3 * Added stabs debugging for the ELF output format, patch from
5 * Fix output/outbin.c to allow origin > 80000000h
10 * Update rdoff - librarian/archiver - common rec - docs!
11 * Fix signed/unsigned problems.
12 * Fix JMP FAR label and CALL FAR label.
13 * Add new multisection support - map files - fix align bug
14 * Fix sysexit, movhps/movlps reg,reg bugs in insns.dat
15 * "Q" or "O" suffixes indicate octal
16 * Support Prescott new instructions.
17 * Cyrix XSTORE instruction.
21 * Fix build failure on 16-bit DOS (Makefile.bc3 workaround for compiler bug.)
22 * Fix dependencies and compiler warnings.
23 * Add "const" in a number of places.
24 * Add -X option to specify error reporting format (use -Xvc to
25 integrate with Microsoft Visual Studio.)
26 * Minor changes for code legibility.
27 * Drop use of tmpnam() in rdoff (security fix.)
31 * Correct additional address-size vs. operand-size confusions.
32 * Generate dependencies for all Makefiles automatically.
33 * Add support for unimplemented (but theoretically available)
34 registers such as tr0 and cr5. Segment registers 6 and 7 are called
35 segr6 and segr7 for the operations which they can be represented.
36 * Correct some disassembler bugs related to redundant address-size prefixes.
37 Some work still remains in this area.
38 * Correctly generate an error for things like "SEG eax".
39 * Add the JMPE instruction, enabled by "CPU IA64".
40 * Correct compilation on newer gcc/glibc platforms.
41 * Issue an error on things like "jmp far eax".
47 * New __NASM_PATCHLEVEL__ and __NASM_VERSION_ID__ standard macros to
48 round out the version-query macros. version.pl now understands
49 X.YYplWW or X.YY.ZZplWW as a version number, equivalent to
50 X.YY.ZZ.WW (or X.YY.0.WW, as appropriate).
51 * New keyword "strict" to disable the optimization of specific
53 * Fix the handing of size overrides with JMP instructions
54 (instructions such as "jmp dword foo".)
55 * Fix the handling of "ABSOLUTE label", where "label" points into a
57 * Fix OBJ output format with lots of externs.
58 * More documentation updates.
59 * Add -Ov option to get verbose information about optimizations.
60 * Undo a braindead change which broke %elif directives.
67 * Fix NASM crashing when %macro directives were left unterminated.
68 * Lots of documentation updates.
69 * Complete rewrite of the PostScript/PDF documentation generator.
70 * The MS Visual C++ Makefile was updated and corrected.
71 * Recognize .rodata as a standard section name in ELF.
72 * Fix some obsolete Perl4-isms in Perl scripts.
73 * Fix configure.in to work with autoconf 2.5x.
74 * Fix a couple of "make cleaner" misses.
75 * Make the normal "./configure && make" work with Cygwin.
81 * Correctly build in a separate object directory again.
82 * Derive all references to the version number from the version file.
83 * New standard macros __NASM_SUBMINOR__ and __NASM_VER__ macros.
84 * Lots of Makefile updates and bug fixes.
85 * New %ifmacro directive to test for multiline macros.
86 * Documentation updates.
87 * Fixes for 16-bit OBJ format output.
88 * Changed the NASM environment variable to NASMENV.
94 * Changed doc files a lot: completely removed old READMExx and
95 Wishlist files, incorporating all information in CHANGES and TODO.
96 * I waited a long time to rename zoutieee.c to (original) outieee.c
97 * moved all output modules to output/ subdirectory.
98 * Added 'make strip' target to strip debug info from nasm & ndisasm.
99 * Added INSTALL file with installation instructions.
100 * Added -v option description to nasm man.
101 * Added dist makefile target to produce source distributions.
102 * 16-bit support for ELF output format (GNU extension, but useful.)
108 * Fastcooked this for Debian's Woody release:
109 Frank applied the INCBIN bug patch to 0.98.25alt and called
110 it 0.98.28 to not confuse poor little apt-get.
116 * Reorganised files even better from 0.98.25alt
122 * Prettified the source tree. Moved files to more reasonable places.
123 * Added findleak.pl script to misc/ directory.
124 * Attempted to fix doc.
129 * Line continuation character '\'
130 * Docs inadvertantly reverted - "dos packaging".
136 * FIXME: Someone, document this please.
142 * Documentation - Ndisasm doc added to Nasm.doc.
148 * Attempted to remove rdoff version1
149 * Lino Mastrodomenico's patches to preproc.c (%$$ bug?).
155 * Update rdoff2 - attempt to remove v1.
161 * Optimization fixes.
167 * Optimization fixes.
173 * H. J. Lu's patch back out.
179 * Added ".rdata" to "-f win32".
185 * H. J. Lu's "bogus elf" patch. (Red Hat problem?)
191 * Fix whitespace before "[section ..." bug.
197 * Fix fixes to memory leaks.
203 * (there was no '.13)
208 * Update optimization (new function of "-O1")
209 * Changes to test/bintest.asm (?).
214 * Optimization changes.
216 * (there was no '.10)
221 * Add multiple sections support to "-f bin".
222 * Changed GLOBAL_TEMP_BASE in outelf.c from 6 to 15.
223 * Add "-v" as an alias to the "-r" switch.
224 * Remove "#ifdef" from Tasm compatibility options.
225 * Remove redundant size-overrides on "mov ds, ex", etc.
226 * Fixes to SSE2, other insns.dat (?).
227 * Enable uppercase "I" and "P" switches.
