NASM - The Netwide Assembler (original) (raw)

Appendix C: NASM Version History

C.1. NASM 3 Series

The NASM 3 series added support for the APX instruction encodings (extended GPRs), as well as preprocessor enhancements meant to make evolving code simpler.

It is the production version of NASM since 2025.

C.1.1. Version 3.01

A new obj2 version of the obj output format, intended for use on OS/2. See section 9.5.
The condition after %if or %elif would be evaluated while output is suppressed after %exitrep or%exitmacro. Although no output would be generated in either case, assembly would fail if evaluating the expression triggered an error.
Fix encoding of TCVTROWPS2PHL, correct multiple AVX512-BF16 instructions' operand formats and typoed mnemonics.
The unofficial but obvious alternate form TEST reg,mem was not accepted by NASM 3.00; corrected.
For the obj output format, multiple GROUPdirectives can now be specified for the same group; the resulting group includes all sections specified in all GROUP directives for the group.
A new %selbits() preprocessor function. Seesection 5.4.19.
A new --bits option as convenience shorthand for--before "BITS ...". Seesection 2.1.31.
The options and pragmas for configuring external label mangling were inconsistent, the former using the spelling postfix and the latter suffix. Furthermore, these were also documented as_directives_ in addition to pragmas. Implement the already documented directives (bracketed forms only) and allow bothpostfix and suffix in all cases.
See section 2.1.28 andsection 8.10.
Define additional permissive patterns and fix several opcode bugs.
Fix parsing of two-operand forms of x87 instructions.
Fix bogus "absolute address can not be RIP-relative" warning.
Hopefully fix building with OpenWatcom.
Generate a warning, promoted to error by default, on the use ofo64 prefixes in 16- or 32-bit mode. If demoted to a warning or suppressed the prefix is ignored, but likely will trigger subsequent, harder to debug, error messages.
More consistent handling of jump and call instructions with specified operand sizes.
Fix an operand size handling bug in the CMPXCHGinstruction.

C.1.2. Version 3.00

Improve the documentation for building from source (appendix D).
Add support for the APX and AVX10 instruction sets, and various miscellaneous new instructions.
Add new preprocessor functions: %b2hs(),%chr(), %depend(), %find(),%findi(), %hs2b(), %null(),%ord(), %pathsearch(), and%realpath(). See section 5.4.
New preprocessor directive %note to insert a note in the list file, without issuing an external diagnosic. Unlike a comment, it is optionally macro-expanded, see section 5.11.
New preprocessor directive %iffile (and corresponding function %isfile()) to test for the existence of a file. Seesection 5.6.12.
New preprocessor directive %ifdirective to test for the existence of a preprocessor directive, assembly directive, or pseudo-instruction; see section 5.6.10.
Fix a number of invalid memory references (usually causing crashes) on various invalid inputs.
Fix multiple bugs in the handling of $–escaped symbols.
The use of $ as a prefix for hexadecimal numbers has been deprecated, and will now issue a warning. A new directive[DOLLARHEX] can be used to disable this syntax entirely, seesection 8.12.
Fix the generation of segment selector references (mainly used in theobj output format.)
Fix crash in the obj backend when code was emitted into the default segment, without any labels having been defined.
Clean up the command-line help text (-h) and break it down into individual topics, as the previous output was just too verbose to be practical as a quick reference.
The implicit DEFAULT ABS in 64-bit mode is deprecated and may be changed to REL in the future. Seesection 8.2. A warning is now emitted for this condition.
It is now possible to set the REL/ABS default for memory accesses using FS: or GS:, seesection 8.2.
The __?DEFAULT?__ standard macro now reflects the settings of the DEFAULT directive. Seesection 6.4.
The NASM preprocessor now assumes that an unknown directive starting with %if or %elif is a misspelled or not yet implemented conditional directive, and tries to match it with a corresponding %endif. Seesection 5.6.14.
The masm macro package now defines a macro for x87 register syntax. See section 7.5.
A new macro package, vtern, to simplify generation of the control immediates for the VPTERNLOGD andVPTERNLOGQ instructions. Seesection 7.6.
A new command line option -LF allows overriding[LIST -] directives.
In the obj output format, allow a segment in theFLAT pseudo-group to also belong to another (real) group. Used on OS/2.
Add a new build_version directive to the Mach-O backend. See section 9.9.6.
Fix a spec violation in the generation of DWARF debugging information on ELF.
Response files can now be nested.
Many documentation improvements.

C.2. NASM 2 Series

The NASM 2 series added support for x86-64, and was the production versions of NASM from 2007 to 2025.

C.2.1. Version 2.16.03

This is a source build machinery and documentation update only. There are no functionality changes.

Fix building from git in a separate directory from the source.
Remove some irrelevant files from the source distribution.
Make the documentation stronger that -O0 or-O1 are probably not what the user wants. Seesection 2.1.24.
Fix configure --enable-lto build option.
Update the included RPM .spec file.

C.2.2. Version 2.16.02

Fix building from the source distribution in a separate directory from the source.
Fix a number of issues when building from source, mostly involvingconfigure or dependency generation.
In particular, more aggressively avoid cross-compilation problems on Unix/Linux systems automatically invoking WINE. We could end up invoking WINE even when we didn't want to, making configure think it was running native when in fact cross-compiling.
Hopefully fix compiling with the latest versions of MSVC/nmake.
Windows host: add embedded manifest file. Without a manifest, Windows applications force a fixed PATH_MAX limit to any pathname; this is unnecessary.
Add support VEX-encoded SM4-NI instructions.
Add support for VEX-encoded SM3-NI instructions.
Add support for VEX-encoded SHA512-NI instructions.
PTWRITE opcode corrected (F3 prefix required.)
Disassembler: the SMAP instructions are NP; notably the prefixed versions of CLAC areERETU/ERETS.
Add support for Flexible Return and Exception Delivery (FRED): theLKGS, ERETS and ERETU instructions.
Fix external references to segments in the obj (OMF) and possibly other output formats.
Always support up to 8 characters, i.e. 64 bits, in a string-to-numeric conversion.
Preprocessor: add %map() function to expand a macro from a list of arguments, see section 5.4.12.
Preprocessor: allow the user to specify the desired radix for an evaluated parameter. It doesn't make any direct difference, but can be nice for debugging or turning into strings. See the = modifier insection 5.2.1.
Update documentation: __USE_package__is now __?USE_package?__.
Documentation: correct a minor problem in the expression grammar forDx statements, seesection 3.2.1.
Preprocessor: correctly handle empty %rep blocks.
Preprocessor: add options for a base prefix to %num(), seesection 5.4.14.
Preprocessor: add a %hex() function, equivalent to%eval() except that it producess hexadecimal values that are nevertheless valid NASM numeric constants, seesection 5.4.9.
Preprocessor: fix the parameter number in error messages (should be 1-based, like %num references to multi-line macro arguments.)
Documentation: be more clear than the bin format is simply a linker built into NASM. See section 9.1.
Adjust the LOCK prefix warning for XCHG.
LOCK XCHG reg,mem would issue a warning for being unlockable, which is incorrect. In this case the_reg_,mem encoding is simply an alias for the_mem_,reg encoding. However, XCHGis always locked, so create a new warning (-w+prefix-lock-xchg) to explicitly flag a user-specifiedLOCK XCHG; default off. Future versions of NASM may remove theLOCK prefix when optimization is enabled.
Fix broken dependency-list generation.
Add optional warnings for specific relocation types (-w+reloc-*, see appendix A), default off.
Some target environments may have specific restrictions on what kinds of relocations are possible or allowed.
Error out on certain bad syntax in Dx statements, such as db 1 2. Seesection 3.2.1.

