1\#
2\# NASM revision history in nasmdoc format
3\#
4
5\H{cl-2.xx} NASM 2 Series
6
7The NASM 2 series supports x86-64, and is the production version of NASM
8since 2007.
9
10\S{cl-2.15.05} Version 2.15.05
11
12\b Correct \c{%ifid $} and \c{%ifid $$} being treated as true. See
13\k{iftyp}.
14
15\b Add \c{--reproducible} option to suppress NASM version numbers and
16timestamps in output files. See \k{opt-reproducible}.
17
18\S{cl-2.15.04} Version 2.15.04
19
20\b More sensible handling of the case where one single-line macro
21definition will shadow another. A warning will be issued, but the
22additional definition will be allowed. For the existing error case
23where both a parameterless and parametered macro are created, that
24warning is promoted to an error by default.
25
26\b Add special preprocessor tokens \c{%*?} and \c{%*??} that expand
27like \c{%?} and \c{%??} in single-line macros only. See
28\k{selfref%*?}.
29
30\b Correct the encoding of the \c{ENQCMDS} and \c{TILELOADT1}
31instructions.
32
33\b Fix case where the COFF backend (the \c{coff}, \c{win32} and
34\c{win64} output formats) would add padding bytes in the middle of a
35section if a \c{SECTION}/\c{SEGMENT} directive was provided which
36repeated an \c{ALIGN=} attribute. This neither matched legacy
37behavior, other backends, or user expectations.
38
39\b Fix SSE instructions not being recognized with an explicit memory
40operation size (e.g. \c{movsd qword [eax],xmm0}).
41
42\b The \c{-L+} option no longer enables \c{-Lw}, which is mainly
43useful to debug NASM crashes. See \k{opt-L}.
44
45\b Document long-standing hazards in the use of \c{$} in \c{Dx}
46statements, see \k{db}.
47
48\b The NASM-only RDOFF output format backend, which has been broken
49since at least NASM 2.14, has been disabled. The RDOFF tools are
50scheduled to be removed from the NASM distribution in NASM 2.16. If
51you have a concrete use case for RDOFF, please file a NASM bug report
52at \W{https://bugs.nasm.us/}\c{https://bugs.nasm.us/} as soon as
53possible. See \k{rdffmt}.
54
55\S{cl-2.15.03} Version 2.15.03
56
57\b Add instructions from the Intel Instruction Set Extensions and
58Future Features Programming Reference, June 2020. This includes
59AVX5512 \c{bfloat16}, AVX512 mask intersect, and Intel Advanced Matrix
60Extensions (AMX).
61
62\b Support for \c{bfloat16} floating-point constants. See \k{fltconst}
63and \k{pkg_fp}.
64
65\b Properly display warnings in preprocess-only mode.
66
67\b Fix copy-and-paste of examples from the PDF documentation.
68
69\b Debug information now properly reflect the line numbers of macro
70invocations (unless declared \c{.nolist}).
71
72\b Fix excessive alignment of sections in the
73\c{coff}/\c{win32}/\c{win64} output formats when the user-specified
74alignment is less than the default alignment for the section or
75section type.
76
77\b Fix explicit token pasting (\c{%+}, \k{concat%+}) for the cases
78where one or more parts result from empty token expansion, resulting
79in \c{%+} tokens at the beginning or end, or multiple ones in a row.
80
81\b Fix macro label capture (\c{%00}, \k{percent00}).
82
83\b Much better documentation for the MASM compatiblity package,
84\c{%use masm} (see \k{pkg_masm}).
85
86\b Fix \c{LEA} without square brackets, for MASM compatibility.
87
88\b Portability fixes.
89
90\S{cl-2.15.02} Version 2.15.02
91
92\b Fix miscompilation when building with \c{clang}.
93
94\b Add \c{db-empty} warning class, see \k{opt-w}.
95
96\b Fix the dependencies in the MSVC NMAKE makefile (\c{Mkfiles/msvc.mak}).
97
98\b Some documentation improvements and cleanups.
99
100\b Fix the handling of macro parameter ranges (\c{%\{:\}}), including
101with brace-enclosed original arguments.
102
103\S{cl-2.15.01} Version 2.15.01
104
105\b Fix building the documentation from the release archive. For 2.15,
106the user has to do \c{make warnings} manually in the main directory in
107order to be able to build the documentation, which means Perl needs to
108be installed on the system.
109
110\b Add instructions for Intel Control Flow Enforcement Technology (CET).
111
112\S{cl-2.15} Version 2.15
113
114\b The comparison and booleanizing operators can now be used in any
115expression context, not just \c{%if}. See \k{expr}.
116
117\b New operator \c{?} ... \c{:}.  See \k{exptri}.
118
119\b Signed shift operators \c{<<<} and \c{>>>}.  See \k{expshift}.
120
121\b The MASM \c{DUP} syntax for data definitions is now supported, in a
122somewhat enhanced form. See \k{db}.
123
124\b Warn for strange legacy behavior regarding empty arguments in
125multi-line macro expansion, but try to match legacy behavior in most
126cases. Legacy behavior can be disabled with the directive \c{%pragma
127preproc sane_empty_expansion}, see \k{mlmacro} and
128\k{pragma-preproc}.
129
130\b A much more sensible limit to expression evaluation depth. The
131previously defined limit would rarely trigger before NASM died with a
132stack overrun error on most systems. See \k{opt-limit}.
133
134\b The state of warnings can now be saved and restored via the
135\c{[WARNING PUSH]} and \c{[WARNING POP]} directives. See
136\k{asmdir-warning}.
137
138\b The \c{sectalign on|off} switch does not affect an explicit directive. See
139\k{sectalign}.
140
141\b Added \c{configure} option to enable building with profiling
142(\c{--enable-profiling}).
143
144\b Attempt to support of long path names, up to 32767 of UTF-16
145characters, on Windows.
146
147\b Fixed 'mismatch in operand sizes' error in the \c{MOVDDUP},
148\c{CMPXCHG8B} and \c{CMPXCHG16B} instructions.
149
150\b Improved error messages in the string transformation routine.
151
152\b Removed obsolete \c{gnu-elf-extensions} warning about 8- and 16-bit
153relocation generation. See \k{elf16}
154
155\b Added group aliases for all prefixed warnings. See \k{opt-w}.
156
157\b Allowed building with MSVC versions older than 1700.
158
159\b Added implicitly sized versions of the \c{K...} instructions, which
160allows the \c{K...} instructions to be specified without a size suffix as
161long as the operands are sized.
162
163\b Added \c{-L} option for additional listing information. See \k{opt-L}.
164
165\b Added some warnings for obsolete instructions for a specified CPU.
166
167\b Deprecated \c{-hf} and \c{-y} options. Use \c{-h} instead.
168
169\b Made DWARF as the default debug format for ELF.
170
171\b Added \c{%pragma list} \e{options...} to set or clear listing options
172(see \c{opt-L}).
173
174\b Allowed immediate syntax for \c{LEA} instruction (ignore operand
175size completely).
176
177\b Added limited functionality MASM compatibility package. See
178\k{pkg_masm}.
179
180\b Add single-line macros aliases using \c{%defalias} or
181\c{%idefalias}. These behave like a kind of "symbolic links" for
182single-line macros. See \k{defalias} and \c{clear}.
183
184\b Added support for stringify, nostrip, evaluating, and greedy
185single-line macro arguments. See \k{define}.
186
187\b Unused single-line macro arguments no longer need to have a
188specified name. See \k{define}.
189
190\b Added conditional comma operator \c{%,}. See \k{cond-comma}.
191
192\b Changed private namespace from \c{__foo__} to \c{__?foo?__}, so a user
193namespace starting from underscore is now clean from symbols. For
194backwards compatibility, the previous names are defined as aliases; see
195\k{defalias}, \k{clear} and \k{stdmac}.
196
197\b Added support of ELF weak symbols and external references. See \k{elfglob}.
198
199\b Changed the behavior of the EXTERN keyword and introduced REQUIRED keyword.
200See \k{required}.
201
202\b Added \c{%ifusable} and \c{%ifusing} directives. See \k{macropkg}.
203
204\b Made various performance improvements and stability fixes in macro
205preprocessor engine.
206
207\b Improved NASM error handling and cleaned up error messages.
208
209\b Many, many bug fixes.
210
211\S{cl-2.14.03} Version 2.14.03
212
213\b Suppress nuisance "\c{label changed during code generation}" messages
214after a real error.
215
216\b Add support for the \c{merge} and \c{strings} attributes on ELF
217sections. See \k{elfsect}.
218
219\b Add support for the \c{note}, \c{preinit_array}, \c{init_array},
220and \c{fini_array} sections type in ELF. See \k{elfsect}.
221
222\b Handle more than 32,633 sections in ELF.
223
224\S{cl-2.14.02} Version 2.14.02
225
226\b Fix crash due to multiple errors or warnings during the code
227generation pass if a list file is specified.
228
229\S{cl-2.14.01} Version 2.14.01
230
231\b Create all system-defined macros defore processing command-line
232given preprocessing directives (\c{-p}, \c{-d}, \c{-u}, \c{--pragma},
233\c{--before}).
234
235\b If debugging is enabled, define a \c{__DEBUG_FORMAT__} predefined
236macro.  See \k{dfmtm}.
237
238\b Fix an assert for the case in the \c{obj} format when a \c{SEG}
239operator refers to an \c{EXTERN} symbol declared further down in the
240code.
241
242\b Fix a corner case in the floating-point code where a binary, octal
243or hexadecimal floating-point having at least 32, 11, or 8 mantissa
244digits could produce slightly incorrect results under very specific
245conditions.
246
247\b Support \c{-MD} without a filename, for \c{gcc}
248compatibility. \c{-MF} can be used to set the dependencies output
249filename. See \k{opt-MD}.
250
251\b Fix \c{-E} in combination with \c{-MD}. See \k{opt-E}.
252
253\b Fix missing errors on redefined labels; would cause convergence
254failure instead which is very slow and not easy to debug.
255
256\b Duplicate definitions of the same label \e{with the same value} is now
257explicitly permitted (2.14 would allow it in some circumstances.)
258
259\b Add the option \c{--no-line} to ignore \c{%line} directives in the
260source. See \k{opt-no-line} and \k{line}.
261
262\S{cl-2.14} Version 2.14
263
264\b Changed \c{-I} option semantics by adding a trailing path separator
265   unconditionally.
266
267\b Fixed null dereference in corrupted invalid single line macros.
268
269\b Fixed division by zero which may happen if source code is malformed.
270
271\b Fixed out of bound access in processing of malformed segment override.
272
273\b Fixed out of bound access in certain \c{EQU} parsing.
274
275\b Fixed buffer underflow in float parsing.
276
277\b Added \c{SGX} (Intel Software Guard Extensions) instructions.
278
279\b Added \c{+n} syntax for multiple contiguous registers.
280
281\b Fixed \c{subsections_via_symbols} for \c{macho} object format.
282
283\b Added the \c{--gprefix}, \c{--gpostfix}, \c{--lprefix}, and
284\c{--lpostfix} command line options, to allow command line base symbol
285renaming. See \k{opt-pfix}.
286
287\b Allow label renaming to be specified by \c{%pragma} in addition to
288from the command line.  See \k{mangling}.
289
290\b Supported generic \c{%pragma} namespaces, \c{output} and \c{debug}. See
291\k{pragma}.
292
293\b Added the \c{--pragma} command line option to inject a \c{%pragma}
294directive.  See \k{opt-pragma}.
295
296\b Added the \c{--before} command line option to accept preprocess
297statement before input. See \k{opt-before}.
298
299\b Added \c{AVX512} \c{VBMI2} (Additional Bit Manipulation), \c{VNNI} (Vector
300Neural Network), \c{BITALG} (Bit Algorithm), and \c{GFNI} (Galois Field New
301Instruction) instructions.
302
303\b Added the \c{STATIC} directive for local symbols that should be
304renamed using global-symbol rules. See \k{static}.
305
306\b Allow a symbol to be defined as \c{EXTERN} and then later
307overridden as \c{GLOBAL} or \c{COMMON}. Furthermore, a symbol declared
308\c{EXTERN} and then defined will be treated as \c{GLOBAL}. See \k{extern}.
309
310\b The \c{GLOBAL} directive no longer is required to precede the
311definition of the symbol.
312
313\b Support \c{private_extern} as \c{macho} specific extension to the
314\c{GLOBAL} directive. See \k{macho-pext}.
315
316\b Updated \c{UD0} encoding to match with the specification
317
318\b Added the \c{--limit-X} command line option to set execution
319limits. See \k{opt-limit}.
320
321\b Updated the \c{Codeview} version number to be aligned with \c{MASM}.
322
323\b Added the \c{--keep-all} command line option to preserve output
324files. See \k{opt-keep-all}.
325
326\b Added the \c{--include} command line option, an alias to \c{-P} (\k{opt-p}).
327
328\b Added the \c{--help} command line option as an alias to \c{-h} (\k{syntax}).
329
330\b Added \c{-W}, \c{-D}, and \c{-Q} suffix aliases for \c{RET}
331   instructions so the operand sizes of these instructions can be
332   encoded without using \c{o16}, \c{o32} or \c{o64}.
333
334\S{cl-2.13.03} Version 2.13.03
335
336\b Added AVX and AVX512 \c{VAES*} and \c{VPCLMULQDQ} instructions.
337
338\b Fixed missing dwarf record in x32 ELF output format.
339
340\S{cl-2.13.02} Version 2.13.02
341
342\b Fix false positive in testing of numeric overflows.
343
344\b Fix generation of \c{PEXTRW} instruction.
345
346\b Fix \c{smartalign} package which could trigger an error during
347   optimization if the alignment code expanded too much due to
348   optimization of the previous code.
349
350\b Fix a case where negative value in \c{TIMES} directive causes
351   panic instead of an error.