228 * Case insinsitive "seg" and "wrt".
229 * Update install.sh (?).
230 * Allocate tokens in blocks.
231 * Improve "invalid effective address" messages.
236 * Add "%strlen" and "%substr" macro operators
237 * Fixed broken c16.mac.
238 * Unterminated string error reported.
239 * Fixed bugs as per 0.98bf
242 0.98.09b with John Coffman patches released 28-Oct-2001
243 -------------------------------------------------------
245 Changes from 0.98.07 release to 98.09b as of 28-Oct-2001
247 * More closely compatible with 0.98 when -O0 is implied
248 or specified. Not strictly identical, since backward
249 branches in range of short offsets are recognized, and signed
250 byte values with no explicit size specification will be
251 assembled as a single byte.
253 * More forgiving with the PUSH instruction. 0.98 requires
254 a size to be specified always. 0.98.09b will imply the size
255 from the current BITS setting (16 or 32).
257 * Changed definition of the optimization flag:
259 -O0 strict two-pass assembly, JMP and Jcc are
260 handled more like 0.98, except that back-
261 ward JMPs are short, if possible.
263 -O1 strict two-pass assembly, but forward
264 branches are assembled with code guaranteed
265 to reach; may produce larger code than
266 -O0, but will produce successful assembly
267 more often if branch offset sizes are not
270 -O2 multi-pass optimization, minimize branch
271 offsets; also will minimize signed immed-
272 iate bytes, overriding size specification.
274 -O3 like -O2, but more passes taken, if needed
277 0.98.07 released 01/28/01
278 -------------------------
280 * Added Stepane Denis' SSE2 instructions to a *working*
281 version of the code - some earlier versions were based on
282 broken code - sorry 'bout that. version "0.98.07"
288 * Cosmetic modifications to nasm.c, nasm.h,
292 0.98.06f released 01/18/01
293 --------------------------
295 * - Add "metalbrain"s jecxz bug fix in insns.dat
296 - alter nasmdoc.src to match - version "0.98.06f"
299 0.98.06e released 01/09/01
300 --------------------------
302 * Removed the "outforms.h" file - it appears to be
303 someone's old backup of "outform.h". version "0.98.06e"
309 * fbk - finally added the fix for the "multiple %includes bug",
310 known since 7/27/99 - reported originally (?) and sent to
311 us by Austin Lunnen - he reports that John Fine had a fix
312 within the day. Here it is...
314 * Nelson Rush resigns from the group. Big thanks to Nelson for
315 his leadership and enthusiasm in getting these changes
316 incorporated into Nasm!
318 * fbk - [list +], [list -] directives - ineptly implemented, should
319 be re-written or removed, perhaps.
321 * Brian Raiter / fbk - "elfso bug" fix - applied to aoutb format
322 as well - testing might be desirable...
328 * James Seter - -postfix, -prefix command line switches.
329 * Yuri Zaporogets - rdoff utility changes.
335 * GAS-like palign (Panos Minos)
336 * FIXME: Someone, fill this in with details
340 ----------------------------------
342 * Fixed - elf and aoutb bug - shared libraries
343 - multiple "%include" bug in "-f obj"
345 - unrecognized option bug in ndisasm
347 0.98.03 with John Coffman's changes released 27-Jul-2000
348 --------------------------------------------------------
350 * Added signed byte optimizations for the 0x81/0x83 class
351 of instructions: ADC, ADD, AND, CMP, OR, SBB, SUB, XOR:
352 when used as 'ADD reg16,imm' or 'ADD reg32,imm.' Also
353 optimization of signed byte form of 'PUSH imm' and 'IMUL
354 reg,imm'/'IMUL reg,reg,imm.' No size specification is needed.
356 * Added multi-pass JMP and Jcc offset optimization. Offsets
357 on forward references will preferentially use the short form,
358 without the need to code a specific size (short or near) for
359 the branch. Added instructions for 'Jcc label' to use the
360 form 'Jnotcc $+3/JMP label', in cases where a short offset
361 is out of bounds. If compiling for a 386 or higher CPU, then
362 the 386 form of Jcc will be used instead.
364 This feature is controlled by a new command-line switch: "O",
365 (upper case letter O). "-O0" reverts the assembler to no
366 extra optimization passes, "-O1" allows up to 5 extra passes,
367 and "-O2"(default), allows up to 10 extra optimization passes.
369 * Added a new directive: 'cpu XXX', where XXX is any of:
370 8086, 186, 286, 386, 486, 586, pentium, 686, PPro, P2, P3 or
371 Katmai. All are case insensitive. All instructions will
372 be selected only if they apply to the selected cpu or lower.
373 Corrected a couple of bugs in cpu-dependence in 'insns.dat'.
375 * Added to 'standard.mac', the "use16" and "use32" forms of
376 the "bits 16/32" directive. This is nothing new, just conforms
377 to a lot of other assemblers. (minor)
379 * Changed label allocation from 320/32 (10000 labels @ 200K+)
380 to 32/37 (1000 labels); makes running under DOS much easier.
381 Since additional label space is allocated dynamically, this
382 should have no effect on large programs with lots of labels.
383 The 37 is a prime, believed to be better for hashing. (minor)
385 * Integrated patchfile 0.98-0.98.01. I call this version
386 0.98.03, for historical reasons: 0.98.02 was trashed.
388 --John Coffman <johninsd@san.rr.com> 27-Jul-2000
391 Kendall Bennett's SciTech MGL changes
392 -------------------------------------
393 Note that you must define "TASM_COMPAT" at compile-time
394 to get the Tasm Ideal Mode compatibility.