C.2.3. Version 2.16.01

This is a documentation update release only. There are no functionality changes.

Fix the creation of the table of contents in the HTML version of the documentation.

C.2.4. Version 2.16

Support for the rdf format has been discontinued and all the RDOFF utilities has been removed.
The --reproducible option now leaves the filename field in the COFF object format blank. This was always rather useless since it is only 18 characters long; as such debug formats have to carry their own filename information anyway.
Fix handling of MASM-syntax reserved memory (e.g. dw ?) when used in structure definitions.
The preprocessor now supports functions, which can be less verbose and more convenient than the equivalent code implemented using directives. Seesection 5.4.
Fix the handling of %00 in the preprocessor.
Fix incorrect handling of path names affecting error messages, dependency generation, and debug format output.
Support for the RDOFF output format and the RDOFF tools have been removed. The RDOFF tools had already been broken since at least NASM 2.14. For flat code the ELF output format recommended; for segmented code theobj (OMF) output format.
New facility: preprocessor functions. Preprocessor functions, which are expanded similarly to single-line macros, can greatly simplify code that in the past would have required a lengthy list of directives and intermediate macros. See section 5.4.
Single-line macros can now declare parameters (using a&& prefix) that creates a quoted string, but does_not_ requote an already quoted string. Seesection 5.2.1.
Instruction table updated per public information available as of November 2022.
All warnings in the preprocessor have now been assigned warning classes. See appendix A.
Fix the invalid use of RELA–type relocations instead of REL–type relocations when generating DWARF debug information for the elf32 output format.
Fix the handling at in istruc when the structure contains local labels. Seesection 6.11.2.
When assembling with --reproducible, don't encode the filename in the COFF header for the coff, win32or win64 output formats. The COFF header only has space for an 18-character filename, which makes this field rather useless in the first place. Debug output data, if enabled, is not affected.
Fix incorrect size calculation when using MASM syntax for non-byte reservations (e.g. dw ?.)
Allow forcing an instruction in 64-bit mode to have a (possibly redundant) REX prefix, using the syntax {rex} as a prefix.
Add a {vex} prefix to enforce VEX (AVX) encoding of an instruction, either using the 2- or 3-byte VEX prefixes.
The CPU directive has been augmented to allow control of generation of VEX (AVX) versus EVEX (AVX-512) instruction formats, seesection 8.11.
Some recent instructions that previously have been only available using EVEX encodings are now also encodable using VEX (AVX) encodings. For backwards compatibility these encodings are not enabled by default, but can be generated either via an explicit {vex} prefix or by specifying either CPU LATEVEX or CPU NOEVEX; seesection 8.11.
Document the already existing %unimacro directive. Seesection 5.5.12.
Fix a code range generation bug in the DWARF debug format (incorrect information in the DW_AT_high_pc field) for the ELF output formats. This bug happened to cancel out with a bug in older versions of the GNU binutils linker, but breaks with other linkers and updated or other linkers that expect the spec to be followed.
Fix segment symbols with addends, e.g. jmp _TEXT+10h:0 in output formats that support segment relocations, e.g. the objformat.
Fix various crashes and hangs on invalid input.

C.2.5. Version 2.15.05

Fix %ifid <math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi mathvariant="normal">‘</mi><mi>a</mi><mi>n</mi><mi>d</mi><mi mathvariant="normal">‘</mi></mrow><annotation encoding="application/x-tex"> and %ifid </annotation></semantics></math>‘and‘$ incorrectly being treated as true. See section 5.6.7.
Add --reproducible option to suppress NASM version numbers and timestamps in output files. Seesection 2.1.35.

C.2.6. Version 2.15.04

More sensible handling of the case where one single-line macro definition will shadow another. A warning will be issued, but the additional definition will be allowed. For the existing error case where both a parameterless and parametered macro are created, that warning is promoted to an error by default.
Add special preprocessor tokens %*? and %*??that expand like %? and %?? in single-line macros only. See section 5.2.6.
Correct the encoding of the ENQCMDS andTILELOADT1 instructions.
Fix case where the COFF backend (the coff,win32 and win64 output formats) would add padding bytes in the middle of a section if aSECTION/SEGMENT directive was provided which repeated an ALIGN= attribute. This neither matched legacy behavior, other backends, or user expectations.
Fix SSE instructions not being recognized with an explicit memory operation size (e.g. movsd qword [eax],xmm0).
The -L+ option no longer enables -Lw, which is mainly useful to debug NASM crashes. Seesection 2.1.4.
Document long-standing hazards in the use of $ inDx statements, seesection 3.2.1.
The NASM-only RDOFF output format backend, which has been broken since at least NASM 2.14, has been disabled. The RDOFF tools are scheduled to be removed from the NASM distribution in NASM 2.16. If you have a concrete use case for RDOFF, please file a NASM bug report athttps://bugs.nasm.us/ as soon as possible.

C.2.7. Version 2.15.03

Add instructions from the Intel Instruction Set Extensions and Future Features Programming Reference, June 2020. This includes AVX5512bfloat16, AVX512 mask intersect, and Intel Advanced Matrix Extensions (AMX).
Support for bfloat16 floating-point constants. Seesection 3.4.6 andsection 7.3.
Properly display warnings in preprocess-only mode.
Fix copy-and-paste of examples from the PDF documentation.
Debug information now properly reflect the line numbers of macro invocations (unless declared .nolist).
Fix excessive alignment of sections in thecoff/win32/win64 output formats when the user-specified alignment is less than the default alignment for the section or section type.
Fix explicit token pasting (%+,section 5.2.4) for the cases where one or more parts result from empty token expansion, resulting in%+ tokens at the beginning or end, or multiple ones in a row.
Fix macro label capture (%00,section 5.5.7).
Much better documentation for the MASM compatibility package,%use masm (see section 7.5).
Fix LEA without square brackets, for MASM compatibility.
Portability fixes.