352
353\b Always finalize \c{.debug_abbrev} section with a null in
354   \c{dwarf} output format.
355
356\b Support \c{debug} flag in section attributes for \c{macho}
357   output format.  See \k{machosect}.
358
359\b Support up to 16 characters in section names for \c{macho}
360   output format.
361
362\b Fix missing update of global \c{BITS} setting if \c{SECTION}
363   directive specified a bit size using output format-specific
364   extensions (e.g. \c{USE32} for the \c{obj} output format.)
365
366\b Fix the incorrect generation of VEX-encoded instruction when static
367   mode decorators are specified on scalar instructions, losing the
368   decorators as they require EVEX encoding.
369
370\b Option \c{-MW} to quote dependency outputs according to Watcom
371   Make conventions instead of POSIX Make conventions.  See \k{opt-MW}.
372
373\b The \c{obj} output format now contains embedded dependency file
374   information, unless disabled with \c{%pragma obj nodepend}.  See
375   \k{objdepend}.
376
377\b Fix generation of dependency lists.
378
379\b Fix a number of null pointer reference and memory allocation errors.
380
381\b Always generate symbol-relative relocations for the \c{macho64}
382   output format; at least some versions of the XCode/LLVM linker fails
383   for section-relative relocations.
384
385\S{cl-2.13.01} Version 2.13.01
386
387\b Fix incorrect output for some types of \c{FAR} or \c{SEG}
388   references in the \c{obj} output format, and possibly other 16-bit
389   output formats.
390
391\b Fix the address in the list file for an instruction containing a
392   \c{TIMES} directive.
393
394\b Fix error with \c{TIMES} used together with an instruction which
395   can vary in size, e.g. \c{JMP}.
396
397\b Fix breakage on some uses of the \c{DZ} pseudo-op.
398
399\S{cl-2.13} Version 2.13
400
401\b Support the official forms of the \c{UD0} and \c{UD1} instructions.
402
403\b Allow self-segment-relative expressions in immediates and
404   displacements, even when combined with an external or otherwise
405   out-of-segment special symbol, e.g.:
406
407\c      extern foo
408\c      mov eax,[foo - $ + ebx]               ; Now legal
409
410\b Handle a 64-bit origin in NDISASM.
411
412\b NASM can now generate sparse output files for relevant output
413   formats, if the underlying operating system supports them.
414
415\b The \c{macho} object format now supports the \c{subsections_via_symbols}
416   and \c{no_dead_strip} directives, see \k{macho-ssvs}.
417
418\b The \c{macho} object format now supports the \c{no_dead_strip},
419  \c{live_support} and \c{strip_static_syms} section flags, see
420  \k{machosect}.
421
422\b The \c{macho} object format now supports the \c{dwarf} debugging
423  format, as required by newer toolchains.
424
425\b All warnings can now be suppressed if desired; warnings not
426   otherwise part of any warning class are now considered its own
427   warning class called \c{other} (e.g. \c{-w-other}).  Furthermore,
428   warning-as-error can now be controlled on a per warning class
429   basis, using the syntax \c{-w+error=}\e{warning-class} and its
430   equivalent for all other warning control options.  See \k{opt-w}
431   for the command-line options and warning classes and
432   \k{asmdir-warning} for the \c{[WARNING]} directive.
433
434\b Fix a number of bugs related to AVX-512 decorators.
435
436\b Significant improvements to building NASM with Microsoft Visual
437   Studio via \c{Mkfiles/msvc.mak}.  It is now possible to build the
438   full Windows installer binary as long as the necessary
439   prerequisites are installed; see \c{Mkfiles/README}
440
441\b To build NASM with custom modifications (table changes) or from the
442   git tree now requires Perl 5.8 at the very minimum, quite possibly
443   a higher version (Perl 5.24.1 tested.)  There is no requirement to
444   have Perl on your system at all if all you want to do is build
445   unmodified NASM from source archives.
446
447\b Fix the \c{\{z\}} decorator on AVX-512 \c{VMOVDQ*} instructions.
448
449\b Add new warnings for certain dangerous constructs which never ought
450   to have been allowed.  In particular, the \c{RESB} family of
451   instructions should have been taking a critical expression all
452   along.
453
454\b Fix the EVEX (AVX-512) versions of the \c{VPBROADCAST}, \c{VPEXTR},
455   and \c{VPINSR} instructions.
456
457\b Support contracted forms of additional instructions.  As a general
458   rule, if an instruction has a non-destructive source immediately
459   after a destination register that isn't used as an input, NASM
460   supports omitting that source register, using the destination
461   register as that value.  This among other things makes it easier to
462   convert SSE code to the equivalent AVX code:
463
464\c      addps xmm1,xmm0                       ; SSE instruction
465\c      vaddps ymm1,ymm1,ymm0                 ; AVX official long form
466\c      vaddps ymm1,ymm0                      ; AVX contracted form
467
468\b Fix Codeview malformed compiler version record.
469
470\b Add the \c{CLWB} and \c{PCOMMIT} instructions.  Note that the
471   \c{PCOMMIT} instruction has been deprecated and will never be
472   included in a shipping product; it is included for completeness
473   only.
474
475\b Add the \c{%pragma} preprocessor directive for soft-error directives.
476
477\b Add the \c{RDPID} instruction.
478
479\S{cl-2.12.02} Version 2.12.02
480
481\b Fix preprocessor errors, especially \c{%error} and \c{%warning},
482   inside \c{%if} statements.
483
484\b Fix relative relocations in 32-bit Mach-O.
485
486\b More Codeview debug format fixes.
487
488\b If the MASM \c{PTR} keyword is encountered, issue a warning.  This is
489   much more likely to indicate a MASM-ism encountered in NASM than it
490   is a valid label.  This warning can be suppressed with \c{-w-ptr},
491   the \c{[warning]} directive (see \k{opt-w}) or by the macro
492   definition \c{%idefine ptr $%?} (see \k{selfref%?}).
493
494\b When an error or a warning comes from the expansion of a multi-line
495   macro, display the file and line numbers for the expanded macros.
496   Macros defined with \c{.nolist} do not get displayed.
497
498\b Add macros \c{ilog2fw()} and \c{ilog2cw()} to the \c{ifunc} macro
499   package.  See \k{ilog2}.
500
501
502\S{cl-2.12.01} Version 2.12.01
503
504\b Portability fixes for some platforms.
505
506\b Fix error when not specifying a list file.
507
508\b Correct the handling of macro-local labels in the Codeview
509   debugging format.
510
511\b Add \c{CLZERO}, \c{MONITORX} and \c{MWAITX} instructions.
512
513
514\S{cl-2.12} Version 2.12
515
516\b Major fixes to the \c{macho} backend (\k{machofmt}); earlier versions
517   would produce invalid symbols and relocations on a regular basis.
518
519\b Support for thread-local storage in Mach-O.
520
521\b Support for arbitrary sections in Mach-O.
522
523\b Fix wrong negative size treated as a big positive value passed into
524   backend causing NASM to crash.
525
526\b Fix handling of zero-extending unsigned relocations, we have been printing
527   wrong message and forgot to assign segment with predefined value before
528   passing it into output format.
529
530\b Fix potential write of oversized (with size greater than allowed in
531   output format) relative relocations.
532
533\b Portability fixes for building NASM with the LLVM compiler.
534
535\b Add support of Codeview version 8 (\c{cv8}) debug format for
536   \c{win32} and \c{win64} formats in the \c{COFF} backend,
537   see \k{codeview}.
538
539\b Allow 64-bit outputs in 16/32-bit only backends.  Unsigned 64-bit
540   relocations are zero-extended from 32-bits with a warning
541   (suppressible via \c{-w-zext-reloc}); signed 64-bit relocations are
542   an error.
543
544\b Line numbers in list files now correspond to the lines in the source
545   files, instead of simply being sequential.
546
547\b There is now an official 64-bit (x64 a.k.a. x86-64) build for Windows.
548
549
550\S{cl-2.11.09} Version 2.11.09
551
552\b Fix potential stack overwrite in \c{macho32} backend.
553
554\b Fix relocation records in \c{macho64} backend.
555
556\b Fix symbol lookup computation in \c{macho64} backend.
557
558\b Adjust \c{.symtab} and \c{.rela.text} sections alignments to 8 bytes
559   in \c{elf64} backed.
560
561\b Fix section length computation in \c{bin} backend which leaded in incorrect
562   relocation records.
563
564\S{cl-2.11.08} Version 2.11.08
565
566\b Fix section length computation in \c{bin} backend which leaded in incorrect
567   relocation records.
568
569\b Add a warning for numeric preprocessor definitions passed via command
570   line which might have unexpected results otherwise.
571
572\b Add ability to specify a module name record in \c{rdoff} linker with
573   \c{-mn} option.
574
575\b Increase label length capacity up to 256 bytes in \c{rdoff} backend for
576   FreePascal sake, which tends to generate very long labels for procedures.
577
578\b Fix segmentation failure when rip addressing is used in \c{macho64} backend.
579
580\b Fix access on out of memory when handling strings with a single
581   grave. We have sixed similar problem in previous release but not
582   all cases were covered.
583
584\b Fix NULL dereference in disassembled on \c{BND} instruction.
585
586\S{cl-2.11.07} Version 2.11.07
587
588\b Fix 256 bit \c{VMOVNTPS} instruction.
589
590\b Fix \c{-MD} option handling, which was rather broken in previous
591release changing command line api.
592
593\b Fix access to unitialized space when handling strings with
594a single grave.
595
596\b Fix nil dereference in handling memory reference parsing.
597
598\S{cl-2.11.06} Version 2.11.06
599
600\b Update AVX512 instructions based on the Extension Reference (319433-021 Sept
6012014).
602
603\b Fix the behavior of \c{-MF} and \c{-MD} options (Bugzilla 3392280)
604
605\b Updated Win32 Makefile to fix issue with build
606
607\S{cl-2.11.05} Version 2.11.05
608
609\b Add \c{--v} as an alias for \c{-v} (see \k{opt-v}), for
610command-line compatibility with Yasm.
611
612\b Fix a bug introduced in 2.11.03 whereby certain instructions would
613contain multiple REX prefixes, and thus be corrupt.
614
615\S{cl-2.11.04} Version 2.11.04
616
617\b Removed an invalid error checking code. Sometimes a memref only with
618a displacement can also set an evex flag. For example:
619
620\c       vmovdqu32 [0xabcd]{k1}, zmm0
621
622\b Fixed a bug in disassembler that EVEX.L'L vector length was not matched
623when EVEX.b was set because it was simply considered as EVEC.RC.
624Separated EVEX.L'L case from EVEX.RC which is ignored in matching.
625
626\S{cl-2.11.03} Version 2.11.03
627
628\b Fix a bug there REX prefixes were missing on instructions inside a
629\c{TIMES} statement.
630
631\S{cl-2.11.02} Version 2.11.02
632
633\b Add the \c{XSAVEC}, \c{XSAVES} and \c{XRSTORS} family instructions.
634
635\b Add the \c{CLFLUSHOPT} instruction.
636
637\S{cl-2.11.01} Version 2.11.01
638
639\b Allow instructions which implicitly uses \c{XMM0} (\c{VBLENDVPD},
640\c{VBLENDVPS}, \c{PBLENDVB} and \c{SHA256RNDS2}) to be specified
641without an explicit \c{xmm0} on the assembly line.  In other words,
642the following two lines produce the same output:
643
644\c      vblendvpd xmm2,xmm1,xmm0      ; Last operand is fixed xmm0
645\c      vblendvpd xmm2,xmm1           ; Implicit xmm0 omitted
646
647\b In the ELF backends, don't crash the assembler if \c{section align}
648is specified without a value.
649
650\S{cl-2.11} Version 2.11
651
652\b Add support for the Intel AVX-512 instruction set:
653
654\b 16 new, 512-bit SIMD registers. Total 32 \c{(ZMM0 ~ ZMM31)}
655
656\b 8 new opmask registers \c{(K0 ~ K7)}. One of 7 registers \c{(K1 ~ K7)} can
657be used as an opmask for conditional execution.
658
659\b A new EVEX encoding prefix. EVEX is based on VEX and provides more
660capabilities: opmasks, broadcasting, embedded rounding and compressed
661displacements.
662
663\c  - opmask
664\c      VDIVPD zmm0{k1}{z}, zmm1, zmm3  ; conditional vector operation
665\c                                      ; using opmask k1.
666\c                                      ; {z} is for zero-masking
667\c  - broadcasting
668\c      VDIVPS zmm4, zmm5, [rbx]{1to16} ; load single-precision float and
669\c                                      ; replicate it 16 times. 32 * 16 = 512
670\c  - embedded rounding
671\c      VCVTSI2SD xmm6, xmm7, {rz-sae}, rax ; round toward zero. note that it
672\c                                       ; is used as if a separate operand.
673\c                                       ; it comes after the last SIMD operand
674
675\b Add support for \c{ZWORD} (512 bits), \c{DZ} and \c{RESZ}.
676
677\b Add support for the MPX and SHA instruction sets.
678
679\b Better handling of section redefinition.
680
681\b Generate manpages when running \c{'make dist'}.
682
683\b Handle all token chains in mmacro params range.
684
685\b Support split [base,index] effective address:
686
687\c      mov eax,[eax+8,ecx*4]   ; eax=base, ecx=index, 4=scale, 8=disp
688
689This is expected to be most useful for the MPX instructions.
690
691\b Support \c{BND} prefix for branch instructions (for MPX).
692
693\b The \c{DEFAULT} directive can now take \c{BND} and \c{NOBND}
694options to indicate whether all relevant branches should be getting
695\c{BND} prefixes.  This is expected to be the normal for use in MPX
696code.