396 All changes can be compiled in and out using the TASM_COMPAT macros,
397 and when compiled without TASM_COMPAT defined we get the exact same
398 binary as the unmodified 0.98 sources.
400 standard.mac, macros.c:
401 . Added macros to ignore TASM directives before first include
404 . Added extern declaration for tasm_compatible_mode
407 . Added global variable tasm_compatible_mode
408 . Added command line switch for TASM compatible mode (-t)
409 . Changed version command line to reflect when compiled with TASM additions
410 . Added response file processing to allow all arguments on a single
411 line (response file is @resp rather than -@resp for NASM format).
414 . Changes islocal() macro to support TASM style @@local labels.
415 . Added islocalchar() macro to support TASM style @@local labels.
418 . Added support for TASM style memory references (ie: mov [DWORD eax],10
419 rather than the NASM style mov DWORD [eax],10).
422 . Added new directives, %arg, %local, %stacksize to directives table
423 . Added support for TASM style directives without a leading % symbol.
425 Integrated a block of changes from Andrew Zabolotny <bit@eltech.ru>:
427 * A new keyword %xdefine and its case-insensitive counterpart %ixdefine.
428 They work almost the same way as %define and %idefine but expand
429 the definition immediately, not on the invocation. Something like a cross
430 between %define and %assign. The "x" suffix stands for "eXpand", so
431 "xdefine" can be deciphered as "expand-and-define". Thus you can do
437 %xdefine %1 dword [esp+ofs]
441 * Changed the place where the expansion of %$name macros are expanded.
442 Now they are converted into ..@ctxnum.name form when detokenizing, so
443 there are no quirks as before when using %$name arguments to macros,
444 in macros etc. For example:
453 Now last line will be expanded into "hello" as expected. This also allows
454 for lots of goodies, a good example are extended "proc" macros included
457 * Added a check for "cstk" in smacro_defined() before calling get_ctx() -
458 this allows for things like:
463 to work without warnings even in no context.
465 * Added a check for "cstk" in %if*ctx and %elif*ctx directives -
466 this allows to use %ifctx without excessive warnings. If there is
467 no active context, %ifctx goes through "false" branch.
469 * Removed "user error: " prefix with %error directive: it just clobbers the
470 output and has absolutely no functionality. Besides, this allows to write
471 macros that does not differ from built-in functions in any way.
473 * Added expansion of string that is output by %error directive. Now you
476 %define hello(x) Hello, x!
479 %error "hello(%$name)"
481 Same happened with %include directive.
483 * Now all directives that expect an identifier will try to expand and
484 concatenate everything without whitespaces in between before usage.
485 For example, with "unfixed" nasm the commands
488 %define __%$abc goodbye
491 would produce "incorrect" output: last line will expand to
495 Not quite what you expected, eh? :-) The answer is that preprocessor
496 treats the %define construct as if it would be
498 %define __ %$abc goodbye
500 (note the white space between __ and %$abc). After my "fix" it
501 will "correctly" expand into
505 as expected. Note that I use quotes around words "correct", "incorrect"
506 etc because this is rather a feature not a bug; however current behaviour
507 is more logical (and allows more advanced macro usage :-).
509 Same change was applied to:
510 %push,%macro,%imacro,%define,%idefine,%xdefine,%ixdefine,
511 %assign,%iassign,%undef
513 * A new directive [WARNING {+|-}warning-id] have been added. It works only
514 if the assembly phase is enabled (i.e. it doesn't work with nasm -e).
516 * A new warning type: macro-selfref. By default this warning is disabled;
517 when enabled NASM warns when a macro self-references itself; for example
518 the following source:
520 [WARNING macro-selfref]
531 will produce a warning, but if we remove the first line we won't see it
532 anymore (which is The Right Thing To Do {tm} IMHO since C preprocessor
533 eats such constructs without warnings at all).
535 * Added a "error" routine to preprocessor which always will set ERR_PASS1
536 bit in severity_code. This removes annoying repeated errors on first
537 and second passes from preprocessor.
539 * Added the %+ operator in single-line macros for concatenating two
540 identifiers. Usage example:
542 %define _myfunc _otherfunc
543 %define cextern(x) _ %+ x
546 After first expansion, third line will become "_myfunc". After this
547 expansion is performed again so it becomes "_otherunc".
549 * Now if preprocessor is in a non-emmitting state, no warning or error
550 will be emmitted. Example:
555 put anything you want between these two brackets,
556 even macro-parameter references %1 or local labels %$zz
557 or macro-local labels %%zz - no warning will be emmitted.
560 * Context-local variables on expansion as a last resort are looked up
561 in outer contexts. For example, the following piece:
571 will expand correctly the fourth line to [esp]; if we'll define another
572 %$a inside the "inner" context, it will take precedence over outer
573 definition. However, this modification has been applied only to
574 expand_smacro and not to smacro_define: as a consequence expansion
575 looks in outer contexts, but %ifdef won't look in outer contexts.
577 This behaviour is needed because we don't want nested contexts to
578 act on already defined local macros. Example:
580 %define %$arg1 [esp+4]
586 In this example the "if" mmacro enters into the "if" context, so %$arg1
587 is not valid anymore inside "if". Of course it could be worked around
588 by using explicitely %$$arg1 but this is ugly IMHO.
590 * Fixed memory leak in %undef. The origline wasn't freed before
593 * Fixed trap in preprocessor when line expanded to empty set of tokens.
594 This happens, for example, in the following case:
603 All changes since NASM 0.98p3 have been produced by H. Peter Anvin <hpa@zytor.com>.
605 * The documentation comment delimiter is \# not #.