C.2.8. Version 2.15.02

Fix miscompilation when building with clang.
Add db-empty warning class, seesection 2.1.26.
Fix the dependencies in the MSVC NMAKE makefile (Mkfiles/msvc.mak).
Some documentation improvements and cleanups.
Fix the handling of macro parameter ranges (%{:}), including with brace-enclosed original arguments.

C.2.9. Version 2.15.01

Fix building the documentation from the release archive. For 2.15, the user has to do make warnings manually in the main directory in order to be able to build the documentation, which means Perl needs to be installed on the system.
Add instructions for Intel Control Flow Enforcement Technology (CET).

C.2.10. Version 2.15

The comparison and booleanizing operators can now be used in any expression context, not just %if. Seesection 3.5.
New operator ? ... :. Seesection 3.5.1.
Signed shift operators <<< and>>>. See section 3.5.9.
The MASM DUP syntax for data definitions is now supported, in a somewhat enhanced form. Seesection 3.2.1.
Warn for strange legacy behavior regarding empty arguments in multi-line macro expansion, but try to match legacy behavior in most cases. Legacy behavior can be disabled with the directive%pragma preproc sane_empty_expansion, seesection 5.5 andsection 5.12.1.
A much more sensible limit to expression evaluation depth. The previously defined limit would rarely trigger before NASM died with a stack overrun error on most systems. Seesection 2.1.32.
The state of warnings can now be saved and restored via the[WARNING PUSH] and [WARNING POP] directives. Seesection 8.14.
The sectalign on|off switch does not affect an explicit directive. See section 6.12.2.
Added configure option to enable building with profiling (--enable-profiling).
Attempt to support of long path names, up to 32767 of UTF-16 characters, on Windows.
Fixed 'mismatch in operand sizes' error in the MOVDDUP,CMPXCHG8B and CMPXCHG16B instructions.
Improved error messages in the string transformation routine.
Removed obsolete gnu-elf-extensions warning about 8- and 16-bit relocation generation. Seesection 9.10.8
Added group aliases for all prefixed warnings. Seesection 2.1.26.
Allowed building with MSVC versions older than 1700.
Added implicitly sized versions of the K... instructions, which allows the K... instructions to be specified without a size suffix as long as the operands are sized.
Added -L option for additional listing information. Seesection 2.1.4.
Added some warnings for obsolete instructions for a specified CPU.
Deprecated -hf and -y options. Use-h instead.
Made DWARF as the default debug format for ELF.
Added %pragma list options... to set or clear listing options (see opt-L).
Allowed immediate syntax for LEA instruction (ignore operand size completely).
Added limited functionality MASM compatibility package. Seesection 7.5.
Add single-line macros aliases using %defalias or%idefalias. These behave like a kind of "symbolic links" for single-line macros. See section 5.2.11 and clear.
Added support for stringify, nostrip, evaluating, and greedy single-line macro arguments. See section 5.2.1.
Unused single-line macro arguments no longer need to have a specified name. See section 5.2.1.
Added conditional comma operator %,. Seesection 5.2.12.
Changed private namespace from __foo__ to__?foo?__, so a user namespace starting from underscore is now clean from symbols. For backwards compatibility, the previous names are defined as aliases; see section 5.2.11, section 5.13.3 andchapter 6.
Added support of ELF weak symbols and external references. Seesection 9.10.5.
Changed the behavior of the EXTERN keyword and introduced REQUIRED keyword. See section 8.6.
Added %ifusable and %ifusing directives. Seechapter 7.
Made various performance improvements and stability fixes in macro preprocessor engine.
Improved NASM error handling and cleaned up error messages.
Many, many bug fixes.

C.2.11. Version 2.14.03

Suppress nuisance "label changed during code generation" messages after a real error.
Add support for the merge and stringsattributes on ELF sections. Seesection 9.10.2.
Add support for the note, preinit_array,init_array, and fini_array sections type in ELF. See section 9.10.2.
Handle more than 32,633 sections in ELF.

C.2.12. Version 2.14.02

Fix crash due to multiple errors or warnings during the code generation pass if a list file is specified.

C.2.13. Version 2.14.01

Create all system-defined macros before processing command-line given preprocessing directives (-p, -d,-u, --pragma, --before).
If debugging is enabled, define a __DEBUG_FORMAT__predefined macro. See section 6.6.
Fix an assert for the case in the obj format when aSEG operator refers to an EXTERN symbol declared further down in the code.
Fix a corner case in the floating-point code where a binary, octal or hexadecimal floating-point having at least 32, 11, or 8 mantissa digits could produce slightly incorrect results under very specific conditions.
Support -MD without a filename, for gcccompatibility. -MF can be used to set the dependencies output filename. See section 2.1.8.
Fix -E in combination with -MD. Seesection 2.1.22.
Fix missing errors on redefined labels; would cause convergence failure instead which is very slow and not easy to debug.
Duplicate definitions of the same label with the same value is now explicitly permitted (2.14 would allow it in some circumstances.)
Add the option --no-line to ignore %linedirectives in the source. See section 2.1.34 and section 5.13.1.

C.2.14. Version 2.14

Changed -I option semantics by adding a trailing path separator unconditionally.
Fixed null dereference in corrupted invalid single line macros.
Fixed division by zero which may happen if source code is malformed.
Fixed out of bound access in processing of malformed segment override.
Fixed out of bound access in certain EQU parsing.
Fixed buffer underflow in float parsing.
Added SGX (Intel Software Guard Extensions) instructions.
Added +n syntax for multiple contiguous registers.
Fixed subsections_via_symbols for macho object format.
Added the --gprefix, --gpostfix,--lprefix, and --lpostfix command line options, to allow command line base symbol renaming. Seesection 2.1.28.
Allow label renaming to be specified by %pragma in addition to from the command line. See section 8.10.
Supported generic %pragma namespaces, outputand debug. See section 5.12.
Added the --pragma command line option to inject a%pragma directive. Seesection 2.1.29.
Added the --before command line option to accept preprocess statement before input. See section 2.1.30.
Added AVX512 VBMI2 (Additional Bit Manipulation), VNNI (Vector Neural Network),BITALG (Bit Algorithm), and GFNI (Galois Field New Instruction) instructions.
Added the STATIC directive for local symbols that should be renamed using global-symbol rules. Seesection 8.9.
Allow a symbol to be defined as EXTERN and then later overridden as GLOBAL or COMMON. Furthermore, a symbol declared EXTERN and then defined will be treated asGLOBAL. See section 8.5.
The GLOBAL directive no longer is required to precede the definition of the symbol.
Support private_extern as macho specific extension to the GLOBAL directive. Seesection 9.9.5.
Updated UD0 encoding to match with the specification
Added the --limit-X command line option to set execution limits. See section 2.1.32.
Updated the Codeview version number to be aligned withMASM.
Added the --keep-all command line option to preserve output files. See section 2.1.33.
Added the --include command line option, an alias to-P (section 2.1.19).
Added the --help command line option as an alias to-h (section 3.1).
Added -W, -D, and -Q suffix aliases for RET instructions so the operand sizes of these instructions can be encoded without using o16,o32 or o64.