697
698\b Add \c{{evex}}, \c{{vex3}} and \c{{vex2}} instruction prefixes to
699have NASM encode the corresponding instruction, if possible, with an EVEX,
7003-byte VEX, or 2-byte VEX prefix, respectively.
701
702\b Support for section names longer than 8 bytes in Win32/Win64 COFF.
703
704\b The \c{NOSPLIT} directive by itself no longer forces a single
705register to become an index register, unless it has an explicit
706multiplier.
707
708\c      mov eax,[nosplit eax]       ; eax as base register
709\c      mov eax,[nosplit eax*1]     ; eax as index register
710
711\S{cl-2.10.09} Version 2.10.09
712
713\b Pregenerate man pages.
714
715\S{cl-2.10.08} Version 2.10.08
716
717\b Fix \c{VMOVNTDQA}, \c{MOVNTDQA} and \c{MOVLPD} instructions.
718
719\b Fix collision for \c{VGATHERQPS}, \c{VPGATHERQD} instructions.
720
721\b Fix \c{VPMOVSXBQ}, \c{VGATHERQPD}, \c{VSPLLW} instructions.
722
723\b Add a bunch of AMD TBM instructions.
724
725\b Fix potential stack overwrite in numbers conversion.
726
727\b Allow byte size in \c{PREFETCHTx} instructions.
728
729\b Make manual pages up to date.
730
731\b Make \c{F3} and \c{F2} SSE prefixes to override \c{66}.
732
733\b Support of AMD SVM instructions in 32 bit mode.
734
735\b Fix near offsets code generation for \c{JMP}, \c{CALL} instrictions
736in long mode.
737
738\b Fix preprocessor parse regression when id is expanding to a whitespace.
739
740\S{cl-2.10.07} Version 2.10.07
741
742\b Fix line continuation parsing being broken in previous version.
743
744\S{cl-2.10.06} Version 2.10.06
745
746\b Always quote the dependency source names when using the automatic
747dependency generation options.
748
749\b If no dependency target name is specified via the \c{-MT} or
750\c{-MQ} options, quote the default output name.
751
752\b Fix assembly of shift operations in \c{CPU 8086} mode.
753
754\b Fix incorrect generation of explicit immediate byte for shift by 1
755under certain circumstances.
756
757\b Fix assembly of the \c{VPCMPGTQ} instruction.
758
759\b Fix RIP-relative relocations in the \c{macho64} backend.
760
761\S{cl-2.10.05} Version 2.10.05
762
763\b Add the \c{CLAC} and \c{STAC} instructions.
764
765\S{cl-2.10.04} Version 2.10.04
766
767\b Add back the inadvertently deleted 256-bit version of the \c{VORPD}
768instruction.
769
770\b Correct disassembly of instructions starting with byte \c{82} hex.
771
772\b Fix corner cases in token pasting, for example:
773
774\c    %define N 1e%++%+ 5
775\c            dd N, 1e+5
776
777\S{cl-2.10.03} Version 2.10.03
778
779\b Correct the assembly of the instruction:
780
781\c XRELEASE MOV [absolute],AL
782
783\> Previous versions would incorrectly generate \c{F3 A2} for this
784instruction and issue a warning; correct behavior is to emit \c{F3 88
78505}.
786
787\S{cl-2.10.02} Version 2.10.02
788
789\b Add the \c{ifunc} macro package with integer functions, currently
790only integer logarithms.  See \k{pkg_ifunc}.
791
792\b Add the \c{RDSEED}, \c{ADCX} and \c{ADOX} instructions.
793
794\S{cl-2.10.01} Version 2.10.01
795
796\b Add missing VPMOVMSKB instruction with reg32, ymmreg operands.
797
798\S{cl-2.10} Version 2.10
799
800\b When optimization is enabled, \c{mov r64,imm} now optimizes to the
801  shortest form possible between:
802
803\c      mov r32,imm32                   ;  5 bytes
804\c      mov r64,imm32                   ;  7 bytes
805\c      mov r64,imm64                   ; 10 bytes
806
807\> To force a specific form, use the \c{STRICT} keyword, see \k{strict}.
808
809\b Add support for the Intel AVX2 instruction set.
810
811\b Add support for Bit Manipulation Instructions 1 and 2.
812
813\b Add support for Intel Transactional Synchronization Extensions (TSX).
814
815\b Add support for x32 ELF (32-bit ELF with the CPU in 64-bit mode.)
816   See \k{elffmt}.
817
818\b Add support for bigendian UTF-16 and UTF-32.  See \k{unicode}.
819
820\S{cl-2.09.10} Version 2.09.10
821
822\b Fix up NSIS script to protect uninstaller against registry keys
823   absence or corruption. It brings in a few additional questions
824   to a user during deinstallation procedure but still it is better
825   than unpredictable file removal.
826
827\S{cl-2.09.09} Version 2.09.09
828
829\b Fix initialization of section attributes of \c{bin} output format.
830
831\b Fix \c{mach64} output format bug that crashes NASM due to NULL symbols.
832
833
834\S{cl-2.09.08} Version 2.09.08
835
836\b Fix \c{__OUTPUT_FORMAT__} assignment when output driver alias
837   is used. For example when \c{-f elf} is used \c{__OUTPUT_FORMAT__}
838   must be set to \c{elf}, if \c{-f elf32} is used \c{__OUTPUT_FORMAT__}
839   must be assigned accordingly, i.e. to \c{elf32}. The rule applies to
840   all output driver aliases. See \k{ofmtm}.
841
842
843\S{cl-2.09.07} Version 2.09.07
844
845\b Fix attempts to close same file several times
846   when \c{-a} option is used.
847
848\b Fixes for VEXTRACTF128, VMASKMOVPS encoding.
849
850
851\S{cl-2.09.06} Version 2.09.06
852
853\b Fix missed section attribute initialization in \c{bin} output target.
854
855
856\S{cl-2.09.05} Version 2.09.05
857
858\b Fix arguments encoding for VPEXTRW instruction.
859
860\b Remove invalid form of VPEXTRW instruction.
861
862\b Add \c{VLDDQU} as alias for \c{VLDQQU} to
863   match specification.
864
865
866\S{cl-2.09.04} Version 2.09.04
867
868\b Fix incorrect labels offset for VEX intructions.
869
870\b Eliminate bogus warning on implicit operand size override.
871
872\b \c{%if} term could not handle 64 bit numbers.
873
874\b The COFF backend was limiting relocations number to 16 bits even if
875   in real there were a way more relocations.
876
877
878\S{cl-2.09.03} Version 2.09.03
879
880\b Print \c{%macro} name inside \c{%rep} blocks on error.
881
882\b Fix preprocessor expansion behaviour. It happened sometime
883   too early and sometime simply wrong. Move behaviour back to
884   the origins (down to NASM 2.05.01).
885
886\b Fix unitialized data dereference on OMF output format.
887
888\b Issue warning on unterminated \c{%{} construct.
889
890\b Fix for documentation typo.
891
892
893\S{cl-2.09.02} Version 2.09.02
894
895\b Fix reversed tokens when \c{%deftok} produces more than one output token.
896
897\b Fix segmentation fault on disassembling some VEX instructions.
898
899\b Missing \c{%endif} did not always cause error.
900
901\b Fix typo in documentation.
902
903\b Compound context local preprocessor single line macro identifiers
904  were not expanded early enough and as result lead to unresolved
905  symbols.
906
907
908\S{cl-2.09.01} Version 2.09.01
909
910\b Fix NULL dereference on missed %deftok second parameter.
911
912\b Fix NULL dereference on invalid %substr parameters.
913
914
915\S{cl-2.09} Version 2.09
916
917\b Fixed assignment the magnitude of \c{%rep} counter. It is limited
918  to 62 bits now.
919
920\b Fixed NULL dereference if argument of \c{%strlen} resolves
921  to whitespace. For example if nonexistent macro parameter is used.
922
923\b \c{%ifenv}, \c{%elifenv}, \c{%ifnenv}, and \c{%elifnenv} directives
924  introduced.  See \k{ifenv}.
925
926\b Fixed NULL dereference if environment variable is missed.
927
928\b Updates of new AVX v7 Intel instructions.
929
930\b \c{PUSH imm32} is now officially documented.
931
932\b Fix for encoding the LFS, LGS and LSS in 64-bit mode.
933
934\b Fixes for compatibility with OpenWatcom compiler and DOS 8.3 file
935  format limitation.
936
937\b Macros parameters range expansion introduced. See \k{mlmacrange}.
938
939\b Backward compatibility on expanging of local sigle macros restored.
940
941\b 8 bit relocations for \c{elf} and \c{bin} output formats are introduced.
942
943\b Short intersegment jumps are permitted now.
944
945\b An alignment more than 64 bytes are allowed for \c{win32},
946  \c{win64} output formats.
947
948\b \c{SECTALIGN} directive introduced. See \k{sectalign}.
949
950\b \c{nojmp} option introduced in \c{smartalign} package. See
951  \k{pkg_smartalign}.
952
953\b Short aliases \c{win}, \c{elf} and \c{macho} for output formats are
954  introduced.  Each stands for \c{win32}, \c{elf32} and \c{macho32}
955  accordingly.
956
957\b Faster handling of missing directives implemented.
958
959\b Various small improvements in documentation.
960
961\b No hang anymore if unable to open malloc.log file.
962
963\b The environments without vsnprintf function are able to build nasm again.
964
965\b AMD LWP instructions updated.
966
967\b Tighten EA checks. We warn a user if there overflow in EA addressing.
968
969\b Make \c{-Ox} the default optimization level.  For the legacy
970  behavior, specify \c{-O0} explicitly.  See \k{opt-O}.
971
972\b Environment variables read with \c{%!} or tested with \c{%ifenv}
973  can now contain non-identifier characters if surrounded by quotes.
974  See \k{getenv}.
975
976\b Add a new standard macro package \c{%use fp} for floating-point
977  convenience macros.  See \k{pkg_fp}.
978
979
980\S{cl-2.08.02} Version 2.08.02
981
982\b Fix crash under certain circumstances when using the \c{%+} operator.
983
984
985\S{cl-2.08.01} Version 2.08.01
986
987\b Fix the \c{%use} statement, which was broken in 2.08.
988
989
990\S{cl-2.08} Version 2.08
991
992\b A number of enhancements/fixes in macros area.
993
994\b Support for converting strings to tokens.  See \k{deftok}.
995
996\b Fuzzy operand size logic introduced.
997
998\b Fix COFF stack overrun on too long export identifiers.
999
1000\b Fix Macho-O alignment bug.
1001
1002\b Fix crashes with -fwin32 on file with many exports.
1003
1004\b Fix stack overrun for too long [DEBUG id].
1005
1006\b Fix incorrect sbyte usage in IMUL (hit only if optimization
1007  flag passed).
1008
1009\b Append ending token for \c{.stabs} records in the ELF output format.
1010
1011\b New NSIS script which uses ModernUI and MultiUser approach.
1012
1013\b Visual Studio 2008 NASM integration (rules file).
1014
1015\b Warn a user if a constant is too long (and as result will be stripped).
1016
1017\b The obsoleted pre-XOP AMD SSE5 instruction set which was never actualized
1018  was removed.
1019
1020\b Fix stack overrun on too long error file name passed from the command line.
1021
1022\b Bind symbols to the .text section by default (ie in case if SECTION
1023  directive was omitted) in the ELF output format.
1024
1025\b Fix sync points array index wrapping.
1026
1027\b A few fixes for FMA4 and XOP instruction templates.
1028
1029\b Add AMD Lightweight Profiling (LWP) instructions.
1030
1031\b Fix the offset for \c{%arg} in 64-bit mode.
1032
1033\b An undefined local macro (\c{%$}) no longer matches a global macro
1034  with the same name.
1035
1036\b Fix NULL dereference on too long local labels.
1037
1038
1039\S{cl-2.07} Version 2.07
1040
1041\b NASM is now under the 2-clause BSD license.  See \k{legal}.
1042
1043\b Fix the section type for the \c{.strtab} section in the \c{elf64}
1044  output format.
1045
1046\b Fix the handling of \c{COMMON} directives in the \c{obj} output format.
1047
1048\b New \c{ith} and \c{srec} output formats; these are variants of the
1049  \c{bin} output format which output Intel hex and Motorola S-records,
1050  respectively.  See \k{ithfmt} and \k{srecfmt}.
1051
1052\b \c{rdf2ihx} replaced with an enhanced \c{rdf2bin}, which can output
1053  binary, COM, Intel hex or Motorola S-records.
1054
1055\b The Windows installer now puts the NASM directory first in the
1056  \c{PATH} of the "NASM Shell".
1057
1058\b Revert the early expansion behavior of \c{%+} to pre-2.06 behavior:
1059  \c{%+} is only expanded late.
1060
1061\b Yet another Mach-O alignment fix.
1062
1063\b Don't delete the list file on errors.  Also, include error and
1064  warning information in the list file.
1065
1066\b Support for 64-bit Mach-O output, see \k{machofmt}.
1067
1068\b Fix assert failure on certain operations that involve strings with
1069  high-bit bytes.
1070
1071
1072\S{cl-2.06} Version 2.06
1073
1074\b This release is dedicated to the memory of Charles A. Crayne, long
1075  time NASM developer as well as moderator of \c{comp.lang.asm.x86} and
1076  author of the book \e{Serious Assembler}.  We miss you, Chuck.
1077
1078\b Support for indirect macro expansion (\c{%[...]}).  See \k{indmacro}.
1079
1080\b \c{%pop} can now take an argument, see \k{pushpop}.
1081
1082\b The argument to \c{%use} is no longer macro-expanded.  Use
1083  \c{%[...]} if macro expansion is desired.
1084
1085\b Support for thread-local storage in ELF32 and ELF64.  See \k{elftls}.
1086
1087\b Fix crash on \c{%ifmacro} without an argument.