606 * Allow EQU definitions to refer to external labels; reported by
608 * Re-enable support for RDOFF v1; reported by Pedro Gimeno.
609 * Updated License file per OK from Simon and Julian.
615 * Update documentation (although the instruction set reference will
616 have to wait; I don't want to hold up the 0.98 release for it.)
617 * Verified that the NASM implementation of the PEXTRW and PMOVMSKB
618 instructions is correct. The encoding differs from what the Intel
619 manuals document, but the Pentium III behaviour matches NASM, not
621 * Fix handling of implicit sizes in PSHUFW and PINSRW, reported by
623 * Resurrect the -s option, which was removed when changing the
624 diagnostic output to stdout.
630 * Fix for "DB" when NASM is running on a bigendian machine.
631 * Invoke insns.pl once for each output script, making Makefile.in
633 * Improve the Unix configure-based makefiles to make package
635 * Included an RPM .spec file for building RPM (RedHat Package Manager)
636 packages on Linux or Unix systems.
637 * Fix Makefile dependency problems.
638 * Change src/rdsrc.pl to include sectioning information in info
639 output; required for install-info to work.
640 * Updated the RDOFF distribution to version 2 from Jules; minor
641 massaging to make it compile in my environment.
642 * Split doc files that can be built by anyone with a Perl interpreter off
643 into a separate archive.
644 * "Dress rehearsal" release!
650 * Fixed opcodes with a third byte-sized immediate argument to not
651 complain if given "byte" on the immediate.
652 * Allow %undef to remove single-line macros with arguments. This
653 matches the behaviour of #undef in the C preprocessor.
654 * Allow -d, -u, -i and -p to be specified as -D, -U, -I and -P for
655 compatibility with most C compilers and preprocessors. This allows
656 Makefile options to be shared between cc and nasm, for example.
658 * Went through the list of Katmai instructions and hopefully fixed the
659 (rather few) mistakes in it.
660 * (Hopefully) fixed a number of disassembler bugs related to ambiguous
661 instructions (disambiguated by -p) and SSE instructions with REP.
662 * Fix for bug reported by Mark Junger: "call dword 0x12345678" should
663 work and may add an OSP (affected CALL, JMP, Jcc).
664 * Fix for environments when "stderr" isn't a compile-time constant.
670 * Took officially over coordination of the 0.98 release; so drop
671 the p3.x notation. Skipped p4 and p5 to avoid confusion with John
672 Fine's J4 and J5 releases.
673 * Update the documentation; however, it still doesn't include
674 documentation for the various new instructions. I somehow wonder if
675 it makes sense to have an instruction set reference in the assembler
676 manual when Intel et al have PDF versions of their manuals online.
677 * Recognize "idt" or "centaur" for the -p option to ndisasm.
678 * Changed error messages back to stderr where they belong, but add an
679 -E option to redirect them elsewhere (the DOS shell cannot redirect
681 * -M option to generate Makefile dependencies (based on code from Alex
683 * %undef preprocessor directive, and -u option, that undefines a
685 * OS/2 Makefile (Mkfiles/Makefile.os2) for Borland under OS/2; from
687 * Various minor bugfixes (reported by):
688 - Dangling %s in preproc.c (Martin Junker)
689 * THERE ARE KNOWN BUGS IN SSE AND THE OTHER KATMAI INSTRUCTIONS. I am
690 on a trip and didn't bring the Katmai instruction reference, so I
691 can't work on them right now.
692 * Updated the License file per agreement with Simon and Jules to
693 include a GPL distribution clause.
699 * (Hopefully) fixed the canned Makefiles to include the outrdf2 and
701 * Renamed changes.asm to changed.asm.
707 * Fixed a bunch of instructions that were added in 0.98p3.5 which had
708 memory operands, and the address-size prefix was missing from the
715 * Merged in changes from John S. Fine's 0.98-J5 release. John's based
716 0.98-J5 on my 0.98p3.3 release; this merges the changes.
717 * Expanded the instructions flag field to a long so we can fit more
718 flags; mark SSE (KNI) and AMD or Katmai-specific instructions as
720 * Fix the "PRIV" flag on a bunch of instructions, and create new
721 "PROT" flag for protected-mode-only instructions (orthogonal to if
722 the instruction is privileged!) and new "SMM" flag for SMM-only
724 * Added AMD-only SYSCALL and SYSRET instructions.
725 * Make SSE actually work, and add new Katmai MMX instructions.
726 * Added a -p (preferred vendor) option to ndisasm so that it can
727 distinguish e.g. Cyrix opcodes also used in SSE. For example:
729 ndisasm -p cyrix aliased.bin
730 00000000 670F514310 paddsiw mm0,[ebx+0x10]
731 00000005 670F514320 paddsiw mm0,[ebx+0x20]
732 ndisasm -p intel aliased.bin
733 00000000 670F514310 sqrtps xmm0,[ebx+0x10]
734 00000005 670F514320 sqrtps xmm0,[ebx+0x20]
735 * Added a bunch of Cyrix-specific instructions.
741 * Made at least an attempt to modify all the additional Makefiles (in
742 the Mkfiles directory). I can't test it, but this was the best I
744 * DOS DJGPP+"Opus Make" Makefile from John S. Fine.
745 * changes.asm changes from John S. Fine.
751 * Patch from Conan Brink to allow nesting of %rep directives.
752 * If we're going to allow INT01 as an alias for INT1/ICEBP (one of
753 Jules 0.98p3 changes), then we should allow INT03 as an alias for INT3
755 * Updated changes.asm to include the latest changes.