C.2.15. Version 2.13.03

Added AVX and AVX512 VAES* and VPCLMULQDQinstructions.
Fixed missing dwarf record in x32 ELF output format.

C.2.16. Version 2.13.02

Fix false positive in testing of numeric overflows.
Fix generation of PEXTRW instruction.
Fix smartalign package which could trigger an error during optimization if the alignment code expanded too much due to optimization of the previous code.
Fix a case where negative value in TIMES directive causes panic instead of an error.
Always finalize .debug_abbrev section with a null indwarf output format.
Support debug flag in section attributes formacho output format. Seesection 9.9.1.
Support up to 16 characters in section names for machooutput format.
Fix missing update of global BITS setting ifSECTION directive specified a bit size using output format-specific extensions (e.g. USE32 for theobj output format.)
Fix the incorrect generation of VEX-encoded instruction when static mode decorators are specified on scalar instructions, losing the decorators as they require EVEX encoding.
Option -MW to quote dependency outputs according to Watcom Make conventions instead of POSIX Make conventions. Seesection 2.1.12.
The obj output format now contains embedded dependency file information, unless disabled with %pragma obj nodepend. Seesection 9.4.9.
Fix generation of dependency lists.
Fix a number of null pointer reference and memory allocation errors.
Always generate symbol-relative relocations for the macho64output format; at least some versions of the XCode/LLVM linker fails for section-relative relocations.

C.2.17. Version 2.13.01

Fix incorrect output for some types of FAR orSEG references in the obj output format, and possibly other 16-bit output formats.
Fix the address in the list file for an instruction containing aTIMES directive.
Fix error with TIMES used together with an instruction which can vary in size, e.g. JMP.
Fix breakage on some uses of the DZ pseudo-op.

C.2.18. Version 2.13

Support the official forms of the UD0 and UD1instructions.
Allow self-segment-relative expressions in immediates and displacements, even when combined with an external or otherwise out-of-segment special symbol, e.g.:
extern foo
mov eax,[foo - $ + ebx] ; Now legal
Handle a 64-bit origin in NDISASM.
NASM can now generate sparse output files for relevant output formats, if the underlying operating system supports them.
The macho object format now supports thesubsections_via_symbols and no_dead_stripdirectives, see section 9.9.4.
The macho object format now supports theno_dead_strip, live_support andstrip_static_syms section flags, seesection 9.9.1.
The macho object format now supports the dwarfdebugging format, as required by newer toolchains.
All warnings can now be suppressed if desired; warnings not otherwise part of any warning class are now considered its own warning class calledother (e.g. -w-other). Furthermore, warning-as-error can now be controlled on a per warning class basis, using the syntax -w+error=warning-class and its equivalent for all other warning control options. Seesection 2.1.26 for the command-line options and warning classes andsection 8.14 for the[WARNING] directive.
Fix a number of bugs related to AVX-512 decorators.
Significant improvements to building NASM with Microsoft Visual Studio via Mkfiles/msvc.mak. It is now possible to build the full Windows installer binary as long as the necessary prerequisites are installed; see Mkfiles/README
To build NASM with custom modifications (table changes) or from the git tree now requires Perl 5.8 at the very minimum, quite possibly a higher version (Perl 5.24.1 tested.) There is no requirement to have Perl on your system at all if all you want to do is build unmodified NASM from source archives.
Fix the {z} decorator on AVX-512 VMOVDQ*instructions.
Add new warnings for certain dangerous constructs which never ought to have been allowed. In particular, the RESx family of instructions should have been taking a critical expression all along.
Fix the EVEX (AVX-512) versions of the VPBROADCAST,VPEXTR, and VPINSR instructions.
Support contracted forms of additional instructions. As a general rule, if an instruction has a non-destructive source immediately after a destination register that isn't used as an input, NASM supports omitting that source register, using the destination register as that value. This among other things makes it easier to convert SSE code to the equivalent AVX code:
addps xmm1,xmm0 ; SSE instruction
vaddps ymm1,ymm1,ymm0 ; AVX official long form
vaddps ymm1,ymm0 ; AVX contracted form
Fix Codeview malformed compiler version record.
Add the CLWB and PCOMMIT instructions. Note that the PCOMMIT instruction has been deprecated and will never be included in a shipping product; it is included for completeness only.
Add the %pragma preprocessor directive for soft-error directives.
Add the RDPID instruction.

C.2.19. Version 2.12.02

Fix preprocessor errors, especially %error and%warning, inside %if statements.
Fix relative relocations in 32-bit Mach-O.
More Codeview debug format fixes.
If the MASM PTR keyword is encountered, issue a warning. This is much more likely to indicate a MASM-ism encountered in NASM than it is a valid label. This warning can be suppressed with -w-ptr, the [warning] directive (seesection 2.1.26) or by the macro definition %idefine ptr $%? (seesection 5.2.5).
When an error or a warning comes from the expansion of a multi-line macro, display the file and line numbers for the expanded macros. Macros defined with .nolist do not get displayed.
Add macros ilog2fw() and ilog2cw() to theifunc macro package. Seesection 7.4.1.

C.2.20. Version 2.12.01

Portability fixes for some platforms.
Fix error when not specifying a list file.
Correct the handling of macro-local labels in the Codeview debugging format.
Add CLZERO, MONITORX and MWAITXinstructions.

C.2.21. Version 2.12

Major fixes to the macho backend (section 9.9); earlier versions would produce invalid symbols and relocations on a regular basis.
Support for thread-local storage in Mach-O.
Support for arbitrary sections in Mach-O.
Fix wrong negative size treated as a big positive value passed into backend causing NASM to crash.
Fix handling of zero-extending unsigned relocations, we have been printing wrong message and forgot to assign segment with predefined value before passing it into output format.
Fix potential write of oversized (with size greater than allowed in output format) relative relocations.
Portability fixes for building NASM with the LLVM compiler.
Add support of Codeview version 8 (cv8) debug format forwin32 and win64 formats in the COFFbackend, see section 9.6.3.
Allow 64-bit outputs in 16/32-bit only backends. Unsigned 64-bit relocations are zero-extended from 32-bits with a warning (suppressible via-w-zext-reloc); signed 64-bit relocations are an error.
Line numbers in list files now correspond to the lines in the source files, instead of simply being sequential.
There is now an official 64-bit (x64 a.k.a. x86-64) build for Windows.