1088
1089\b Correct the arguments to the \c{POPCNT} instruction.
1090
1091\b Fix section alignment in the Mach-O format.
1092
1093\b Update AVX support to version 5 of the Intel specification.
1094
1095\b Fix the handling of accesses to context-local macros from higher
1096  levels in the context stack.
1097
1098\b Treat \c{WAIT} as a prefix rather than as an instruction, thereby
1099  allowing constructs like \c{O16 FSAVE} to work correctly.
1100
1101\b Support for structures with a non-zero base offset. See \k{struc}.
1102
1103\b Correctly handle preprocessor token concatenation (see \k{concat})
1104   involving floating-point numbers.
1105
1106\b The \c{PINSR} series of instructions have been corrected and
1107   rationalized.
1108
1109\b Removed AMD SSE5, replaced with the new XOP/FMA4/CVT16 (rev 3.03)
1110   spec.
1111
1112\b The ELF backends no longer automatically generate a \c{.comment} section.
1113
1114\b Add additional "well-known" ELF sections with default attributes.  See
1115   \k{elfsect}.
1116
1117
1118\S{cl-2.05.01} Version 2.05.01
1119
1120\b Fix the \c{-w}/\c{-W} option parsing, which was broken in NASM 2.05.
1121
1122
1123\S{cl-2.05} Version 2.05
1124
1125\b Fix redundant REX.W prefix on \c{JMP reg64}.
1126
1127\b Make the behaviour of \c{-O0} match NASM 0.98 legacy behavior.
1128  See \k{opt-O}.
1129
1130\b \c{-w-user} can be used to suppress the output of \c{%warning} directives.
1131  See \k{opt-w}.
1132
1133\b Fix bug where \c{ALIGN} would issue a full alignment datum instead of
1134  zero bytes.
1135
1136\b Fix offsets in list files.
1137
1138\b Fix \c{%include} inside multi-line macros or loops.
1139
1140\b Fix error where NASM would generate a spurious warning on valid
1141  optimizations of immediate values.
1142
1143\b Fix arguments to a number of the \c{CVT} SSE instructions.
1144
1145\b Fix RIP-relative offsets when the instruction carries an immediate.
1146
1147\b Massive overhaul of the ELF64 backend for spec compliance.
1148
1149\b Fix the Geode \c{PFRCPV} and \c{PFRSQRTV} instruction.
1150
1151\b Fix the SSE 4.2 \c{CRC32} instruction.
1152
1153
1154\S{cl-2.04} Version 2.04
1155
1156\b Sanitize macro handing in the \c{%error} directive.
1157
1158\b New \c{%warning} directive to issue user-controlled warnings.
1159
1160\b \c{%error} directives are now deferred to the final assembly phase.
1161
1162\b New \c{%fatal} directive to immediately terminate assembly.
1163
1164\b New \c{%strcat} directive to join quoted strings together.
1165
1166\b New \c{%use} macro directive to support standard macro directives.  See
1167  \k{use}.
1168
1169\b Excess default parameters to \c{%macro} now issues a warning by default.
1170  See \k{mlmacro}.
1171
1172\b Fix \c{%ifn} and \c{%elifn}.
1173
1174\b Fix nested \c{%else} clauses.
1175
1176\b Correct the handling of nested \c{%rep}s.
1177
1178\b New \c{%unmacro} directive to undeclare a multi-line macro.
1179  See \k{unmacro}.
1180
1181\b Builtin macro \c{__PASS__} which expands to the current assembly pass.
1182  See \k{pass_macro}.
1183
1184\b \c{__utf16__} and \c{__utf32__} operators to generate UTF-16 and UTF-32
1185  strings.  See \k{unicode}.
1186
1187\b Fix bug in case-insensitive matching when compiled on platforms that
1188  don't use the \c{configure} script.  Of the official release binaries,
1189  that only affected the OS/2 binary.
1190
1191\b Support for x87 packed BCD constants.  See \k{bcdconst}.
1192
1193\b Correct the \c{LTR} and \c{SLDT} instructions in 64-bit mode.
1194
1195\b Fix unnecessary REX.W prefix on indirect jumps in 64-bit mode.
1196
1197\b Add AVX versions of the AES instructions (\c{VAES}...).
1198
1199\b Fix the 256-bit FMA instructions.
1200
1201\b Add 256-bit AVX stores per the latest AVX spec.
1202
1203\b VIA XCRYPT instructions can now be written either with or without
1204  \c{REP}, apparently different versions of the VIA spec wrote them
1205  differently.
1206
1207\b Add missing 64-bit \c{MOVNTI} instruction.
1208
1209\b Fix the operand size of \c{VMREAD} and \c{VMWRITE}.
1210
1211\b Numerous bug fixes, especially to the AES, AVX and VTX instructions.
1212
1213\b The optimizer now always runs until it converges.  It also runs even
1214  when disabled, but doesn't optimize.  This allows most forward references
1215  to be resolved properly.
1216
1217\b \c{%push} no longer needs a context identifier; omitting the context
1218  identifier results in an anonymous context.
1219
1220
1221\S{cl-2.03.01} Version 2.03.01
1222
1223\b Fix buffer overflow in the listing module.
1224
1225\b Fix the handling of hexadecimal escape codes in `...` strings.
1226
1227\b The Postscript/PDF documentation has been reformatted.
1228
1229\b The \c{-F} option now implies \c{-g}.
1230
1231
1232\S{cl-2.03} Version 2.03
1233
1234\b Add support for Intel AVX, CLMUL and FMA instructions,
1235including YMM registers.
1236
1237\b \c{dy}, \c{resy} and \c{yword} for 32-byte operands.
1238
1239\b Fix some SSE5 instructions.
1240
1241\b Intel \c{INVEPT}, \c{INVVPID} and \c{MOVBE} instructions.
1242
1243\b Fix checking for critical expressions when the optimizer is enabled.
1244
1245\b Support the DWARF debugging format for ELF targets.
1246
1247\b Fix optimizations of signed bytes.
1248
1249\b Fix operation on bigendian machines.
1250
1251\b Fix buffer overflow in the preprocessor.
1252
1253\b \c{SAFESEH} support for Win32, \c{IMAGEREL} for Win64 (SEH).
1254
1255\b \c{%?} and \c{%??} to refer to the name of a macro itself.  In particular,
1256\c{%idefine keyword $%?} can be used to make a keyword "disappear".
1257
1258\b New options for dependency generation: \c{-MD}, \c{-MF},
1259\c{-MP}, \c{-MT}, \c{-MQ}.
1260
1261\b New preprocessor directives \c{%pathsearch} and \c{%depend}; INCBIN
1262reimplemented as a macro.
1263
1264\b \c{%include} now resolves macros in a sane manner.
1265
1266\b \c{%substr} can now be used to get other than one-character substrings.
1267
1268\b New type of character/string constants, using backquotes (\c{`...`}),
1269which support C-style escape sequences.
1270
1271\b \c{%defstr} and \c{%idefstr} to stringize macro definitions before
1272creation.
1273
1274\b Fix forward references used in \c{EQU} statements.
1275
1276
1277\S{cl-2.02} Version 2.02
1278
1279\b Additional fixes for MMX operands with explicit \c{qword}, as well as
1280  (hopefully) SSE operands with \c{oword}.
1281
1282\b Fix handling of truncated strings with \c{DO}.
1283
1284\b Fix segfaults due to memory overwrites when floating-point constants
1285  were used.
1286
1287\b Fix segfaults due to missing include files.
1288
1289\b Fix OpenWatcom Makefiles for DOS and OS/2.
1290
1291\b Add autogenerated instruction list back into the documentation.
1292
1293\b ELF: Fix segfault when generating stabs, and no symbols have been
1294  defined.
1295
1296\b ELF: Experimental support for DWARF debugging information.
1297
1298\b New compile date and time standard macros.
1299
1300\b \c{%ifnum} now returns true for negative numbers.
1301
1302\b New \c{%iftoken} test for a single token.
1303
1304\b New \c{%ifempty} test for empty expansion.
1305
1306\b Add support for the \c{XSAVE} instruction group.
1307
1308\b Makefile for Netware/gcc.
1309
1310\b Fix issue with some warnings getting emitted way too many times.
1311
1312\b Autogenerated instruction list added to the documentation.
1313
1314
1315\S{cl-2.01} Version 2.01
1316
1317\b Fix the handling of MMX registers with explicit \c{qword} tags on
1318  memory (broken in 2.00 due to 64-bit changes.)
1319
1320\b Fix the PREFETCH instructions.
1321
1322\b Fix the documentation.
1323
1324\b Fix debugging info when using \c{-f elf}
1325(backwards compatibility alias for \c{-f elf32}).
1326
1327\b Man pages for rdoff tools (from the Debian project.)
1328
1329\b ELF: handle large numbers of sections.
1330
1331\b Fix corrupt output when the optimizer runs out of passes.
1332
1333
1334\S{cl-2.00} Version 2.00
1335
1336\b Added c99 data-type compliance.
1337
1338\b Added general x86-64 support.
1339
1340\b Added win64 (x86-64 COFF) output format.
1341
1342\b Added \c{__BITS__} standard macro.
1343
1344\b Renamed the \c{elf} output format to \c{elf32} for clarity.
1345
1346\b Added \c{elf64} and \c{macho} (MacOS X) output formats.
1347
1348\b Added Numeric constants in \c{dq} directive.
1349
1350\b Added \c{oword}, \c{do} and \c{reso} pseudo operands.
1351
1352\b Allow underscores in numbers.
1353
1354\b Added 8-, 16- and 128-bit floating-point formats.
1355
1356\b Added binary, octal and hexadecimal floating-point.
1357
1358\b Correct the generation of floating-point constants.
1359
1360\b Added floating-point option control.
1361
1362\b Added Infinity and NaN floating point support.
1363
1364\b Added ELF Symbol Visibility support.
1365
1366\b Added setting OSABI value in ELF header directive.
1367
1368\b Added Generate Makefile Dependencies option.
1369
1370\b Added Unlimited Optimization Passes option.
1371
1372\b Added \c{%IFN} and \c{%ELIFN} support.
1373
1374\b Added Logical Negation Operator.
1375
1376\b Enhanced Stack Relative Preprocessor Directives.
1377
1378\b Enhanced ELF Debug Formats.
1379
1380\b Enhanced Send Errors to a File option.
1381
1382\b Added SSSE3, SSE4.1, SSE4.2, SSE5 support.
1383
1384\b Added a large number of additional instructions.
1385
1386\b Significant performance improvements.
1387
1388\b \c{-w+warning} and \c{-w-warning} can now be written as -Wwarning and
1389 -Wno-warning, respectively.  See \k{opt-w}.
1390
1391\b Add \c{-w+error} to treat warnings as errors.  See \k{opt-w}.
1392
1393\b Add \c{-w+all} and \c{-w-all} to enable or disable all suppressible
1394 warnings.  See \k{opt-w}.
1395
1396
1397\H{cl-0.98.xx} NASM 0.98 Series
1398
1399The 0.98 series was the production versions of NASM from 1999 to 2007.
1400
1401
1402\S{cl-0.98.39} Version 0.98.39
1403
1404\b fix buffer overflow
1405
1406\b fix outas86's \c{.bss} handling
1407
1408\b "make spotless" no longer deletes config.h.in.
1409
1410\b \c{%(el)if(n)idn} insensitivity to string quotes difference (#809300).
1411
1412\b (nasm.c)\c{__OUTPUT_FORMAT__} changed to string value instead of symbol.
1413
1414\S{cl-0.98.38} Version 0.98.38
1415
1416
1417\b Add Makefile for 16-bit DOS binaries under OpenWatcom, and modify
1418  \c{mkdep.pl} to be able to generate completely pathless dependencies, as
1419  required by OpenWatcom wmake (it supports path searches, but not
1420  explicit paths.)
1421
1422\b Fix the \c{STR} instruction.
1423
1424\b Fix the ELF output format, which was broken under certain
1425  circumstances due to the addition of stabs support.
1426
1427\b Quick-fix Borland format debug-info for \c{-f obj}
1428
1429\b Fix for \c{%rep} with no arguments (#560568)
1430
1431\b Fix concatenation of preprocessor function call (#794686)
1432
1433\b Fix long label causes coredump (#677841)
1434
1435\b Use autoheader as well as autoconf to keep configure from generating
1436  ridiculously long command lines.
1437
1438\b Make sure that all of the formats which support debugging output
1439  actually will suppress debugging output when \c{-g} not specified.
1440
1441\S{cl-0.98.37} Version 0.98.37
1442
1443
1444\b Paths given in \c{-I} switch searched for \c{incbin}-ed as
1445  well as \c{%include}-ed files.
1446
1447\b Added stabs debugging for the ELF output format, patch from
1448  Martin Wawro.
1449
1450\b Fix \c{output/outbin.c} to allow origin > 80000000h.
1451
1452\b Make \c{-U} switch work.
1453
1454\b Fix the use of relative offsets with explicit prefixes, e.g.
1455\c{a32 loop foo}.
1456
1457\b Remove \c{backslash()}.
1458
1459\b Fix the \c{SMSW} and \c{SLDT} instructions.
1460
1461\b \c{-O2} and \c{-O3} are no longer aliases for \c{-O10} and \c{-O15}.
1462If you mean the latter, please say so! :)
1463
1464\S{cl-0.98.36} Version 0.98.36
1465
1466
1467\b Update rdoff - librarian/archiver - common rec - docs!
1468
1469\b Fix signed/unsigned problems.
1470
1471\b Fix \c{JMP FAR label} and \c{CALL FAR label}.
1472
1473\b Add new multisection support - map files - fix align bug
1474
1475\b Fix sysexit, movhps/movlps reg,reg bugs in insns.dat
1476
1477\b \c{Q} or \c{O} suffixes indicate octal
1478
1479\b Support Prescott new instructions (PNI).