756 * Tried to clean up the <CR>s that had snuck in from a DOS/Windows
757 environment into my Unix environment, and try to make sure than
758 DOS/Windows users get them back.
759 * We would silently generate broken tools if insns.dat wasn't sorted
760 properly. Change insns.pl so that the order doesn't matter.
761 * Fix bug in insns.pl (introduced by me) which would cause conditional
762 instructions to have an extra "cc" in disassembly, e.g. "jnz"
763 disassembled as "jccnz".
769 * Merged in John S. Fine's changes from his 0.98-J4 prerelease; see
770 http://www.csoft.net/cz/johnfine/
771 * Changed previous "spotless" Makefile target (appropriate for distribution)
772 to "distclean", and added "cleaner" target which is same as "clean"
773 except deletes files generated by Perl scripts; "spotless" is union.
774 * Removed BASIC programs from distribution. Get a Perl interpreter
776 * Calling this "pre-release 3.2" rather than "p3-hpa2" because of
777 John's contributions.
778 * Actually link in the IEEE output format (zoutieee.c); fix a bunch of
779 compiler warnings in that file. Note I don't know what IEEE output
780 is supposed to look like, so these changes were made "blind".
786 * Merged nasm098p3.zip with nasm-0.97.tar.gz to create a fully
787 buildable version for Unix systems (Makefile.in updates, etc.)
788 * Changed insns.pl to create the instruction tables in nasm.h and
789 names.c, so that a new instruction can be added by adding it *only*
791 * Added the following new instructions: SYSENTER, SYSEXIT, FXSAVE,
792 FXRSTOR, UD1, UD2 (the latter two are two opcodes that Intel
793 guarantee will never be used; one of them is documented as UD2 in
794 Intel documentation, the other one just as "Undefined Opcode" --
795 calling it UD1 seemed to make sense.)
796 * MAX_SYMBOL was defined to be 9, but LOADALL286 and LOADALL386 are 10
797 characters long. Now MAX_SYMBOL is derived from insns.dat.
798 * A note on the BASIC programs included: forget them. insns.bas is
799 already out of date. Get yourself a Perl interpreter for your
800 platform of choice at:
802 http://www.cpan.org/ports/index.html
808 added response file support, improved command line handling, new layout
811 fixed limit checking bug, 'OUT byte nn, reg' bug, and a couple of rdoff
812 related bugs, updated Wishlist; 0.98 Prerelease 3.
818 fixed bug in outcoff.c to do with truncating section names longer
819 than 8 characters, referencing beyond end of string; 0.98 pre-release 2
822 0.98 pre-released May 1999
823 --------------------------
825 Fixed a bug whereby STRUC didn't work at all in RDF.
827 Fixed a problem with group specification in PUBDEFs in OBJ.
829 Improved ease of adding new output formats. Contribution due to
832 Fixed a bug in relocations in the `bin' format: was showing up when
833 a relocatable reference crossed an 8192-byte boundary in any output
836 Fixed a bug in local labels: local-label lookups were inconsistent
837 between passes one and two if an EQU occurred between the definition
838 of a global label and the subsequent use of a local label local to
841 Fixed a seg-fault in the preprocessor (again) which happened when
842 you use a blank line as the first line of a multi-line macro
843 definition and then defined a label on the same line as a call to
846 Fixed a stale-pointer bug in the handling of the NASM environment
847 variable. Thanks to Thomas McWilliams.
849 ELF had a hard limit on the number of sections which caused
850 segfaults when transgressed. Fixed.
852 Added ability for ndisasm to read from stdin by using `-' as the
855 ndisasm wasn't outputting the TO keyword. Fixed.
857 Fixed error cascade on bogus expression in %if - an error in
858 evaluation was causing the entire %if to be discarded, thus creating
859 trouble later when the %else or %endif was encountered.
861 Forward reference tracking was instruction-granular not operand-
862 granular, which was causing 286-specific code to be generated
863 needlessly on code of the form `shr word [forwardref],1'. Thanks to
864 Jim Hague for sending a patch.
866 All messages now appear on stdout, as sending them to stderr serves
867 no useful purpose other than to make redirection difficult.
869 Fixed the problem with EQUs pointing to an external symbol - this
870 now generates an error message.
872 Allowed multiple size prefixes to an operand, of which only the first
873 is taken into account.
875 Incorporated John Fine's changes, including fixes of a large number
876 of preprocessor bugs, some small problems in OBJ, and a reworking of
877 label handling to define labels before their line is assembled, rather
880 Reformatted a lot of the source code to be more readable. Included
881 'coding.txt' as a guideline for how to format code for contributors.
883 Stopped nested %reps causing a panic - they now cause a slightly more
884 friendly error message instead.
886 Fixed floating point constant problems (patch by Pedro Gimeno)
888 Fixed the return value of insn_size() not being checked for -1, indicating
891 Incorporated 3D now instructions.
893 Fixed the 'mov eax, eax + ebx' bug.
895 Fixed the GLOBAL EQU bug in ELF. Released developers release 3.
897 Incorporated John Fine's command line parsing changes
899 Incorporated David Lindauer's OMF debug support
901 Made changes for LCC 4.0 support (__NASM_CDecl__, removed register size
902 specification warning when sizes agree).
904 Released NASM 0.98 Pre-release 1
907 0.97 released December 1997
908 ---------------------------
910 This was entirely a bug-fix release to 0.96, which seems to have got
913 Fixed stupid mistake in OBJ which caused `MOV EAX,<constant>' to
914 fail. Caused by an error in the `MOV EAX,<segment>' support.