C.2.22. Version 2.11.09

Fix potential stack overwrite in macho32 backend.
Fix relocation records in macho64 backend.
Fix symbol lookup computation in macho64 backend.
Adjust .symtab and .rela.text sections alignments to 8 bytes in elf64 backed.
Fix section length computation in bin backend which leaded in incorrect relocation records.

C.2.23. Version 2.11.08

Fix section length computation in bin backend which leaded in incorrect relocation records.
Add a warning for numeric preprocessor definitions passed via command line which might have unexpected results otherwise.
Add ability to specify a module name record in rdoff linker with -mn option.
Increase label length capacity up to 256 bytes in rdoffbackend for FreePascal sake, which tends to generate very long labels for procedures.
Fix segmentation failure when rip addressing is used inmacho64 backend.
Fix access on out of memory when handling strings with a single grave. We have sixed similar problem in previous release but not all cases were covered.
Fix NULL dereference in disassembled on BND instruction.

C.2.24. Version 2.11.07

Fix 256 bit VMOVNTPS instruction.
Fix -MD option handling, which was rather broken in previous release changing command line api.
Fix access to uninitialized space when handling strings with a single grave.
Fix nil dereference in handling memory reference parsing.

C.2.25. Version 2.11.06

Update AVX512 instructions based on the Extension Reference (319433-021 Sept 2014).
Fix the behavior of -MF and -MD options (Bugzilla 3392280)
Updated Win32 Makefile to fix issue with build

C.2.26. Version 2.11.05

Add --v as an alias for -v (seesection 2.1.27), for command-line compatibility with Yasm.
Fix a bug introduced in 2.11.03 whereby certain instructions would contain multiple REX prefixes, and thus be corrupt.

C.2.27. Version 2.11.04

Removed an invalid error checking code. Sometimes a memref only with a displacement can also set an evex flag. For example:
vmovdqu32 [0xabcd]{k1}, zmm0
Fixed a bug in disassembler that EVEX.L'L vector length was not matched when EVEX.b was set because it was simply considered as EVEC.RC. Separated EVEX.L'L case from EVEX.RC which is ignored in matching.

C.2.28. Version 2.11.03

Fix a bug there REX prefixes were missing on instructions inside aTIMES statement.

C.2.29. Version 2.11.02

Add the XSAVEC, XSAVES andXRSTORS family instructions.
Add the CLFLUSHOPT instruction.

C.2.30. Version 2.11.01

Allow instructions which implicitly uses XMM0(VBLENDVPD, VBLENDVPS, PBLENDVB andSHA256RNDS2) to be specified without an explicitxmm0 on the assembly line. In other words, the following two lines produce the same output:
vblendvpd xmm2,xmm1,xmm0 ; Last operand is fixed xmm0
vblendvpd xmm2,xmm1 ; Implicit xmm0 omitted
In the ELF backends, don't crash the assembler ifsection align is specified without a value.

C.2.31. Version 2.11

Add support for the Intel AVX-512 instruction set:
16 new, 512-bit SIMD registers. Total 32 (ZMM0 ~ ZMM31)
8 new opmask registers (K0 ~ K7). One of 7 registers(K1 ~ K7) can be used as an opmask for conditional execution.
A new EVEX encoding prefix. EVEX is based on VEX and provides more capabilities: opmasks, broadcasting, embedded rounding and compressed displacements.

opmask
VDIVPD zmm0{k1}{z}, zmm1, zmm3 ; conditional vector operation
; using opmask k1.
; {z} is for zero-masking
broadcasting
VDIVPS zmm4, zmm5, [rbx]{1to16} ; load single-precision float and
; replicate it 16 times. 32 * 16 = 512
embedded rounding
VCVTSI2SD xmm6, xmm7, {rz-sae}, rax ; round toward zero. note that it
; is used as if a separate operand.
; it comes after the last SIMD operand

Add support for ZWORD (512 bits), DZ andRESZ.
Add support for the MPX and SHA instruction sets.
Better handling of section redefinition.
Generate manpages when running 'make dist'.
Handle all token chains in mmacro params range.
Support split [base,index] effective address:
mov eax,[eax+8,ecx*4] ; eax=base, ecx=index, 4=scale, 8=disp

This is expected to be most useful for the MPX instructions.

Support BND prefix for branch instructions (for MPX).
The DEFAULT directive can now take BND andNOBND options to indicate whether all relevant branches should be getting BND prefixes. This is expected to be the normal for use in MPX code.
Add {evex}, {vex3} and {vex2}instruction prefixes to have NASM encode the corresponding instruction, if possible, with an EVEX, 3-byte VEX, or 2-byte VEX prefix, respectively.
Support for section names longer than 8 bytes in Win32/Win64 COFF.
The NOSPLIT directive by itself no longer forces a single register to become an index register, unless it has an explicit multiplier.
mov eax,[nosplit eax] ; eax as base register
mov eax,[nosplit eax*1] ; eax as index register

C.2.32. Version 2.10.09

Pregenerate man pages.

C.2.33. Version 2.10.08

Fix VMOVNTDQA, MOVNTDQA andMOVLPD instructions.
Fix collision for VGATHERQPS, VPGATHERQDinstructions.
Fix VPMOVSXBQ, VGATHERQPD, VSPLLWinstructions.
Add a bunch of AMD TBM instructions.
Fix potential stack overwrite in numbers conversion.
Allow byte size in PREFETCHTx instructions.
Make manual pages up to date.
Make F3 and F2 SSE prefixes to override66.
Support of AMD SVM instructions in 32 bit mode.
Fix near offsets code generation for JMP, CALLinstrictions in long mode.
Fix preprocessor parse regression when id is expanding to a whitespace.

C.2.34. Version 2.10.07

Fix line continuation parsing being broken in previous version.

C.2.35. Version 2.10.06

Always quote the dependency source names when using the automatic dependency generation options.
If no dependency target name is specified via the -MT or-MQ options, quote the default output name.
Fix assembly of shift operations in CPU 8086 mode.
Fix incorrect generation of explicit immediate byte for shift by 1 under certain circumstances.
Fix assembly of the VPCMPGTQ instruction.
Fix RIP-relative relocations in the macho64 backend.

