbf0 Resolve types whenever possible. 2011-04-03 13:02:04 +09:00
Bill Currie
9b62c8c126 Rename ty_type_e to ty_meta_e and type_t.ty to type_t.meta 2011-03-22 16:06:47 +09:00
Bill Currie
f1d9570dc8 Fix constant value type handling.
Rename immediate.[ch] to value.[ch] and clean up the mess.

Add convert_value() to convert the type of a value (only scalar types).
2011-03-22 12:24:39 +09:00
Bill Currie
505bde308e Fix the handling of @system defs.
When treating a def as external, actually say it is external.
2011-03-07 20:53:27 +09:00
Bill Currie
8dbf8a97ea Be more informative with ext/def internal errors. 2011-03-07 20:52:31 +09:00
Bill Currie
49e446b637 Make field defs global.
It turns out that field defs need to be global after all, so use separate
field def tables in the linker.
2011-03-07 14:45:50 +09:00
Bill Currie
c4e9a45d91 Do not lose the entity field defs. 2011-03-06 15:40:07 +09:00
Bill Currie
97b1ceceb0 Fix line number info.
line number info is now 1 based rather than 0 based (to better detect when
a function does not have line number info).
2011-03-05 15:40:08 +09:00
Bill Currie
7c95913c61 Avoid segfaulting with bad string relocs. 2011-03-05 09:07:16 +09:00
Bill Currie
fe7be543f8 Correctly update the offset of code relocs. 2011-03-05 08:39:47 +09:00
Bill Currie
c3f1c186f8 Resolve types of .return etc. 2011-03-05 08:36:55 +09:00
Bill Currie
e62163e569 Nuke any invalidated relocs. 2011-03-04 21:46:19 +09:00
Bill Currie
2565e0db32 Update reloc targets when linking. 2011-03-04 21:29:47 +09:00
Bill Currie
e04f6266b1 Nuke def_list from defref_t.
It didn't work :/.
2011-03-04 20:40:01 +09:00
Bill Currie
2211a807d4 Do not relocate builtin functions. 2011-03-04 18:46:25 +09:00
Bill Currie
9a96febdf5 Fix the hordes of bogus defs.
They were caused by unused external defs, so not processing unused externs
in the first place is the simplest solution.
2011-03-04 14:51:09 +09:00
Bill Currie
7b7f4c70e0 Correct non-external def processing. 2011-03-04 14:33:03 +09:00
Bill Currie
c5dac6d1d1 Transfer the type encoding.
oops
2011-03-04 14:29:22 +09:00
Bill Currie
e525727a24 Add a null type descriptor to the linker output. 2011-03-04 11:38:33 +09:00
Bill Currie
3a828ca5c2 Ensure work spaces are always connected to their data. 2011-03-04 11:18:56 +09:00
Bill Currie
29985efcef Expose some linker functions to the rest of the compiler.
They proved to be useful for cleaning up the final compilation stages.
2011-03-02 20:48:55 +09:00
Bill Currie
065d3e0dd7 Partial linking for a single qfo file now seems to work. 2011-03-02 14:03:41 +09:00
Bill Currie
bfeece3255 Fix the bogus declarations of obj_msgSend and obj_msgSend_super. 2011-03-02 12:35:10 +09:00
Bill Currie
fda46a4aca Handle the special defs. 2011-03-02 12:26:53 +09:00
Bill Currie
54317965cc Handle relocs.
With this, the building of the work qfo seems to be almost finished, just
the special defs to go. Next step will be to stitch all the parts of the
work qfo into a real qfo.

However, still have to fix the relocs for the type descriptors, but they
are not critical to the linker's functioning.
2011-03-02 11:16:35 +09:00
Bill Currie
8669dc373f Add some documentation. 2011-03-02 11:13:44 +09:00
Bill Currie
65bd215058 Minor rearrangements. 2011-03-01 15:35:14 +09:00
Bill Currie
2cb1c74718 More progress on the linker.
Need to finish up defs and handle relocs, and then take care of special
defs (@self etc), but things are getting very close.
2011-03-01 15:18:47 +09:00
Bill Currie
c36919aae7 Fix verbosity and diagnostic messages. 2011-03-01 11:08:32 +09:00
Bill Currie
cfff2b1037 Add defs from near and far data spaces. 2011-03-01 10:52:27 +09:00
Bill Currie
21558864b9 Make defrefs a little more stupid.
Stupid code does little well :P