916 ndisasm hung at EOF when compiled with lcc on Linux because lcc on
917 Linux somehow breaks feof(). ndisasm now does not rely on feof().
919 A heading in the documentation was missing due to a markup error in
922 Fixed failure to update all pointers on realloc() within extended-
923 operand code in parser.c. Was causing wrong behaviour and seg faults
924 on lines such as `dd 0.0,0.0,0.0,0.0,...'
926 Fixed a subtle preprocessor bug whereby invoking one multi-line
927 macro on the first line of the expansion of another, when the second
928 had been invoked with a label defined before it, didn't expand the
931 Added internal.doc back in to the distribution archives - it was
932 missing in 0.96 *blush*
934 Fixed bug causing 0.96 to be unable to assemble its own test files,
935 specifically objtest.asm. *blush again*
937 Fixed seg-faults and bogus error messages caused by mismatching
938 %rep and %endrep within multi-line macro definitions.
940 Fixed a problem with buffer overrun in OBJ, which was causing
941 corruption at ends of long PUBDEF records.
943 Separated DOS archives into main-program and documentation to reduce
947 0.96 released November 1997
948 ---------------------------
950 Fixed a bug whereby, if `nasm sourcefile' would cause a filename
951 collision warning and put output into `nasm.out', then `nasm
952 sourcefile -o outputfile' still gave the warning even though the
955 Fixed name pollution under Digital UNIX: one of its header files
956 defined R_SP, which broke the enum in nasm.h.
958 Fixed minor instruction table problems: FUCOM and FUCOMP didn't have
959 two-operand forms; NDISASM didn't recognise the longer register
960 forms of PUSH and POP (eg FF F3 for PUSH BX); TEST mem,imm32 was
961 flagged as undocumented; the 32-bit forms of CMOV had 16-bit operand
962 size prefixes; `AAD imm' and `AAM imm' are no longer flagged as
963 undocumented because the Intel Architecture reference documents
966 Fixed a problem with the local-label mechanism, whereby strange
967 types of symbol (EQUs, auto-defined OBJ segment base symbols)
968 interfered with the `previous global label' value and screwed up
971 Fixed a bug whereby the stub preprocessor didn't communicate with
972 the listing file generator, so that the -a and -l options in
973 conjunction would produce a useless listing file.
975 Merged `os2' object file format back into `obj', after discovering
976 that `obj' _also_ shouldn't have a link pass separator in a module
977 containing a non-trivial MODEND. Flat segments are now declared
978 using the FLAT attribute. `os2' is no longer a valid object format
981 Removed the fixed-size temporary storage in the evaluator. Very very
982 long expressions (like `mov ax,1+1+1+1+...' for two hundred 1s or
983 so) should now no longer crash NASM.
985 Fixed a bug involving segfaults on disassembly of MMX instructions,
986 by changing the meaning of one of the operand-type flags in nasm.h.
987 This may cause other apparently unrelated MMX problems; it needs to
988 be tested thoroughly.
990 Fixed some buffer overrun problems with large OBJ output files.
991 Thanks to DJ Delorie for the bug report and fix.
993 Made preprocess-only mode actually listen to the %line markers as it
994 prints them, so that it can report errors more sanely.
996 Re-designed the evaluator to keep more sensible track of expressions
997 involving forward references: can now cope with previously-nightmare
1003 Added the ALIGN and ALIGNB standard macros.
1005 Added PIC support in ELF: use of WRT to obtain the four extra
1006 relocation types needed.
1008 Added the ability for output file formats to define their own
1009 extensions to the GLOBAL, COMMON and EXTERN directives.
1011 Implemented common-variable alignment, and global-symbol type and
1012 size declarations, in ELF.
1014 Implemented NEAR and FAR keywords for common variables, plus
1015 far-common element size specification, in OBJ.
1017 Added a feature whereby EXTERNs and COMMONs in OBJ can be given a
1018 default WRT specification (either a segment or a group).
1020 Transformed the Unix NASM archive into an auto-configuring package.
1022 Added a sanity-check for people applying SEG to things which are
1023 already segment bases: this previously went unnoticed by the SEG
1024 processing and caused OBJ-driver panics later.
1026 Added the ability, in OBJ format, to deal with `MOV EAX,<segment>'
1027 type references: OBJ doesn't directly support dword-size segment
1028 base fixups, but as long as the low two bytes of the constant term
1029 are zero, a word-size fixup can be generated instead and it will
1032 Added the ability to specify sections' alignment requirements in
1033 Win32 object files and pure binary files.
1035 Added preprocess-time expression evaluation: the %assign (and
1036 %iassign) directive and the bare %if (and %elif) conditional. Added
1037 relational operators to the evaluator, for use only in %if
1038 constructs: the standard relationals = < > <= >= <> (and C-like
1039 synonyms == and !=) plus low-precedence logical operators &&, ^^ and
1042 Added a preprocessor repeat construct: %rep / %exitrep / %endrep.
1044 Added the __FILE__ and __LINE__ standard macros.
1046 Added a sanity check for number constants being greater than
1047 0xFFFFFFFF. The warning can be disabled.
1049 Added the %0 token whereby a variadic multi-line macro can tell how
1050 many parameters it's been given in a specific invocation.
1052 Added %rotate, allowing multi-line macro parameters to be cycled.
1054 Added the `*' option for the maximum parameter count on multi-line
1055 macros, allowing them to take arbitrarily many parameters.