C.2.36. Version 2.10.05

Add the CLAC and STAC instructions.

C.2.37. Version 2.10.04

Add back the inadvertently deleted 256-bit version of theVORPD instruction.
Correct disassembly of instructions starting with byte 82hex.
Fix corner cases in token pasting, for example:
%define N 1e%++%+ 5
dd N, 1e+5

C.2.38. Version 2.10.03

Correct the assembly of the instruction:
XRELEASE MOV [absolute],AL
Previous versions would incorrectly generate F3 A2 for this instruction and issue a warning; correct behavior is to emitF3 88 05.

C.2.39. Version 2.10.02

Add the ifunc macro package with integer functions, currently only integer logarithms. Seesection 7.4.
Add the RDSEED, ADCX and ADOXinstructions.

C.2.40. Version 2.10.01

Add missing VPMOVMSKB instruction with reg32, ymmreg operands.

C.2.41. Version 2.10

When optimization is enabled, mov r64,imm now optimizes to the shortest form possible between:
mov r32,imm32 ; 5 bytes
mov r64,imm32 ; 7 bytes
mov r64,imm64 ; 10 bytes
To force a specific form, use the STRICT keyword, seesection 3.7.
Add support for the Intel AVX2 instruction set.
Add support for Bit Manipulation Instructions 1 and 2.
Add support for Intel Transactional Synchronization Extensions (TSX).
Add support for x32 ELF (32-bit ELF with the CPU in 64-bit mode.) Seesection 9.10.
Add support for bigendian UTF-16 and UTF-32. Seesection 3.4.5.

C.2.42. Version 2.09.10

Fix up NSIS script to protect uninstaller against registry keys absence or corruption. It brings in a few additional questions to a user during deinstallation procedure but still it is better than unpredictable file removal.

C.2.43. Version 2.09.09

Fix initialization of section attributes of bin output format.
Fix mach64 output format bug that crashes NASM due to NULL symbols.

C.2.44. Version 2.09.08

Fix __OUTPUT_FORMAT__ assignment when output driver alias is used. For example when -f elf is used__OUTPUT_FORMAT__ must be set to elf, if-f elf32 is used __OUTPUT_FORMAT__ must be assigned accordingly, i.e. to elf32. The rule applies to all output driver aliases. See section 6.5.

C.2.45. Version 2.09.07

Fix attempts to close same file several times when -aoption is used.
Fixes for VEXTRACTF128, VMASKMOVPS encoding.

C.2.46. Version 2.09.06

Fix missed section attribute initialization in bin output target.

C.2.47. Version 2.09.05

Fix arguments encoding for VPEXTRW instruction.
Remove invalid form of VPEXTRW instruction.
Add VLDDQU as alias for VLDQQU to match specification.

C.2.48. Version 2.09.04

Fix incorrect labels offset for VEX instructions.
Eliminate bogus warning on implicit operand size override.
%if term could not handle 64 bit numbers.
The COFF backend was limiting relocations number to 16 bits even if in real there were a way more relocations.

C.2.49. Version 2.09.03

Print %macro name inside %rep blocks on error.
Fix preprocessor expansion behaviour. It happened sometime too early and sometime simply wrong. Move behaviour back to the origins (down to NASM 2.05.01).
Fix uninitialized data dereference on OMF output format.
Issue warning on unterminated %{ construct.
Fix for documentation typo.

C.2.50. Version 2.09.02

Fix reversed tokens when %deftok produces more than one output token.
Fix segmentation fault on disassembling some VEX instructions.
Missing %endif did not always cause error.
Fix typo in documentation.
Compound context local preprocessor single line macro identifiers were not expanded early enough and as result lead to unresolved symbols.

C.2.51. Version 2.09.01

Fix NULL dereference on missed %deftok second parameter.
Fix NULL dereference on invalid %substr parameters.

C.2.52. Version 2.09

Fixed assignment the magnitude of %rep counter. It is limited to 62 bits now.
Fixed NULL dereference if argument of %strlen resolves to whitespace. For example if nonexistent macro parameter is used.
%ifenv, %elifenv, %ifnenv, and%elifnenv directives introduced. Seesection 5.6.13.
Fixed NULL dereference if environment variable is missed.
Updates of new AVX v7 Intel instructions.
PUSH imm32 is now officially documented.
Fix for encoding the LFS, LGS and LSS in 64-bit mode.
Fixes for compatibility with OpenWatcom compiler and DOS 8.3 file format limitation.
Macros parameters range expansion introduced. Seesection 5.5.4.
Backward compatibility on expanging of local single macros restored.
8 bit relocations for elf and bin output formats are introduced.
Short intersegment jumps are permitted now.
An alignment more than 64 bytes are allowed for win32,win64 output formats.
SECTALIGN directive introduced. Seesection 6.12.2.
nojmp option introduced in smartalign package. See section 7.2.
Short aliases win, elf and machofor output formats are introduced. Each stands for win32,elf32 and macho32 accordingly.
Faster handling of missing directives implemented.
Various small improvements in documentation.
No hang anymore if unable to open malloc.log file.
The environments without vsnprintf function are able to build nasm again.
AMD LWP instructions updated.
Tighten EA checks. We warn a user if there overflow in EA addressing.
Make -Ox the default optimization level. For the legacy behavior, specify -O0 explicitly. Seesection 2.1.24.
Environment variables read with %! or tested with%ifenv can now contain non-identifier characters if surrounded by quotes. See section 5.13.2.
Add a new standard macro package %use fp for floating-point convenience macros. See section 7.3.

C.2.53. Version 2.08.02

Fix crash under certain circumstances when using the %+operator.

C.2.54. Version 2.08.01

Fix the %use statement, which was broken in 2.08.

C.2.55. Version 2.08

A number of enhancements/fixes in macros area.
Support for converting strings to tokens. Seesection 5.2.10.
Fuzzy operand size logic introduced.
Fix COFF stack overrun on too long export identifiers.
Fix Macho-O alignment bug.
Fix crashes with –fwin32 on file with many exports.
Fix stack overrun for too long [DEBUG id].
Fix incorrect sbyte usage in IMUL (hit only if optimization flag passed).
Append ending token for .stabs records in the ELF output format.
New NSIS script which uses ModernUI and MultiUser approach.
Visual Studio 2008 NASM integration (rules file).
Warn a user if a constant is too long (and as result will be stripped).
The obsoleted pre-XOP AMD SSE5 instruction set which was never actualized was removed.
Fix stack overrun on too long error file name passed from the command line.
Bind symbols to the .text section by default (ie in case if SECTION directive was omitted) in the ELF output format.
Fix sync points array index wrapping.
A few fixes for FMA4 and XOP instruction templates.
Add AMD Lightweight Profiling (LWP) instructions.
Fix the offset for %arg in 64-bit mode.
An undefined local macro (%$) no longer matches a global macro with the same name.
Fix NULL dereference on too long local labels.