I forgot that spaces get realloced, and I don't remember why I had the def
list pointer separate anyway.
2011-03-01 09:19:04 +09:00
Bill Currie
1416b996d3 Ensure the type space pointer is cleared with each object file. 2011-03-01 09:18:09 +09:00
Bill Currie
30272968c9 Update the def's type pointer. 2011-03-01 09:17:32 +09:00
Bill Currie
1837518d05 Split the space processing into two passes.
This allows strings and types to be handled before code and data.
2011-03-01 08:48:58 +09:00
Bill Currie
eb11a19ed4 Split the type processing into two passes.
This allows all duplicate types to be "merged" before adding any new
types, thus avoiding any accidental duplication caused by possible
differences in order between defs and the type descriptors.
2011-03-01 08:46:02 +09:00
Bill Currie
05e6f00e72 Initialize the new type def.
Oops :P
2011-03-01 08:16:14 +09:00
Bill Currie
3761cddca6 Finally, some real progress on the linker.
There is still much to be done, but this seems to be a workable approach.
2011-02-28 23:18:07 +09:00
Bill Currie
8b661bbb89 Rename qfo_reloc_t's def to target and correct its usage. 2011-02-27 11:37:24 +09:00
Bill Currie
f362e74f98 Update the header. 2011-02-24 15:24:35 +09:00
Bill Currie
c10a551c84 Rename new_defspace() and defspace_new_loc(). 2011-02-24 13:07:36 +09:00
Bill Currie
66b255afbf First stage of the linker rewrite.
Strings, code and data spaces are "merged". Various offsets are updated.
Still to do:
 o  Merge reloc, def, function and line data.
 o  Merge type data.
 o  Update def and function type pointers.
 o  Resolve external defs and check for duplicate defs.
 o  Perform relocations.
 o  anything else I've forgotten :)
2011-02-24 12:18:35 +09:00
Bill Currie
a1ea492889 Remove some redundant fields from qfo objects. 2011-02-23 10:40:35 +09:00
Bill Currie
bceb98e5a4 Create structures for new qfo format.
This breaks a lot of existing code, so that code has been commented out
for now.
2011-02-17 15:02:09 +09:00
Bill Currie
0624408317 Move the diagnostic functions into their own file. 2011-01-24 21:54:57 +09:00
Bill Currie
ea3895805b Rewrite much to use symbols and symtabs. Gut emit.c. Massive breakage.
That which isn't rewritten is horribly broken. However, this does include a
nice mechanism for building QC structs for emitting data.

emit.c has been gutted in the spirit of "throw one away".

There is much work to be done to get even variables emitted, let alone
code. Things should be a little more fun from here on.
2011-01-17 22:34:41 +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
9ac15436d9 More docs.
Enhance qfprogs qfo debugging.