1480
1481\b Cyrix \c{XSTORE} instruction.
1482
1483
1484\S{cl-0.98.35} Version 0.98.35
1485
1486\b Fix build failure on 16-bit DOS (Makefile.bc3 workaround for compiler bug.)
1487
1488\b Fix dependencies and compiler warnings.
1489
1490\b Add "const" in a number of places.
1491
1492\b Add -X option to specify error reporting format (use -Xvc to
1493  integrate with Microsoft Visual Studio.)
1494
1495\b Minor changes for code legibility.
1496
1497\b Drop use of tmpnam() in rdoff (security fix.)
1498
1499
1500\S{cl-0.98.34} Version 0.98.34
1501
1502\b Correct additional address-size vs. operand-size confusions.
1503
1504\b Generate dependencies for all Makefiles automatically.
1505
1506\b Add support for unimplemented (but theoretically available)
1507  registers such as tr0 and cr5.  Segment registers 6 and 7 are called
1508  segr6 and segr7 for the operations which they can be represented.
1509
1510\b Correct some disassembler bugs related to redundant address-size prefixes.
1511  Some work still remains in this area.
1512
1513\b Correctly generate an error for things like "SEG eax".
1514
1515\b Add the JMPE instruction, enabled by "CPU IA64".
1516
1517\b Correct compilation on newer gcc/glibc platforms.
1518
1519\b Issue an error on things like "jmp far eax".
1520
1521
1522\S{cl-0.98.33} Version 0.98.33
1523
1524\b New __NASM_PATCHLEVEL__ and __NASM_VERSION_ID__ standard macros to
1525  round out the version-query macros.  version.pl now understands
1526  X.YYplWW or X.YY.ZZplWW as a version number, equivalent to
1527  X.YY.ZZ.WW (or X.YY.0.WW, as appropriate).
1528
1529\b New keyword "strict" to disable the optimization of specific
1530  operands.
1531
1532\b Fix the handing of size overrides with JMP instructions
1533  (instructions such as "jmp dword foo".)
1534
1535\b Fix the handling of "ABSOLUTE label", where "label" points into a
1536  relocatable segment.
1537
1538\b Fix OBJ output format with lots of externs.
1539
1540\b More documentation updates.
1541
1542\b Add -Ov option to get verbose information about optimizations.
1543
1544\b Undo a braindead change which broke \c{%elif} directives.
1545
1546\b Makefile updates.
1547
1548
1549\S{cl-0.98.32} Version 0.98.32
1550
1551\b Fix NASM crashing when \c{%macro} directives were left unterminated.
1552
1553\b Lots of documentation updates.
1554
1555\b Complete rewrite of the PostScript/PDF documentation generator.
1556
1557\b The MS Visual C++ Makefile was updated and corrected.
1558
1559\b Recognize .rodata as a standard section name in ELF.
1560
1561\b Fix some obsolete Perl4-isms in Perl scripts.
1562
1563\b Fix configure.in to work with autoconf 2.5x.
1564
1565\b Fix a couple of "make cleaner" misses.
1566
1567\b Make the normal "./configure && make" work with Cygwin.
1568
1569
1570\S{cl-0.98.31} Version 0.98.31
1571
1572\b Correctly build in a separate object directory again.
1573
1574\b Derive all references to the version number from the version file.
1575
1576\b New standard macros __NASM_SUBMINOR__ and __NASM_VER__ macros.
1577
1578\b Lots of Makefile updates and bug fixes.
1579
1580\b New \c{%ifmacro} directive to test for multiline macros.
1581
1582\b Documentation updates.
1583
1584\b Fixes for 16-bit OBJ format output.
1585
1586\b Changed the NASM environment variable to NASMENV.
1587
1588
1589\S{cl-0.98.30} Version 0.98.30
1590
1591\b Changed doc files a lot: completely removed old READMExx and
1592  Wishlist files, incorporating all information in CHANGES and TODO.
1593
1594\b I waited a long time to rename zoutieee.c to (original) outieee.c
1595
1596\b moved all output modules to output/ subdirectory.
1597
1598\b Added 'make strip' target to strip debug info from nasm & ndisasm.
1599
1600\b Added INSTALL file with installation instructions.
1601
1602\b Added -v option description to nasm man.
1603
1604\b Added dist makefile target to produce source distributions.
1605
1606\b 16-bit support for ELF output format (GNU extension, but useful.)
1607
1608
1609\S{cl-0.98.28} Version 0.98.28
1610
1611\b Fastcooked this for Debian's Woody release:
1612Frank applied the INCBIN bug patch to 0.98.25alt and called
1613it 0.98.28 to not confuse poor little apt-get.
1614
1615
1616\S{cl-0.98.26} Version 0.98.26
1617
1618\b Reorganised files even better from 0.98.25alt
1619
1620
1621\S{cl-0.98.25alt} Version 0.98.25alt
1622
1623\b Prettified the source tree. Moved files to more reasonable places.
1624
1625\b Added findleak.pl script to misc/ directory.
1626
1627\b Attempted to fix doc.
1628
1629
1630\S{cl-0.98.25} Version 0.98.25
1631
1632\b Line continuation character \c{\\}.
1633
1634\b Docs inadvertantly reverted - "dos packaging".
1635
1636
1637\S{cl-0.98.24p1} Version 0.98.24p1
1638
1639\b FIXME: Someone, document this please.
1640
1641
1642\S{cl-0.98.24} Version 0.98.24
1643
1644\b Documentation - Ndisasm doc added to Nasm.doc.
1645
1646
1647\S{cl-0.98.23} Version 0.98.23
1648
1649\b Attempted to remove rdoff version1
1650
1651\b Lino Mastrodomenico's patches to preproc.c (%$$ bug?).
1652
1653
1654\S{cl-0.98.22} Version 0.98.22
1655
1656\b Update rdoff2 - attempt to remove v1.
1657
1658
1659\S{cl-0.98.21} Version 0.98.21
1660
1661\b Optimization fixes.
1662
1663
1664\S{cl-0.98.20} Version 0.98.20
1665
1666\b Optimization fixes.
1667
1668
1669\S{cl-0.98.19} Version 0.98.19
1670
1671\b H. J. Lu's patch back out.
1672
1673
1674\S{cl-0.98.18} Version 0.98.18
1675
1676\b Added ".rdata" to "-f win32".
1677
1678
1679\S{cl-0.98.17} Version 0.98.17
1680
1681\b H. J. Lu's "bogus elf" patch. (Red Hat problem?)
1682
1683
1684\S{cl-0.98.16} Version 0.98.16
1685
1686\b Fix whitespace before "[section ..." bug.
1687
1688
1689\S{cl-0.98.15} Version 0.98.15
1690
1691\b Rdoff changes (?).
1692
1693\b Fix fixes to memory leaks.
1694
1695
1696\S{cl-0.98.14} Version 0.98.14
1697
1698\b Fix memory leaks.
1699
1700
1701\S{cl-0.98.13} Version 0.98.13
1702
1703\b There was no 0.98.13
1704
1705
1706\S{cl-0.98.12} Version 0.98.12
1707
1708\b Update optimization (new function of "-O1")
1709
1710\b Changes to test/bintest.asm (?).
1711
1712
1713\S{cl-0.98.11} Version 0.98.11
1714
1715\b Optimization changes.
1716
1717\b Ndisasm fixed.
1718
1719
1720\S{cl-0.98.10} Version 0.98.10
1721
1722\b There was no 0.98.10
1723
1724
1725\S{cl-0.98.09} Version 0.98.09
1726
1727\b Add multiple sections support to "-f bin".
1728
1729\b Changed GLOBAL_TEMP_BASE in outelf.c from 6 to 15.
1730
1731\b Add "-v" as an alias to the "-r" switch.
1732
1733\b Remove "#ifdef" from Tasm compatibility options.
1734
1735\b Remove redundant size-overrides on "mov ds, ex", etc.
1736
1737\b Fixes to SSE2, other insns.dat (?).
1738
1739\b Enable uppercase "I" and "P" switches.
1740
1741\b Case insinsitive "seg" and "wrt".
1742
1743\b Update install.sh (?).
1744
1745\b Allocate tokens in blocks.
1746
1747\b Improve "invalid effective address" messages.
1748
1749
1750\S{cl-0.98.08} Version 0.98.08
1751
1752\b Add "\c{%strlen}" and "\c{%substr}" macro operators
1753
1754\b Fixed broken c16.mac.
1755
1756\b Unterminated string error reported.
1757
1758\b Fixed bugs as per 0.98bf
1759
1760
1761\S{cl-0.98.09b with John Coffman patches released 28-Oct-2001} Version 0.98.09b with John Coffman patches released 28-Oct-2001
1762
1763Changes from 0.98.07 release to 98.09b as of 28-Oct-2001
1764
1765\b More closely compatible with 0.98 when -O0 is implied
1766or specified.  Not strictly identical, since backward
1767branches in range of short offsets are recognized, and signed
1768byte values with no explicit size specification will be
1769assembled as a single byte.
1770
1771\b More forgiving with the PUSH instruction.  0.98 requires
1772a size to be specified always.  0.98.09b will imply the size
1773from the current BITS setting (16 or 32).
1774
1775\b Changed definition of the optimization flag:
1776
1777\c    -O0     strict two-pass assembly, JMP and Jcc are
1778\c            handled more like 0.98, except that back-
1779\c            ward JMPs are short, if possible.
1780\c
1781\c    -O1     strict two-pass assembly, but forward
1782\c            branches are assembled with code guaranteed
1783\c            to reach; may produce larger code than
1784\c            -O0, but will produce successful assembly
1785\c            more often if branch offset sizes are not
1786\c            specified.
1787\c
1788\c    -O2     multi-pass optimization, minimize branch
1789\c            offsets; also will minimize signed immed-
1790\c            iate bytes, overriding size specification.
1791\c
1792\c    -O3     like -O2, but more passes taken, if needed
1793
1794
1795\S{cl-0.98.07 released 01/28/01} Version 0.98.07 released 01/28/01
1796
1797\b      Added Stepane Denis' SSE2 instructions to a *working*
1798        version of the code - some earlier versions were based on
1799        broken code - sorry 'bout that. version "0.98.07"
1800
1801\b       Cosmetic modifications to nasm.c, nasm.h,
1802        AUTHORS, MODIFIED
1803
1804
1805\S{cl-0.98.06f released 01/18/01} Version 0.98.06f released 01/18/01
1806
1807
1808\b Add "metalbrain"s jecxz bug fix in insns.dat
1809
1810\b Alter nasmdoc.src to match - version "0.98.06f"
1811
1812
1813\S{cl-0.98.06e released 01/09/01} Version 0.98.06e released 01/09/01
1814
1815
1816\b       Removed the "outforms.h" file - it appears to be
1817        someone's old backup of "outform.h". version "0.98.06e"
1818
1819\b fbk - finally added the fix for the "multiple %includes bug",
1820        known since 7/27/99 - reported originally (?) and sent to
1821        us by Austin Lunnen - he reports that John Fine had a fix
1822        within the day. Here it is...
1823
1824\b Nelson Rush resigns from the group. Big thanks to Nelson for
1825  his leadership and enthusiasm in getting these changes
1826  incorporated into Nasm!
1827
1828\b fbk - [list +], [list -] directives - ineptly implemented, should
1829        be re-written or removed, perhaps.
1830
1831\b Brian Raiter / fbk - "elfso bug" fix - applied to aoutb format
1832                       as well - testing might be desirable...
1833
1834\b James Seter - -postfix, -prefix command line switches.
1835
1836\b Yuri Zaporozhets - rdoff utility changes.
1837
1838
1839\S{cl-0.98p1} Version 0.98p1
1840
1841\b GAS-like palign (Panos Minos)
1842
1843\b FIXME: Someone, fill this in with details
1844
1845
1846\S{cl-0.98bf (bug-fixed)} Version 0.98bf (bug-fixed)
1847
1848\b Fixed - elf and aoutb bug - shared libraries
1849        - multiple "%include" bug in "-f obj"
1850        - jcxz, jecxz bug
1851        - unrecognized option bug in ndisasm
1852
1853\S{cl-0.98.03 with John Coffman's changes released 27-Jul-2000} Version 0.98.03 with John Coffman's changes released 27-Jul-2000
1854
1855\b Added signed byte optimizations for the 0x81/0x83 class
1856of instructions: ADC, ADD, AND, CMP, OR, SBB, SUB, XOR:
1857when used as 'ADD reg16,imm' or 'ADD reg32,imm.'  Also
1858optimization of signed byte form of 'PUSH imm' and 'IMUL
1859reg,imm'/'IMUL reg,reg,imm.'  No size specification is needed.
1860
1861\b Added multi-pass JMP and Jcc offset optimization.  Offsets
1862on forward references will preferentially use the short form,
1863without the need to code a specific size (short or near) for
1864the branch.  Added instructions for 'Jcc label' to use the
1865form 'Jnotcc $+3/JMP label', in cases where a short offset
1866is out of bounds.  If compiling for a 386 or higher CPU, then
1867the 386 form of Jcc will be used instead.
1868
1869\> This feature is controlled by a new command-line switch: "O",
1870(upper case letter O).  "-O0" reverts the assembler to no
1871extra optimization passes, "-O1" allows up to 5 extra passes,
1872and "-O2"(default), allows up to 10 extra optimization passes.
1873
1874\b Added a new directive:  'cpu XXX', where XXX is any of:
18758086, 186, 286, 386, 486, 586, pentium, 686, PPro, P2, P3 or
1876Katmai.  All are case insensitive.  All instructions will
1877be selected only if they apply to the selected cpu or lower.
1878Corrected a couple of bugs in cpu-dependence in 'insns.dat'.