C.2.56. Version 2.07

NASM is now under the 2-clause BSD license. Seesection 1.1.1.
Fix the section type for the .strtab section in theelf64 output format.
Fix the handling of COMMON directives in theobj output format.
New ith and srec output formats; these are variants of the bin output format which output Intel hex and Motorola S-records, respectively. Seesection 9.2 andsection 9.3.
rdf2ihx replaced with an enhanced rdf2bin, which can output binary, COM, Intel hex or Motorola S-records.
The Windows installer now puts the NASM directory first in thePATH of the "NASM Shell".
Revert the early expansion behavior of %+ to pre-2.06 behavior: %+ is only expanded late.
Yet another Mach-O alignment fix.
Don't delete the list file on errors. Also, include error and warning information in the list file.
Support for 64-bit Mach-O output, seesection 9.9.
Fix assert failure on certain operations that involve strings with high-bit bytes.

C.2.57. Version 2.06

This release is dedicated to the memory of Charles A. Crayne, long time NASM developer as well as moderator of comp.lang.asm.x86 and author of the book Serious Assembler. We miss you, Chuck.
Support for indirect macro expansion (%[...]). Seesection 5.2.3.
%pop can now take an argument, seesection 5.9.1.
The argument to %use is no longer macro-expanded. Use%[...] if macro expansion is desired.
Support for thread-local storage in ELF32 and ELF64. Seesection 9.10.4.
Fix crash on %ifmacro without an argument.
Correct the arguments to the POPCNT instruction.
Fix section alignment in the Mach-O format.
Update AVX support to version 5 of the Intel specification.
Fix the handling of accesses to context-local macros from higher levels in the context stack.
Treat WAIT as a prefix rather than as an instruction, thereby allowing constructs like O16 FSAVE to work correctly.
Support for structures with a non-zero base offset. Seesection 6.11.1.
Correctly handle preprocessor token concatenation (seesection 5.5.9) involving floating-point numbers.
The PINSR series of instructions have been corrected and rationalized.
Removed AMD SSE5, replaced with the new XOP/FMA4/CVT16 (rev 3.03) spec.
The ELF backends no longer automatically generate a.comment section.
Add additional "well-known" ELF sections with default attributes. Seesection 9.10.2.

C.2.58. Version 2.05.01

Fix the -w/-W option parsing, which was broken in NASM 2.05.

C.2.59. Version 2.05

Fix redundant REX.W prefix on JMP reg64.
Make the behaviour of -O0 match NASM 0.98 legacy behavior. See section 2.1.24.
-w-user can be used to suppress the output of%warning directives. Seesection 2.1.26.
Fix bug where ALIGN would issue a full alignment datum instead of zero bytes.
Fix offsets in list files.
Fix %include inside multi-line macros or loops.
Fix error where NASM would generate a spurious warning on valid optimizations of immediate values.
Fix arguments to a number of the CVT SSE instructions.
Fix RIP-relative offsets when the instruction carries an immediate.
Massive overhaul of the ELF64 backend for spec compliance.
Fix the Geode PFRCPV and PFRSQRTV instruction.
Fix the SSE 4.2 CRC32 instruction.

C.2.60. Version 2.04

Sanitize macro handing in the %error directive.
New %warning directive to issue user-controlled warnings.
%error directives are now deferred to the final assembly phase.
New %fatal directive to immediately terminate assembly.
New %strcat directive to join quoted strings together.
New %use macro directive to support standard macro directives. See section 5.8.4.
Excess default parameters to %macro now issues a warning by default. See section 5.5.
Fix %ifn and %elifn.
Fix nested %else clauses.
Correct the handling of nested %reps.
New %unmacro directive to undeclare a multi-line macro. Seesection 5.5.12.
Builtin macro __PASS__ which expands to the current assembly pass. See section 6.10.
__utf16__ and __utf32__ operators to generate UTF-16 and UTF-32 strings. See section 3.4.5.
Fix bug in case-insensitive matching when compiled on platforms that don't use the configure script. Of the official release binaries, that only affected the OS/2 binary.
Support for x87 packed BCD constants. Seesection 3.4.7.
Correct the LTR and SLDT instructions in 64-bit mode.
Fix unnecessary REX.W prefix on indirect jumps in 64-bit mode.
Add AVX versions of the AES instructions (VAES...).
Fix the 256-bit FMA instructions.
Add 256-bit AVX stores per the latest AVX spec.
VIA XCRYPT instructions can now be written either with or withoutREP, apparently different versions of the VIA spec wrote them differently.
Add missing 64-bit MOVNTI instruction.
Fix the operand size of VMREAD and VMWRITE.
Numerous bug fixes, especially to the AES, AVX and VTX instructions.
The optimizer now always runs until it converges. It also runs even when disabled, but doesn't optimize. This allows most forward references to be resolved properly.
%push no longer needs a context identifier; omitting the context identifier results in an anonymous context.

C.2.61. Version 2.03.01

Fix buffer overflow in the listing module.
Fix the handling of hexadecimal escape codes in `...` strings.
The Postscript/PDF documentation has been reformatted.
The -F option now implies -g.

C.2.62. Version 2.03

Add support for Intel AVX, CLMUL and FMA instructions, including YMM registers.
dy, resy and yword for 32-byte operands.
Fix some SSE5 instructions.
Intel INVEPT, INVVPID and MOVBEinstructions.
Fix checking for critical expressions when the optimizer is enabled.
Support the DWARF debugging format for ELF targets.
Fix optimizations of signed bytes.
Fix operation on bigendian machines.
Fix buffer overflow in the preprocessor.
SAFESEH support for Win32, IMAGEREL for Win64 (SEH).
%? and %?? to refer to the name of a macro itself. In particular, %idefine keyword $%? can be used to make a keyword "disappear".
New options for dependency generation: -MD,-MF, -MP, -MT, -MQ.
New preprocessor directives %pathsearch and%depend; INCBIN reimplemented as a macro.
%include now resolves macros in a sane manner.
%substr can now be used to get other than one-character substrings.
New type of character/string constants, using backquotes (`...`), which support C-style escape sequences.
%defstr and %idefstr to stringize macro definitions before creation.
Fix forward references used in EQU statements.