1057 Added the ability for the user-level forms of EXTERN, GLOBAL and
1058 COMMON to take more than one argument.
1060 Added the IMPORT and EXPORT directives in OBJ format, to deal with
1063 Added some more preprocessor %if constructs: %ifidn / %ifidni (exact
1064 textual identity), and %ifid / %ifnum / %ifstr (token type testing).
1066 Added the ability to distinguish SHL AX,1 (the 8086 version) from
1067 SHL AX,BYTE 1 (the 286-and-upwards version whose constant happens to
1070 Added NetBSD/FreeBSD/OpenBSD's variant of a.out format, complete
1071 with PIC shared library features.
1073 Changed NASM's idiosyncratic handling of FCLEX, FDISI, FENI, FINIT,
1074 FSAVE, FSTCW, FSTENV, and FSTSW to bring it into line with the
1075 otherwise accepted standard. The previous behaviour, though it was a
1076 deliberate feature, was a deliberate feature based on a
1077 misunderstanding. Apologies for the inconvenience.
1079 Improved the flexibility of ABSOLUTE: you can now give it an
1080 expression rather than being restricted to a constant, and it can
1081 take relocatable arguments as well.
1083 Added the ability for a variable to be declared as EXTERN multiple
1084 times, and the subsequent definitions are just ignored.
1086 We now allow instruction prefixes (CS, DS, LOCK, REPZ etc) to be
1087 alone on a line (without a following instruction).
1089 Improved sanity checks on whether the arguments to EXTERN, GLOBAL
1090 and COMMON are valid identifiers.
1092 Added misc/exebin.mac to allow direct generation of .EXE files by
1093 hacking up an EXE header using DB and DW; also added test/binexe.asm
1094 to demonstrate the use of this. Thanks to Yann Guidon for
1095 contributing the EXE header code.
1097 ndisasm forgot to check whether the input file had been successfully
1098 opened. Now it does. Doh!
1100 Added the Cyrix extensions to the MMX instruction set.
1102 Added a hinting mechanism to allow [EAX+EBX] and [EBX+EAX] to be
1103 assembled differently. This is important since [ESI+EBP] and
1104 [EBP+ESI] have different default base segment registers.
1106 Added support for the PharLap OMF extension for 4096-byte segment
1110 0.95 released July 1997
1111 -----------------------
1113 Fixed yet another ELF bug. This one manifested if the user relied on
1114 the default segment, and attempted to define global symbols without
1115 first explicitly declaring the target segment.
1117 Added makefiles (for NASM and the RDF tools) to build Win32 console
1118 apps under Symantec C++. Donated by Mark Junker.
1120 Added `macros.bas' and `insns.bas', QBasic versions of the Perl
1121 scripts that convert `standard.mac' to `macros.c' and convert
1122 `insns.dat' to `insnsa.c' and `insnsd.c'. Also thanks to Mark
1125 Changed the diassembled forms of the conditional instructions so
1126 that JB is now emitted as JC, and other similar changes. Suggested
1127 list by Ulrich Doewich.
1129 Added `@' to the list of valid characters to begin an identifier
1132 Documentary changes, notably the addition of the `Common Problems'
1133 section in nasm.doc.
1135 Fixed a bug relating to 32-bit PC-relative fixups in OBJ.
1137 Fixed a bug in perm_copy() in labels.c which was causing exceptions
1138 in cleanup_labels() on some systems.
1140 Positivity sanity check in TIMES argument changed from a warning to
1141 an error following a further complaint.
1143 Changed the acceptable limits on byte and word operands to allow
1144 things like `~10111001b' to work.
1146 Fixed a major problem in the preprocessor which caused seg-faults if
1147 macro definitions contained blank lines or comment-only lines.
1149 Fixed inadequate error checking on the commas separating the
1150 arguments to `db', `dw' etc.
1152 Fixed a crippling bug in the handling of macros with operand counts
1153 defined with a `+' modifier.
1155 Fixed a bug whereby object file formats which stored the input file
1156 name in the output file (such as OBJ and COFF) weren't doing so
1157 correctly when the output file name was specified on the command
1160 Removed [INC] and [INCLUDE] support for good, since they were
1163 Fixed a bug in OBJ which caused all fixups to be output in 16-bit
1164 (old-format) FIXUPP records, rather than putting the 32-bit ones in
1165 FIXUPP32 (new-format) records.
1167 Added, tentatively, OS/2 object file support (as a minor variant on
1170 Updates to Fox Cutter's Borland C makefile, Makefile.bc2.
1172 Removed a spurious second fclose() on the output file.
1174 Added the `-s' command line option to redirect all messages which
1175 would go to stderr (errors, help text) to stdout instead.
1177 Added the `-w' command line option to selectively suppress some
1178 classes of assembly warning messages.
1180 Added the `-p' pre-include and `-d' pre-define command-line options.
1182 Added an include file search path: the `-i' command line option.
1184 Fixed a silly little preprocessor bug whereby starting a line with a
1185 `%!' environment-variable reference caused an `unknown directive'
1188 Added the long-awaited listing file support: the `-l' command line
1191 Fixed a problem with OBJ format whereby, in the absence of any
1192 explicit segment definition, non-global symbols declared in the
1193 implicit default segment generated spurious EXTDEF records in the
1196 Added the NASM environment variable.
1198 From this version forward, Win32 console-mode binaries will be
1199 included in the DOS distribution in addition to the 16-bit binaries.