Make sure functions marked as extern don't emit anything. Fixes the segfault
when building klik.
2008-08-01 13:54:24 +00:00
Bill Currie
3b7c83a96f uncomment some lines that shouldn't have been commented (doesn't fix the bug, though) 2007-04-09 14:32:47 +00:00
Bill Currie
3cd01223dd don't try to fix up function refs too early. fixes klik's segfault on firing 2007-04-09 08:56:23 +00:00
Bill Currie
9509d0fd5e make qfo files typesize safe 2007-04-07 12:35:03 +00:00
Bill Currie
0382b05780 various unsigned long -> uintptr_t conversions 2007-04-04 11:22:48 +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
4d044bfea4 make entity.vector.x work properly by making relative field offset work. also don't emit a field def for the resulting immediate
This is an imperfect revision of history.
2004-11-12 10:49:00 +00:00
Bill Currie
ca46503607 add -Wsign-compare (default for -Wall only in C++) and fix up the warnings 2004-07-13 19:14:01 +00:00
Bill Currie
8a7bb9c4a0 fix a silly that messed up type mismatch errors 2004-04-12 23:46:21 +00:00
Bill Currie
0b61dae402 allow .imm fields (need a better test, I think) 2004-04-09 03:45:55 +00:00
Bill Currie
8f440c7911 linker.c:
I think I found the real reason doxygen gets confused
expr.c:
	add a check for x & y == 2 (or similar) doesn't work for constant y at
	the moment.
2004-01-22 01:04:48 +00:00
Bill Currie
41106a1f01 hopefully shut doxygen up about a "const char" function 2004-01-11 05:52:19 +00:00
Bill Currie
ef4ca0c798 some doxygen tweaks 2003-08-31 05:24:24 +00:00
Bill Currie
b38486d405 make rel_def_op work properly 2003-08-22 19:48:14 +00:00
Bill Currie
11c7c27f70 better relocations: pointers to structure members now work 2003-08-21 02:21:30 +00:00
Bill Currie
a2300387de fix a bunch of segfaults 2003-08-20 22:36:02 +00:00
Bill Currie
4a680bf981 grovel around for the source file and line of an undefined def to give
better diagnostics
2003-08-06 21:36:07 +00:00
Bill Currie
bfaa07c726 better type mismatch error reporting 2003-05-15 05:19:05 +00:00
Bill Currie
0ebf9d6afc more inline avoidance 2003-05-07 15:00:17 +00:00
Bill Currie
b6b5089ccc add a "nosave" flag to defs making it easier to make defs that don't get
saved automaticly
2003-04-25 17:00:22 +00:00
Bill Currie
f7d806e8d5 can now handle > vector sized parameters 2003-04-22 20:11:16 +00:00
Bill Currie
2f6fc5d80c general revamp so params and the return value use proper defs rather than
hardcoded locations. this will allow use of quaternions in the future.
2003-04-22 15:29:32 +00:00
Bill Currie
6a79ee57c4 the correct fix for overzealous linking: preemptively turf unused externs 2003-03-13 03:18:34 +00:00
Bill Currie
bb1b7c481c don't pull in a lib object file due to an /unused/ undefined def. 2003-03-13 01:50:49 +00:00
Bill Currie
b928f94440 produce a warning if both @self and self are used in the same progs 2003-02-24 20:10:33 +00:00
Bill Currie
5f75aa024a don't relocate a null function constant 2003-02-24 15:33:57 +00:00
Bill Currie
af44b129d6 don't store pointers to defs in the hash table as the defs can move around
due to realloc. can now link large progs
2003-02-21 22:31:40 +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
1cba65f6ce fix a gcc 2.95 warning 2002-11-04 02:24:28 +00:00
Bill Currie
7f69f593d3 configure.ac:
define default include and lib paths for qfcc
ruamoko:
	{cl_menu,game}/Makefile.am:
		conform to the new qfcc library linking rules
	lib/Makefile.am:
		install the libs to ${prefix}/lib/ruamoko
qfcc:
	linker.[ch]:
		support path searching for -llib and make linker_add_lib search for
		libfoo.a for -lfoo in the paths, or just a single dir search for the
		libname otherwise.
	options.c:
		support -L libpath and setup the default include and lib paths
		also change most strdup calls to save_string
	qfcc.c:
		check for foo.a as well as -lfoo when deciding whether the file is
		an object file or lib file.
2002-11-01 18:05:12 +00:00
Bill Currie
7af99f422f expr.c:
fix a braino that caused only sizeof expr to work
linker.c:
	print an error message if unable to open the lib
qc-parse.y:
	revamp expression parsing so casting is done via (type)expr rather than
	type(expr)
	make it so pointers to pointers don't need ()s
2002-10-31 22:58:59 +00:00
Bill Currie
9d8d570269 using Qseek on a pak sub-file should work better. this should make oggs in
paks work.
2002-09-27 04:27:19 +00:00
Bill Currie
d8a38ae648 don't fuss with function defs that aren't initialized 2002-09-19 06:01:06 +00:00
Bill Currie
ece9660099 get @system working (Closes: #52). I'm not happy with the name of this
operator and I'm very open to suggestions, but it seems to work. What
@system does is to mark a def in such a way that the linker will allow it
to be redefined /once/ giving the programmer control over where and when
a def is defined, but initializing it elsewhere. This is most useful when
working with v6 progs and separate compliation as the `system' defs need to
be in a certain order in certain locations or the resuting progs.dat file
will be incompatable with non-qf 0.5 servers.
2002-09-17 18:42:35 +00:00
Bill Currie
4a34960d95 a bit of code cleanup 2002-09-16 17:37:22 +00:00
Bill Currie
842a9d2eb0 linker.c:
don't seg on uninitialized fields
qc-parse.y:
	support @extern and @static blocks (eg, @extern { ... };)
2002-09-12 17:13:19 +00:00
Bill Currie
ec47e08e1c go back to using QFile, quakeio.h and quakefs.h 2002-08-27 07:16:28 +00:00
Bill Currie
36ca7b55bd fix several problems caused by insufficient error checking and some double
Qclose calls
2002-08-15 18:53:03 +00:00
Bill Currie
09ce2b0662 make sure a class' super class gets referenced when the class is implemented
to ensure the super class' object file gets linked in when it's in a library
2002-08-14 03:55:23 +00:00
Bill Currie
fabaf84745 fix a missing reloc ref for .this 2002-08-14 03:44:41 +00:00
Bill Currie
c2a7564d55 implement -l, but -L isn't supported yet and the full lib name must be given 2002-07-21 07:12:17 +00:00
Bill Currie
fb4ef47af6 make the filename output high verbosity 2002-07-17 19:00:16 +00:00
Bill Currie
8d6c491390 fix branches (been broken for a while :/) 2002-07-17 18:45:17 +00:00
Bill Currie
8a96b5bbbc oops, forgot to remove the debug print :) 2002-07-16 22:00:55 +00:00
Bill Currie
678a45cfe2 entity fields seem to work. I guess this means separate compilation is now
useful :)
2002-07-16 21:59:47 +00:00
Bill Currie
c36b7e36d2 fix external vectors and only complain about undefined externals that are
actually used
2002-07-16 20:09:14 +00:00
Bill Currie
ac1392eb74 object.r:
define __obj_exec_class, obj_msgSend and obj_msgSend_super for the way
	qfcc now handles these defs
pr_edict.c:
	PR_ValueString now checks that the function is valid before trying to
	print its name.
class.c:
	__obj_exec_class is now extern
linker.c:
	fix more mangled relocs
method.c:
	obj_msgSend and obj_msgSend_super are now extern
qfcc.c:
	fixes so obj_module_t doesn't get zapped
type.c:
	fixes so obj_module_t doesn't get zapped
2002-07-16 19:03:56 +00:00