1879
1880\b Added to 'standard.mac', the "use16" and "use32" forms of
1881the "bits 16/32" directive. This is nothing new, just conforms
1882to a lot of other assemblers. (minor)
1883
1884\b Changed label allocation from 320/32 (10000 labels @ 200K+)
1885to 32/37 (1000 labels); makes running under DOS much easier.
1886Since additional label space is allocated dynamically, this
1887should have no effect on large programs with lots of labels.
1888The 37 is a prime, believed to be better for hashing. (minor)
1889
1890
1891\S{cl-0.98.03} Version 0.98.03
1892
1893"Integrated patchfile 0.98-0.98.01.  I call this version 0.98.03 for
1894historical reasons: 0.98.02 was trashed." --John Coffman
1895<johninsd@san.rr.com>, 27-Jul-2000
1896
1897\b Kendall Bennett's SciTech MGL changes
1898
1899\b Note that you must define "TASM_COMPAT" at compile-time
1900to get the Tasm Ideal Mode compatibility.
1901
1902\b All changes can be compiled in and out using the TASM_COMPAT macros,
1903and when compiled without TASM_COMPAT defined we get the exact same
1904binary as the unmodified 0.98 sources.
1905
1906\b standard.mac, macros.c: Added macros to ignore TASM directives before
1907first include
1908
1909\b nasm.h: Added extern declaration for tasm_compatible_mode
1910
1911\b nasm.c: Added global variable tasm_compatible_mode
1912
1913\b Added command line switch for TASM compatible mode (-t)
1914
1915\b Changed version command line to reflect when compiled with TASM additions
1916
1917\b Added response file processing to allow all arguments on a single
1918line (response file is @resp rather than -@resp for NASM format).
1919
1920\b labels.c: Changes islocal() macro to support TASM style @@local labels.
1921
1922\b Added islocalchar() macro to support TASM style @@local labels.
1923
1924\b parser.c: Added support for TASM style memory references (ie: mov
1925[DWORD eax],10 rather than the NASM style mov DWORD [eax],10).
1926
1927\b preproc.c: Added new directives, \c{%arg}, \c{%local}, \c{%stacksize} to directives
1928table
1929
1930\b Added support for TASM style directives without a leading % symbol.
1931
1932\b Integrated a block of changes from Andrew Zabolotny <bit@eltech.ru>:
1933
1934\b A new keyword \c{%xdefine} and its case-insensitive counterpart \c{%ixdefine}.
1935They work almost the same way as \c{%define} and \c{%idefine} but expand
1936the definition immediately, not on the invocation. Something like a cross
1937between \c{%define} and \c{%assign}. The "x" suffix stands for "eXpand", so
1938"xdefine" can be deciphered as "expand-and-define". Thus you can do
1939things like this:
1940
1941\c      %assign ofs     0
1942\c
1943\c      %macro  arg     1
1944\c              %xdefine %1 dword [esp+ofs]
1945\c              %assign ofs ofs+4
1946\c      %endmacro
1947
1948\b Changed the place where the expansion of %$name macros are expanded.
1949Now they are converted into ..@ctxnum.name form when detokenizing, so
1950there are no quirks as before when using %$name arguments to macros,
1951in macros etc. For example:
1952
1953\c      %macro  abc     1
1954\c              %define %1 hello
1955\c      %endm
1956\c
1957\c      abc     %$here
1958\c      %$here
1959
1960\>    Now last line will be expanded into "hello" as expected. This also allows
1961    for lots of goodies, a good example are extended "proc" macros included
1962    in this archive.
1963
1964\b Added a check for "cstk" in smacro_defined() before calling get_ctx() -
1965    this allows for things like:
1966
1967\c      %ifdef %$abc
1968\c      %endif
1969
1970\>    to work without warnings even in no context.
1971
1972\b Added a check for "cstk" in %if*ctx and %elif*ctx directives -
1973    this allows to use \c{%ifctx} without excessive warnings. If there is
1974    no active context, \c{%ifctx} goes through "false" branch.
1975
1976\b Removed "user error: " prefix with \c{%error} directive: it just clobbers the
1977    output and has absolutely no functionality. Besides, this allows to write
1978    macros that does not differ from built-in functions in any way.
1979
1980\b Added expansion of string that is output by \c{%error} directive. Now you
1981    can do things like:
1982
1983\c      %define hello(x) Hello, x!
1984\c
1985\c      %define %$name andy
1986\c      %error "hello(%$name)"
1987
1988\> Same happened with \c{%include} directive.
1989
1990\b Now all directives that expect an identifier will try to expand and
1991    concatenate everything without whitespaces in between before usage.
1992    For example, with "unfixed" nasm the commands
1993
1994\c      %define %$abc hello
1995\c      %define __%$abc goodbye
1996\c      __%$abc
1997
1998\>    would produce "incorrect" output: last line will expand to
1999
2000\c      hello goodbyehello
2001
2002\>    Not quite what you expected, eh? :-) The answer is that preprocessor
2003    treats the \c{%define} construct as if it would be
2004
2005\c      %define __ %$abc goodbye
2006
2007\>    (note the white space between __ and %$abc). After my "fix" it
2008    will "correctly" expand into
2009
2010\c      goodbye
2011
2012\>    as expected. Note that I use quotes around words "correct", "incorrect"
2013    etc because this is rather a feature not a bug; however current behaviour
2014    is more logical (and allows more advanced macro usage :-).
2015
2016    Same change was applied to:
2017        \c{%push},\c{%macro},\c{%imacro},\c{%define},\c{%idefine},\c{%xdefine},\c{%ixdefine},
2018        \c{%assign},\c{%iassign},\c{%undef}
2019
2020\b A new directive [WARNING {+|-}warning-id] have been added. It works only
2021    if the assembly phase is enabled (i.e. it doesn't work with nasm -e).
2022
2023\b A new warning type: macro-selfref. By default this warning is disabled;
2024    when enabled NASM warns when a macro self-references itself; for example
2025    the following source:
2026
2027\c        [WARNING macro-selfref]
2028\c
2029\c        %macro          push    1-*
2030\c                %rep    %0
2031\c                        push    %1
2032\c                        %rotate 1
2033\c                %endrep
2034\c        %endmacro
2035\c
2036\c                        push    eax,ebx,ecx
2037
2038\>  will produce a warning, but if we remove the first line we won't see it
2039    anymore (which is The Right Thing To Do {tm} IMHO since C preprocessor
2040    eats such constructs without warnings at all).
2041
2042\b Added a "error" routine to preprocessor which always will set ERR_PASS1
2043    bit in severity_code. This removes annoying repeated errors on first
2044    and second passes from preprocessor.
2045
2046\b Added the %+ operator in single-line macros for concatenating two
2047    identifiers. Usage example:
2048
2049\c        %define _myfunc _otherfunc
2050\c        %define cextern(x) _ %+ x
2051\c        cextern (myfunc)
2052
2053\>    After first expansion, third line will become "_myfunc". After this
2054    expansion is performed again so it becomes "_otherunc".
2055
2056\b Now if preprocessor is in a non-emitting state, no warning or error
2057    will be emitted. Example:
2058
2059\c        %if 1
2060\c                mov     eax,ebx
2061\c        %else
2062\c                put anything you want between these two brackets,
2063\c                even macro-parameter references %1 or local
2064\c                labels %$zz or macro-local labels %%zz - no
2065\c                warning will be emitted.
2066\c        %endif
2067
2068\b Context-local variables on expansion as a last resort are looked up
2069    in outer contexts. For example, the following piece:
2070
2071\c        %push   outer
2072\c        %define %$a [esp]
2073\c
2074\c                %push   inner
2075\c                %$a
2076\c                %pop
2077\c        %pop
2078
2079\>    will expand correctly the fourth line to [esp]; if we'll define another
2080    %$a inside the "inner" context, it will take precedence over outer
2081    definition. However, this modification has been applied only to
2082    expand_smacro and not to smacro_define: as a consequence expansion
2083    looks in outer contexts, but \c{%ifdef} won't look in outer contexts.
2084
2085\>    This behaviour is needed because we don't want nested contexts to
2086    act on already defined local macros. Example:
2087
2088\c        %define %$arg1  [esp+4]
2089\c        test    eax,eax
2090\c        if      nz
2091\c                mov     eax,%$arg1
2092\c        endif
2093
2094\>    In this example the "if" mmacro enters into the "if" context, so %$arg1
2095    is not valid anymore inside "if". Of course it could be worked around
2096    by using explicitely %$$arg1 but this is ugly IMHO.
2097
2098\b Fixed memory leak in \c{%undef}. The origline wasn't freed before
2099    exiting on success.
2100
2101\b Fixed trap in preprocessor when line expanded to empty set of tokens.
2102    This happens, for example, in the following case:
2103
2104\c        #define SOMETHING
2105\c        SOMETHING
2106
2107
2108\S{cl-0.98} Version 0.98
2109
2110All changes since NASM 0.98p3 have been produced by H. Peter Anvin <hpa@zytor.com>.
2111
2112\b The documentation comment delimiter is \# not #.
2113
2114\b Allow EQU definitions to refer to external labels; reported by
2115  Pedro Gimeno.
2116
2117\b Re-enable support for RDOFF v1; reported by Pedro Gimeno.
2118
2119\b Updated License file per OK from Simon and Julian.
2120
2121
2122\S{cl-0.98p9} Version 0.98p9
2123
2124\b Update documentation (although the instruction set reference will
2125  have to wait; I don't want to hold up the 0.98 release for it.)
2126
2127\b Verified that the NASM implementation of the PEXTRW and PMOVMSKB
2128  instructions is correct.  The encoding differs from what the Intel
2129  manuals document, but the Pentium III behaviour matches NASM, not
2130  the Intel manuals.
2131
2132\b Fix handling of implicit sizes in PSHUFW and PINSRW, reported by
2133  Stefan Hoffmeister.
2134
2135\b Resurrect the -s option, which was removed when changing the
2136  diagnostic output to stdout.
2137
2138
2139\S{cl-0.98p8} Version 0.98p8
2140
2141\b Fix for "DB" when NASM is running on a bigendian machine.
2142
2143\b Invoke insns.pl once for each output script, making Makefile.in
2144  legal for "make -j".
2145
2146\b Improve the Unix configure-based makefiles to make package
2147  creation easier.
2148
2149\b Included an RPM .spec file for building RPM (RedHat Package Manager)
2150  packages on Linux or Unix systems.
2151
2152\b Fix Makefile dependency problems.
2153
2154\b Change src/rdsrc.pl to include sectioning information in info
2155  output; required for install-info to work.
2156
2157\b Updated the RDOFF distribution to version 2 from Jules; minor
2158  massaging to make it compile in my environment.
2159
2160\b Split doc files that can be built by anyone with a Perl interpreter off
2161  into a separate archive.
2162
2163\b "Dress rehearsal" release!
2164
2165
2166\S{cl-0.98p7} Version 0.98p7
2167
2168\b Fixed opcodes with a third byte-sized immediate argument to not
2169  complain if given "byte" on the immediate.
2170
2171\b Allow \c{%undef} to remove single-line macros with arguments.  This
2172  matches the behaviour of #undef in the C preprocessor.
2173
2174\b Allow -d, -u, -i and -p to be specified as -D, -U, -I and -P for
2175  compatibility with most C compilers and preprocessors.  This allows
2176  Makefile options to be shared between cc and nasm, for example.
2177
2178\b Minor cleanups.
2179
2180\b Went through the list of Katmai instructions and hopefully fixed the
2181  (rather few) mistakes in it.
2182
2183\b (Hopefully) fixed a number of disassembler bugs related to ambiguous
2184  instructions (disambiguated by -p) and SSE instructions with REP.
2185
2186\b Fix for bug reported by Mark Junger: "call dword 0x12345678" should
2187  work and may add an OSP (affected CALL, JMP, Jcc).
2188
2189\b Fix for environments when "stderr" isn't a compile-time constant.
2190
2191
2192\S{cl-0.98p6} Version 0.98p6
2193
2194
2195\b Took officially over coordination of the 0.98 release; so drop
2196  the p3.x notation. Skipped p4 and p5 to avoid confusion with John
2197  Fine's J4 and J5 releases.
2198
2199\b Update the documentation; however, it still doesn't include
2200  documentation for the various new instructions.  I somehow wonder if
2201  it makes sense to have an instruction set reference in the assembler
2202  manual when Intel et al have PDF versions of their manuals online.
2203
2204\b Recognize "idt" or "centaur" for the -p option to ndisasm.
2205
2206\b Changed error messages back to stderr where they belong, but add an
2207  -E option to redirect them elsewhere (the DOS shell cannot redirect
2208  stderr.)
2209
2210\b -M option to generate Makefile dependencies (based on code from Alex
2211  Verstak.)
2212
2213\b \c{%undef} preprocessor directive, and -u option, that undefines a
2214  single-line macro.
2215
2216\b OS/2 Makefile (Mkfiles/Makefile.os2) for Borland under OS/2; from
2217  Chuck Crayne.
2218
2219\b Various minor bugfixes (reported by):
2220  - Dangling \c{%s} in preproc.c (Martin Junker)
2221
2222\b THERE ARE KNOWN BUGS IN SSE AND THE OTHER KATMAI INSTRUCTIONS.  I am
2223  on a trip and didn't bring the Katmai instruction reference, so I
2224  can't work on them right now.
2225
2226\b Updated the License file per agreement with Simon and Jules to
2227  include a GPL distribution clause.
2228
2229
2230\S{cl-0.98p3.7} Version 0.98p3.7
2231
2232\b (Hopefully) fixed the canned Makefiles to include the outrdf2 and
2233  zoutieee modules.
2234
2235\b Renamed changes.asm to changed.asm.
2236
2237
2238\S{cl-0.98p3.6} Version 0.98p3.6
2239
2240\b Fixed a bunch of instructions that were added in 0.98p3.5 which had
2241  memory operands, and the address-size prefix was missing from the
2242  instruction pattern.