1200 Added Makefile.vc for this purpose.
1202 Added `return 0;' to test/objlink.c to prevent compiler warnings.
1204 Added the __NASM_MAJOR__ and __NASM_MINOR__ standard defines.
1206 Added an alternative memory-reference syntax in which prefixing an
1207 operand with `&' is equivalent to enclosing it in square brackets,
1208 at the request of Fox Cutter.
1210 Errors in pass two now cause the program to return a non-zero error
1211 code, which they didn't before.
1213 Fixed the single-line macro cycle detection, which didn't work at
1214 all on macros with no parameters (caused an infinite loop). Also
1215 changed the behaviour of single-line macro cycle detection to work
1216 like cpp, so that macros like `extrn' as given in the documentation
1219 Fixed the implementation of WRT, which was too restrictive in that
1220 you couldn't do `mov ax,[di+abc wrt dgroup]' because (di+abc) wasn't
1221 a relocatable reference.
1224 0.94 released April 1997
1225 ------------------------
1227 Major item: added the macro processor.
1229 Added undocumented instructions SMI, IBTS, XBTS and LOADALL286. Also
1230 reorganised CMPXCHG instruction into early-486 and Pentium forms.
1231 Thanks to Thobias Jones for the information.
1233 Fixed two more stupid bugs in ELF, which were causing `ld' to
1234 continue to seg-fault in a lot of non-trivial cases.
1236 Fixed a seg-fault in the label manager.
1238 Stopped FBLD and FBSTP from _requiring_ the TWORD keyword, which is
1239 the only option for BCD loads/stores in any case.
1241 Ensured FLDCW, FSTCW and FSTSW can cope with the WORD keyword, if
1242 anyone bothers to provide it. Previously they complained unless no
1243 keyword at all was present.
1245 Some forms of FDIV/FDIVR and FSUB/FSUBR were still inverted: a
1246 vestige of a bug that I thought had been fixed in 0.92. This was
1247 fixed, hopefully for good this time...
1249 Another minor phase error (insofar as a phase error can _ever_ be
1250 minor) fixed, this one occurring in code of the form
1251 rol ax,forward_reference
1252 forward_reference equ 1
1254 The number supplied to TIMES is now sanity-checked for positivity,
1255 and also may be greater than 64K (which previously didn't work on
1258 Added Watcom C makefiles, and misc/pmw.bat, donated by Dominik Behr.
1260 Added the INCBIN pseudo-opcode.
1262 Due to the advent of the preprocessor, the [INCLUDE] and [INC]
1263 directives have become obsolete. They are still supported in this
1264 version, with a warning, but won't be in the next.
1266 Fixed a bug in OBJ format, which caused incorrect object records to
1267 be output when absolute labels were made global.
1269 Updates to RDOFF subdirectory, and changes to outrdf.c.
1272 0.93 released January 1997
1273 --------------------------
1275 This release went out in a great hurry after semi-crippling bugs
1278 Really _did_ fix the stack overflows this time. *blush*
1280 Had problems with EA instruction sizes changing between passes, when
1281 an offset contained a forward reference and so 4 bytes were
1282 allocated for the offset in pass one; by pass two the symbol had
1283 been defined and happened to be a small absolute value, so only 1
1284 byte got allocated, causing instruction size mismatch between passes
1285 and hence incorrect address calculations. Fixed.
1287 Stupid bug in the revised ELF section generation fixed (associated
1288 string-table section for .symtab was hard-coded as 7, even when this
1289 didn't fit with the real section table). Was causing `ld' to
1290 seg-fault under Linux.
1292 Included a new Borland C makefile, Makefile.bc2, donated by Fox
1293 Cutter <lmb@comtch.iea.com>.
1296 0.92 released January 1997
1297 --------------------------
1299 The FDIVP/FDIVRP and FSUBP/FSUBRP pairs had been inverted: this was
1300 fixed. This also affected the LCC driver.
1302 Fixed a bug regarding 32-bit effective addresses of the form
1303 [other_register+ESP].
1305 Documentary changes, notably documentation of the fact that Borland
1306 Win32 compilers use `obj' rather than `win32' object format.
1308 Fixed the COMENT record in OBJ files, which was formatted
1311 Fixed a bug causing segfaults in large RDF files.
1313 OBJ format now strips initial periods from segment and group
1314 definitions, in order to avoid complications with the local label
1317 Fixed a bug in disassembling far calls and jumps in NDISASM.
1319 Added support for user-defined sections in COFF and ELF files.
1321 Compiled the DOS binaries with a sensible amount of stack, to
1322 prevent stack overflows on any arithmetic expression containing
1325 Fixed a bug in handling of files that do not terminate in a newline.
1328 0.91 released November 1996
1329 ---------------------------
1332 Support for RDF added.
1333 Support for DBG debugging format added.
1334 Support for 32-bit extensions to Microsoft OBJ format added.
1335 Revised for Borland C: some variable names changed, makefile added.
1336 LCC support revised to actually work.
1337 JMP/CALL NEAR/FAR notation added.
1338 `a16', `o16', `a32' and `o32' prefixes added.
1339 Range checking on short jumps implemented.
1340 MMX instruction support added.
1341 Negative floating point constant support added.
1342 Memory handling improved to bypass 64K barrier under DOS.
1343 $ prefix to force treatment of reserved words as identifiers added.
1344 Default-size mechanism for object formats added.
1345 Compile-time configurability added.
1346 `#', `@', `~' and `?' are now valid characters in labels.
1347 `-e' and `-k' options in NDISASM added.
1350 0.90 released October 1996
1351 --------------------------
1353 First release version. First support for object file output. Other
1354 changes from previous version (0.3x) too numerous to document.