2243
2244
2245\S{cl-0.98p3.5} Version 0.98p3.5
2246
2247\b Merged in changes from John S. Fine's 0.98-J5 release.  John's based
2248  0.98-J5 on my 0.98p3.3 release; this merges the changes.
2249
2250\b Expanded the instructions flag field to a long so we can fit more
2251  flags; mark SSE (KNI) and AMD or Katmai-specific instructions as
2252  such.
2253
2254\b Fix the "PRIV" flag on a bunch of instructions, and create new
2255  "PROT" flag for protected-mode-only instructions (orthogonal to if
2256  the instruction is privileged!) and new "SMM" flag for SMM-only
2257  instructions.
2258
2259\b Added AMD-only SYSCALL and SYSRET instructions.
2260
2261\b Make SSE actually work, and add new Katmai MMX instructions.
2262
2263\b Added a -p (preferred vendor) option to ndisasm so that it can
2264  distinguish e.g. Cyrix opcodes also used in SSE.  For example:
2265
2266\c      ndisasm -p cyrix aliased.bin
2267\c      00000000  670F514310        paddsiw mm0,[ebx+0x10]
2268\c      00000005  670F514320        paddsiw mm0,[ebx+0x20]
2269\c      ndisasm -p intel aliased.bin
2270\c      00000000  670F514310        sqrtps xmm0,[ebx+0x10]
2271\c      00000005  670F514320        sqrtps xmm0,[ebx+0x20]
2272
2273\b Added a bunch of Cyrix-specific instructions.
2274
2275
2276\S{cl-0.98p3.4} Version 0.98p3.4
2277
2278\b Made at least an attempt to modify all the additional Makefiles (in
2279  the Mkfiles directory).  I can't test it, but this was the best I
2280  could do.
2281
2282\b DOS DJGPP+"Opus Make" Makefile from John S. Fine.
2283
2284\b changes.asm changes from John S. Fine.
2285
2286
2287\S{cl-0.98p3.3} Version 0.98p3.3
2288
2289\b Patch from Conan Brink to allow nesting of \c{%rep} directives.
2290
2291\b If we're going to allow INT01 as an alias for INT1/ICEBP (one of
2292  Jules 0.98p3 changes), then we should allow INT03 as an alias for INT3
2293  as well.
2294
2295\b Updated changes.asm to include the latest changes.
2296
2297\b Tried to clean up the <CR>s that had snuck in from a DOS/Windows
2298  environment into my Unix environment, and try to make sure than
2299  DOS/Windows users get them back.
2300
2301\b We would silently generate broken tools if insns.dat wasn't sorted
2302  properly.  Change insns.pl so that the order doesn't matter.
2303
2304\b Fix bug in insns.pl (introduced by me) which would cause conditional
2305  instructions to have an extra "cc" in disassembly, e.g. "jnz"
2306  disassembled as "jccnz".
2307
2308
2309\S{cl-0.98p3.2} Version 0.98p3.2
2310
2311\b Merged in John S. Fine's changes from his 0.98-J4 prerelease; see
2312  http://www.csoft.net/cz/johnfine/
2313
2314\b Changed previous "spotless" Makefile target (appropriate for distribution)
2315  to "distclean", and added "cleaner" target which is same as "clean"
2316  except deletes files generated by Perl scripts; "spotless" is union.
2317
2318\b Removed BASIC programs from distribution.  Get a Perl interpreter
2319  instead (see below.)
2320
2321\b Calling this "pre-release 3.2" rather than "p3-hpa2" because of
2322  John's contributions.
2323
2324\b Actually link in the IEEE output format (zoutieee.c); fix a bunch of
2325  compiler warnings in that file.  Note I don't know what IEEE output
2326  is supposed to look like, so these changes were made "blind".
2327
2328
2329\S{cl-0.98p3-hpa} Version 0.98p3-hpa
2330
2331\b Merged nasm098p3.zip with nasm-0.97.tar.gz to create a fully
2332  buildable version for Unix systems (Makefile.in updates, etc.)
2333
2334\b Changed insns.pl to create the instruction tables in nasm.h and
2335  names.c, so that a new instruction can be added by adding it *only*
2336  to insns.dat.
2337
2338\b Added the following new instructions: SYSENTER, SYSEXIT, FXSAVE,
2339  FXRSTOR, UD1, UD2 (the latter two are two opcodes that Intel
2340  guarantee will never be used; one of them is documented as UD2 in
2341  Intel documentation, the other one just as "Undefined Opcode" --
2342  calling it UD1 seemed to make sense.)
2343
2344\b MAX_SYMBOL was defined to be 9, but LOADALL286 and LOADALL386 are 10
2345  characters long.  Now MAX_SYMBOL is derived from insns.dat.
2346
2347\b A note on the BASIC programs included: forget them.  insns.bas is
2348  already out of date.  Get yourself a Perl interpreter for your
2349  platform of choice at
2350  \W{http://www.cpan.org/ports/index.html}{http://www.cpan.org/ports/index.html}.
2351
2352
2353\S{cl-0.98p3} Version 0.98 pre-release 3
2354
2355\b added response file support, improved command line handling, new layout
2356help screen
2357
2358\b fixed limit checking bug, 'OUT byte nn, reg' bug, and a couple of rdoff
2359related bugs, updated Wishlist; 0.98 Prerelease 3.
2360
2361
2362\S{cl-0.98p2} Version 0.98 pre-release 2
2363
2364\b fixed bug in outcoff.c to do with truncating section names longer
2365than 8 characters, referencing beyond end of string; 0.98 pre-release 2
2366
2367
2368\S{cl-0.98p1} Version 0.98 pre-release 1
2369
2370\b Fixed a bug whereby STRUC didn't work at all in RDF.
2371
2372\b Fixed a problem with group specification in PUBDEFs in OBJ.
2373
2374\b Improved ease of adding new output formats. Contribution due to
2375Fox Cutter.
2376
2377\b Fixed a bug in relocations in the `bin' format: was showing up when
2378a relocatable reference crossed an 8192-byte boundary in any output
2379section.
2380
2381\b Fixed a bug in local labels: local-label lookups were inconsistent
2382between passes one and two if an EQU occurred between the definition
2383of a global label and the subsequent use of a local label local to
2384that global.
2385
2386\b Fixed a seg-fault in the preprocessor (again) which happened when
2387you use a blank line as the first line of a multi-line macro
2388definition and then defined a label on the same line as a call to
2389that macro.
2390
2391\b Fixed a stale-pointer bug in the handling of the NASM environment
2392variable. Thanks to Thomas McWilliams.
2393
2394\b ELF had a hard limit on the number of sections which caused
2395segfaults when transgressed. Fixed.
2396
2397\b Added ability for ndisasm to read from stdin by using `-' as the
2398filename.
2399
2400\b ndisasm wasn't outputting the TO keyword. Fixed.
2401
2402\b Fixed error cascade on bogus expression in \c{%if} - an error in
2403evaluation was causing the entire \c{%if} to be discarded, thus creating
2404trouble later when the \c{%else} or \c{%endif} was encountered.
2405
2406\b Forward reference tracking was instruction-granular not operand-
2407granular, which was causing 286-specific code to be generated
2408needlessly on code of the form `shr word [forwardref],1'. Thanks to
2409Jim Hague for sending a patch.
2410
2411\b All messages now appear on stdout, as sending them to stderr serves
2412no useful purpose other than to make redirection difficult.
2413
2414\b Fixed the problem with EQUs pointing to an external symbol - this
2415now generates an error message.
2416
2417\b Allowed multiple size prefixes to an operand, of which only the first
2418is taken into account.
2419
2420\b Incorporated John Fine's changes, including fixes of a large number
2421of preprocessor bugs, some small problems in OBJ, and a reworking of
2422label handling to define labels before their line is assembled, rather
2423than after.
2424
2425\b Reformatted a lot of the source code to be more readable. Included
2426'coding.txt' as a guideline for how to format code for contributors.
2427
2428\b Stopped nested \c{%reps} causing a panic - they now cause a slightly more
2429friendly error message instead.
2430
2431\b Fixed floating point constant problems (patch by Pedro Gimeno)
2432
2433\b Fixed the return value of insn_size() not being checked for -1, indicating
2434an error.
2435
2436\b Incorporated 3Dnow! instructions.
2437
2438\b Fixed the 'mov eax, eax + ebx' bug.
2439
2440\b Fixed the GLOBAL EQU bug in ELF. Released developers release 3.
2441
2442\b Incorporated John Fine's command line parsing changes
2443
2444\b Incorporated David Lindauer's OMF debug support
2445
2446\b Made changes for LCC 4.0 support (\c{__NASM_CDecl__}, removed register size
2447specification warning when sizes agree).
2448
2449
2450\H{cl-0.9x} NASM 0.9 Series
2451
2452Revisions before 0.98.
2453
2454
2455\S{cl-0.97} Version 0.97 released December 1997
2456
2457\b This was entirely a bug-fix release to 0.96, which seems to have got
2458cursed. Silly me.
2459
2460\b Fixed stupid mistake in OBJ which caused `MOV EAX,<constant>' to
2461fail. Caused by an error in the `MOV EAX,<segment>' support.
2462
2463\b ndisasm hung at EOF when compiled with lcc on Linux because lcc on
2464Linux somehow breaks feof(). ndisasm now does not rely on feof().
2465
2466\b A heading in the documentation was missing due to a markup error in
2467the indexing. Fixed.
2468
2469\b Fixed failure to update all pointers on realloc() within extended-
2470operand code in parser.c. Was causing wrong behaviour and seg faults
2471on lines such as `dd 0.0,0.0,0.0,0.0,...'
2472
2473\b Fixed a subtle preprocessor bug whereby invoking one multi-line
2474macro on the first line of the expansion of another, when the second
2475had been invoked with a label defined before it, didn't expand the
2476inner macro.
2477
2478\b Added internal.doc back in to the distribution archives - it was
2479missing in 0.96 *blush*
2480
2481\b Fixed bug causing 0.96 to be unable to assemble its own test files,
2482specifically objtest.asm. *blush again*
2483
2484\b Fixed seg-faults and bogus error messages caused by mismatching
2485\c{%rep} and \c{%endrep} within multi-line macro definitions.
2486
2487\b Fixed a problem with buffer overrun in OBJ, which was causing
2488corruption at ends of long PUBDEF records.
2489
2490\b Separated DOS archives into main-program and documentation to reduce
2491download size.
2492
2493
2494\S{cl-0.96} Version 0.96 released November 1997
2495
2496\b Fixed a bug whereby, if `nasm sourcefile' would cause a filename
2497collision warning and put output into `nasm.out', then `nasm
2498sourcefile -o outputfile' still gave the warning even though the
2499`-o' was honoured.
2500Fixed name pollution under Digital UNIX: one of its header files
2501defined R_SP, which broke the enum in nasm.h.
2502
2503\b Fixed minor instruction table problems: FUCOM and FUCOMP didn't have
2504two-operand forms; NDISASM didn't recognise the longer register
2505forms of PUSH and POP (eg FF F3 for PUSH BX); TEST mem,imm32 was
2506flagged as undocumented; the 32-bit forms of CMOV had 16-bit operand
2507size prefixes; `AAD imm' and `AAM imm' are no longer flagged as
2508undocumented because the Intel Architecture reference documents
2509them.
2510
2511\b Fixed a problem with the local-label mechanism, whereby strange
2512types of symbol (EQUs, auto-defined OBJ segment base symbols)
2513interfered with the `previous global label' value and screwed up
2514local labels.
2515
2516\b Fixed a bug whereby the stub preprocessor didn't communicate with
2517the listing file generator, so that the -a and -l options in
2518conjunction would produce a useless listing file.
2519
2520\b Merged `os2' object file format back into `obj', after discovering
2521that `obj' _also_ shouldn't have a link pass separator in a module
2522containing a non-trivial MODEND. Flat segments are now declared
2523using the FLAT attribute. `os2' is no longer a valid object format
2524name: use `obj'.
2525
2526\b Removed the fixed-size temporary storage in the evaluator. Very very
2527long expressions (like `mov ax,1+1+1+1+...' for two hundred 1s or
2528so) should now no longer crash NASM.
2529
2530\b Fixed a bug involving segfaults on disassembly of MMX instructions,
2531by changing the meaning of one of the operand-type flags in nasm.h.
2532This may cause other apparently unrelated MMX problems; it needs to
2533be tested thoroughly.
2534
2535\b Fixed some buffer overrun problems with large OBJ output files.
2536Thanks to DJ Delorie for the bug report and fix.
2537
2538\b Made preprocess-only mode actually listen to the \c{%line} markers as it
2539prints them, so that it can report errors more sanely.
2540
2541\b Re-designed the evaluator to keep more sensible track of expressions
2542involving forward references: can now cope with previously-nightmare
2543situations such as:
2544
2545\c   mov ax,foo | bar
2546\c   foo equ 1
2547\c   bar equ 2
2548
2549\b Added the ALIGN and ALIGNB standard macros.
2550
2551\b Added PIC support in ELF: use of WRT to obtain the four extra
2552relocation types needed.
2553
2554\b Added the ability for output file formats to define their own
2555extensions to the GLOBAL, COMMON and EXTERN directives.
2556
2557\b Implemented common-variable alignment, and global-symbol type and
2558size declarations, in ELF.
2559
2560\b Implemented NEAR and FAR keywords for common variables, plus
2561far-common element size specification, in OBJ.
2562
2563\b Added a feature whereby EXTERNs and COMMONs in OBJ can be given a
2564default WRT specification (either a segment or a group).
2565
2566\b Transformed the Unix NASM archive into an auto-configuring package.
2567
2568\b Added a sanity-check for people applying SEG to things which are
2569already segment bases: this previously went unnoticed by the SEG
2570processing and caused OBJ-driver panics later.
2571
2572\b Added the ability, in OBJ format, to deal with `MOV EAX,<segment>'
2573type references: OBJ doesn't directly support dword-size segment
2574base fixups, but as long as the low two bytes of the constant term
2575are zero, a word-size fixup can be generated instead and it will
2576work.
2577
2578\b Added the ability to specify sections' alignment requirements in
2579Win32 object files and pure binary files.
2580
2581\b Added preprocess-time expression evaluation: the \c{%assign} (and
2582\c{%iassign}) directive and the bare \c{%if} (and \c{%elif}) conditional. Added
2583relational operators to the evaluator, for use only in \c{%if}
2584constructs: the standard relationals = < > <= >= <> (and C-like
2585synonyms == and !=) plus low-precedence logical operators &&, ^^ and
2586||.
2587
2588\b Added a preprocessor repeat construct: \c{%rep} / \c{%exitrep} / \c{%endrep}.
2589
2590\b Added the __FILE__ and __LINE__ standard macros.
2591
2592\b Added a sanity check for number constants being greater than
25930xFFFFFFFF. The warning can be disabled.
2594
2595\b Added the %0 token whereby a variadic multi-line macro can tell how
2596many parameters it's been given in a specific invocation.
2597
2598\b Added \c{%rotate}, allowing multi-line macro parameters to be cycled.
2599
2600\b Added the `*' option for the maximum parameter count on multi-line
2601macros, allowing them to take arbitrarily many parameters.
2602
2603\b Added the ability for the user-level forms of EXTERN, GLOBAL and
2604COMMON to take more than one argument.
2605
2606\b Added the IMPORT and EXPORT directives in OBJ format, to deal with
2607Windows DLLs.
2608
2609\b Added some more preprocessor \c{%if} constructs: \c{%ifidn} / \c{%ifidni} (exact
2610textual identity), and \c{%ifid} / \c{%ifnum} / \c{%ifstr} (token type testing).
2611
2612\b Added the ability to distinguish SHL AX,1 (the 8086 version) from
2613SHL AX,BYTE 1 (the 286-and-upwards version whose constant happens to
2614be 1).
2615
2616\b Added NetBSD/FreeBSD/OpenBSD's variant of a.out format, complete
2617with PIC shared library features.
2618
2619\b Changed NASM's idiosyncratic handling of FCLEX, FDISI, FENI, FINIT,
2620FSAVE, FSTCW, FSTENV, and FSTSW to bring it into line with the
2621otherwise accepted standard. The previous behaviour, though it was a
2622deliberate feature, was a deliberate feature based on a
2623misunderstanding. Apologies for the inconvenience.
2624
2625\b Improved the flexibility of ABSOLUTE: you can now give it an
2626expression rather than being restricted to a constant, and it can
2627take relocatable arguments as well.
2628
2629\b Added the ability for a variable to be declared as EXTERN multiple
2630times, and the subsequent definitions are just ignored.
2631
2632\b We now allow instruction prefixes (CS, DS, LOCK, REPZ etc) to be
2633alone on a line (without a following instruction).
2634
2635\b Improved sanity checks on whether the arguments to EXTERN, GLOBAL
2636and COMMON are valid identifiers.
2637
2638\b Added misc/exebin.mac to allow direct generation of .EXE files by
2639hacking up an EXE header using DB and DW; also added test/binexe.asm
2640to demonstrate the use of this. Thanks to Yann Guidon for
2641contributing the EXE header code.
2642
2643\b ndisasm forgot to check whether the input file had been successfully
2644opened. Now it does. Doh!
2645
2646\b Added the Cyrix extensions to the MMX instruction set.
2647
2648\b Added a hinting mechanism to allow [EAX+EBX] and [EBX+EAX] to be
2649assembled differently. This is important since [ESI+EBP] and
2650[EBP+ESI] have different default base segment registers.
2651
2652\b Added support for the PharLap OMF extension for 4096-byte segment
2653alignment.
2654
2655
2656\S{cl-0.95 released July 1997} Version 0.95 released July 1997
2657
2658\b Fixed yet another ELF bug. This one manifested if the user relied on
2659the default segment, and attempted to define global symbols without
2660first explicitly declaring the target segment.
2661
2662\b Added makefiles (for NASM and the RDF tools) to build Win32 console
2663apps under Symantec C++. Donated by Mark Junker.
2664
2665\b Added `macros.bas' and `insns.bas', QBasic versions of the Perl
2666scripts that convert `standard.mac' to `macros.c' and convert
2667`insns.dat' to `insnsa.c' and `insnsd.c'. Also thanks to Mark
2668Junker.
2669
2670\b Changed the diassembled forms of the conditional instructions so
2671that JB is now emitted as JC, and other similar changes. Suggested
2672list by Ulrich Doewich.
2673
2674\b Added `@' to the list of valid characters to begin an identifier
2675with.
2676
2677\b Documentary changes, notably the addition of the `Common Problems'
2678section in nasm.doc.
2679
2680\b Fixed a bug relating to 32-bit PC-relative fixups in OBJ.
2681
2682\b Fixed a bug in perm_copy() in labels.c which was causing exceptions
2683in cleanup_labels() on some systems.
2684
2685\b Positivity sanity check in TIMES argument changed from a warning to
2686an error following a further complaint.
2687
2688\b Changed the acceptable limits on byte and word operands to allow
2689things like `~10111001b' to work.
2690
2691\b Fixed a major problem in the preprocessor which caused seg-faults if
2692macro definitions contained blank lines or comment-only lines.
2693
2694\b Fixed inadequate error checking on the commas separating the
2695arguments to `db', `dw' etc.
2696
2697\b Fixed a crippling bug in the handling of macros with operand counts
2698defined with a `+' modifier.
2699
2700\b Fixed a bug whereby object file formats which stored the input file
2701name in the output file (such as OBJ and COFF) weren't doing so
2702correctly when the output file name was specified on the command
2703line.
2704
2705\b Removed [INC] and [INCLUDE] support for good, since they were
2706obsolete anyway.
2707
2708\b Fixed a bug in OBJ which caused all fixups to be output in 16-bit
2709(old-format) FIXUPP records, rather than putting the 32-bit ones in
2710FIXUPP32 (new-format) records.
2711
2712\b Added, tentatively, OS/2 object file support (as a minor variant on
2713OBJ).
2714
2715\b Updates to Fox Cutter's Borland C makefile, Makefile.bc2.
2716
2717\b Removed a spurious second fclose() on the output file.
2718
2719\b Added the `-s' command line option to redirect all messages which
2720would go to stderr (errors, help text) to stdout instead.
2721
2722\b Added the `-w' command line option to selectively suppress some
2723classes of assembly warning messages.
2724
2725\b Added the `-p' pre-include and `-d' pre-define command-line options.
2726
2727\b Added an include file search path: the `-i' command line option.
2728
2729\b Fixed a silly little preprocessor bug whereby starting a line with a
2730`%!' environment-variable reference caused an `unknown directive'
2731error.
2732
2733\b Added the long-awaited listing file support: the `-l' command line
2734option.
2735
2736\b Fixed a problem with OBJ format whereby, in the absence of any
2737explicit segment definition, non-global symbols declared in the
2738implicit default segment generated spurious EXTDEF records in the
2739output.
2740
2741\b Added the NASM environment variable.
2742
2743\b From this version forward, Win32 console-mode binaries will be
2744included in the DOS distribution in addition to the 16-bit binaries.
2745Added Makefile.vc for this purpose.
2746
2747\b Added `return 0;' to test/objlink.c to prevent compiler warnings.
2748
2749\b Added the __NASM_MAJOR__ and __NASM_MINOR__ standard defines.
2750
2751\b Added an alternative memory-reference syntax in which prefixing an
2752operand with `&' is equivalent to enclosing it in square brackets,
2753at the request of Fox Cutter.
2754
2755\b Errors in pass two now cause the program to return a non-zero error
2756code, which they didn't before.
2757
2758\b Fixed the single-line macro cycle detection, which didn't work at
2759all on macros with no parameters (caused an infinite loop). Also
2760changed the behaviour of single-line macro cycle detection to work
2761like cpp, so that macros like `extrn' as given in the documentation
2762can be implemented.
2763
2764\b Fixed the implementation of WRT, which was too restrictive in that
2765you couldn't do `mov ax,[di+abc wrt dgroup]' because (di+abc) wasn't
2766a relocatable reference.
2767
2768
2769\S{cl-0.94 released April 1997} Version 0.94 released April 1997
2770
2771
2772\b Major item: added the macro processor.
2773
2774\b Added undocumented instructions SMI, IBTS, XBTS and LOADALL286. Also
2775reorganised CMPXCHG instruction into early-486 and Pentium forms.
2776Thanks to Thobias Jones for the information.
2777
2778\b Fixed two more stupid bugs in ELF, which were causing `ld' to
2779continue to seg-fault in a lot of non-trivial cases.
2780
2781\b Fixed a seg-fault in the label manager.
2782
2783\b Stopped FBLD and FBSTP from _requiring_ the TWORD keyword, which is
2784the only option for BCD loads/stores in any case.
2785
2786\b Ensured FLDCW, FSTCW and FSTSW can cope with the WORD keyword, if
2787anyone bothers to provide it. Previously they complained unless no
2788keyword at all was present.
2789
2790\b Some forms of FDIV/FDIVR and FSUB/FSUBR were still inverted: a
2791vestige of a bug that I thought had been fixed in 0.92. This was
2792fixed, hopefully for good this time...
2793
2794\b Another minor phase error (insofar as a phase error can _ever_ be
2795minor) fixed, this one occurring in code of the form
2796
2797\c   rol ax,forward_reference
2798\c   forward_reference equ 1
2799
2800\b The number supplied to TIMES is now sanity-checked for positivity,
2801and also may be greater than 64K (which previously didn't work on
280216-bit systems).
2803
2804\b Added Watcom C makefiles, and misc/pmw.bat, donated by Dominik Behr.
2805
2806\b Added the INCBIN pseudo-opcode.
2807
2808\b Due to the advent of the preprocessor, the [INCLUDE] and [INC]
2809directives have become obsolete. They are still supported in this
2810version, with a warning, but won't be in the next.
2811
2812\b Fixed a bug in OBJ format, which caused incorrect object records to
2813be output when absolute labels were made global.
2814
2815\b Updates to RDOFF subdirectory, and changes to outrdf.c.
2816
2817
2818\S{cl-0.93 released January 1997} Version 0.93 released January 1997
2819
2820This release went out in a great hurry after semi-crippling bugs
2821were found in 0.92.
2822
2823\b Really \e{did} fix the stack overflows this time. *blush*
2824
2825\b Had problems with EA instruction sizes changing between passes, when
2826an offset contained a forward reference and so 4 bytes were
2827allocated for the offset in pass one; by pass two the symbol had
2828been defined and happened to be a small absolute value, so only 1
2829byte got allocated, causing instruction size mismatch between passes
2830and hence incorrect address calculations. Fixed.
2831
2832\b Stupid bug in the revised ELF section generation fixed (associated
2833string-table section for .symtab was hard-coded as 7, even when this
2834didn't fit with the real section table). Was causing `ld' to
2835seg-fault under Linux.
2836
2837\b Included a new Borland C makefile, Makefile.bc2, donated by Fox
2838Cutter <lmb@comtch.iea.com>.
2839
2840
2841\S{cl-0.92 released January 1997} Version 0.92 released January 1997
2842
2843\b The FDIVP/FDIVRP and FSUBP/FSUBRP pairs had been inverted: this was
2844fixed. This also affected the LCC driver.
2845
2846\b Fixed a bug regarding 32-bit effective addresses of the form
2847\c{[other_register+ESP]}.
2848
2849\b Documentary changes, notably documentation of the fact that Borland
2850Win32 compilers use `obj' rather than `win32' object format.
2851
2852\b Fixed the COMENT record in OBJ files, which was formatted
2853incorrectly.
2854
2855\b Fixed a bug causing segfaults in large RDF files.
2856
2857\b OBJ format now strips initial periods from segment and group
2858definitions, in order to avoid complications with the local label
2859syntax.
2860
2861\b Fixed a bug in disassembling far calls and jumps in NDISASM.
2862
2863\b Added support for user-defined sections in COFF and ELF files.
2864
2865\b Compiled the DOS binaries with a sensible amount of stack, to
2866prevent stack overflows on any arithmetic expression containing
2867parentheses.
2868
2869\b Fixed a bug in handling of files that do not terminate in a newline.
2870
2871
2872\S{cl-0.91 released November 1996} Version 0.91 released November 1996
2873
2874\b Loads of bug fixes.
2875
2876\b Support for RDF added.
2877
2878\b Support for DBG debugging format added.
2879
2880\b Support for 32-bit extensions to Microsoft OBJ format added.
2881
2882\b Revised for Borland C: some variable names changed, makefile added.
2883
2884\b LCC support revised to actually work.
2885
2886\b JMP/CALL NEAR/FAR notation added.
2887
2888\b `a16', `o16', `a32' and `o32' prefixes added.
2889
2890\b Range checking on short jumps implemented.
2891
2892\b MMX instruction support added.
2893
2894\b Negative floating point constant support added.
2895
2896\b Memory handling improved to bypass 64K barrier under DOS.
2897
2898\b \c{$} prefix to force treatment of reserved words as identifiers added.
2899
2900\b Default-size mechanism for object formats added.
2901
2902\b Compile-time configurability added.
2903
2904\b \c{#}, \c{@}, \c{~} and c\{?} are now valid characters in labels.
2905
2906\b \c{-e} and \c{-k} options in NDISASM added.
2907
2908
2909\S{cl-0.90 released October 1996} Version 0.90 released October 1996
2910
2911First release version. First support for object file output. Other
2912changes from previous version (0.3x) too numerous to document.
2913