C.2.63. Version 2.02

Additional fixes for MMX operands with explicit qword, as well as (hopefully) SSE operands with oword.
Fix handling of truncated strings with DO.
Fix segfaults due to memory overwrites when floating-point constants were used.
Fix segfaults due to missing include files.
Fix OpenWatcom Makefiles for DOS and OS/2.
Add autogenerated instruction list back into the documentation.
ELF: Fix segfault when generating stabs, and no symbols have been defined.
ELF: Experimental support for DWARF debugging information.
New compile date and time standard macros.
%ifnum now returns true for negative numbers.
New %iftoken test for a single token.
New %ifempty test for empty expansion.
Add support for the XSAVE instruction group.
Makefile for Netware/gcc.
Fix issue with some warnings getting emitted way too many times.
Autogenerated instruction list added to the documentation.

C.2.64. Version 2.01

Fix the handling of MMX registers with explicit qword tags on memory (broken in 2.00 due to 64-bit changes.)
Fix the PREFETCH instructions.
Fix the documentation.
Fix debugging info when using -f elf (backwards compatibility alias for -f elf32).
Man pages for rdoff tools (from the Debian project.)
ELF: handle large numbers of sections.
Fix corrupt output when the optimizer runs out of passes.

C.2.65. Version 2.00

Added c99 data-type compliance.
Added general x86-64 support.
Added win64 (x86-64 COFF) output format.
Added __BITS__ standard macro.
Renamed the elf output format to elf32 for clarity.
Added elf64 and macho (MacOS X) output formats.
Added Numeric constants in dq directive.
Added oword, do and reso pseudo operands.
Allow underscores in numbers.
Added 8-, 16- and 128-bit floating-point formats.
Added binary, octal and hexadecimal floating-point.
Correct the generation of floating-point constants.
Added floating-point option control.
Added Infinity and NaN floating point support.
Added ELF Symbol Visibility support.
Added setting OSABI value in ELF header directive.
Added Generate Makefile Dependencies option.
Added Unlimited Optimization Passes option.
Added %IFN and %ELIFN support.
Added Logical Negation Operator.
Enhanced Stack Relative Preprocessor Directives.
Enhanced ELF Debug Formats.
Enhanced Send Errors to a File option.
Added SSSE3, SSE4.1, SSE4.2, SSE5 support.
Added a large number of additional instructions.
Significant performance improvements.
-w+warning and -w-warning can now be written as –Wwarning and –Wno-warning, respectively. Seesection 2.1.26.
Add -w+error to treat warnings as errors. Seesection 2.1.26.
Add -w+all and -w-all to enable or disable all suppressible warnings. See section 2.1.26.

C.3. NASM 0.98 Series

The 0.98 series was the production versions of NASM from 1999 to 2007.

C.3.1. Version 0.98.39

fix buffer overflow
fix outas86's .bss handling
"make spotless" no longer deletes config.h.in.
%(el)if(n)idn insensitivity to string quotes difference (#809300).
(nasm.c)__OUTPUT_FORMAT__ changed to string value instead of symbol.

C.3.2. Version 0.98.38

Add Makefile for 16-bit DOS binaries under OpenWatcom, and modifymkdep.pl to be able to generate completely pathless dependencies, as required by OpenWatcom wmake (it supports path searches, but not explicit paths.)
Fix the STR instruction.
Fix the ELF output format, which was broken under certain circumstances due to the addition of stabs support.
Quick-fix Borland format debug-info for -f obj
Fix for %rep with no arguments (#560568)
Fix concatenation of preprocessor function call (#794686)
Fix long label causes coredump (#677841)
Use autoheader as well as autoconf to keep configure from generating ridiculously long command lines.
Make sure that all of the formats which support debugging output actually will suppress debugging output when -g not specified.

C.3.3. Version 0.98.37

Paths given in -I switch searched forincbin–ed as well as %include–ed files.
Added stabs debugging for the ELF output format, patch from Martin Wawro.
Fix output/outbin.c to allow origin > 80000000h.
Make -U switch work.
Fix the use of relative offsets with explicit prefixes, e.g.a32 loop foo.
Remove backslash().
Fix the SMSW and SLDT instructions.
-O2 and -O3 are no longer aliases for-O10 and -O15. If you mean the latter, please say so! :)

C.3.4. Version 0.98.36

Update rdoff – librarian/archiver – common rec – docs!
Fix signed/unsigned problems.
Fix JMP FAR label and CALL FAR label.
Add new multisection support – map files – fix align bug
Fix sysexit, movhps/movlps reg,reg bugs in insns.dat
Q or O suffixes indicate octal
Support Prescott new instructions (PNI).
Cyrix XSTORE instruction.

C.3.5. Version 0.98.35

Fix build failure on 16-bit DOS (Makefile.bc3 workaround for compiler bug.)
Fix dependencies and compiler warnings.
Add "const" in a number of places.
Add –X option to specify error reporting format (use –Xvc to integrate with Microsoft Visual Studio.)
Minor changes for code legibility.
Drop use of tmpnam() in rdoff (security fix.)

C.3.6. Version 0.98.34

Correct additional address-size vs. operand-size confusions.
Generate dependencies for all Makefiles automatically.
Add support for unimplemented (but theoretically available) registers such as tr0 and cr5. Segment registers 6 and 7 are called segr6 and segr7 for the operations which they can be represented.
Correct some disassembler bugs related to redundant address-size prefixes. Some work still remains in this area.
Correctly generate an error for things like "SEG eax".
Add the JMPE instruction, enabled by "CPU IA64".
Correct compilation on newer gcc/glibc platforms.
Issue an error on things like "jmp far eax".

C.3.7. Version 0.98.33

New __NASM_PATCHLEVEL__ and __NASM_VERSION_ID__ standard macros to round out the version-query macros. version.pl now understands X.YYplWW or X.YY.ZZplWW as a version number, equivalent to X.YY.ZZ.WW (or X.YY.0.WW, as appropriate).
New keyword "strict" to disable the optimization of specific operands.
Fix the handing of size overrides with JMP instructions (instructions such as "jmp dword foo".)
Fix the handling of "ABSOLUTE label", where "label" points into a relocatable segment.
Fix OBJ output format with lots of externs.
More documentation updates.
Add –Ov option to get verbose information about optimizations.
Undo a braindead change which broke %elif directives.
Makefile updates.

C.3.8. Version 0.98.32

Fix NASM crashing when %macro directives were left unterminated.
Lots of documentation updates.
Complete rewrite of the PostScript/PDF documentation generator.
The MS Visual C++ Makefile was updated and corrected.
Recognize .rodata as a standard section name in ELF.
Fix some obsolete Perl4-isms in Perl scripts.
Fix configure.in to work with autoconf 2.5x.
Fix a couple of "make cleaner" misses.
Make the normal "./configure && make" work with Cygwin.

C.3.9. Version 0.98.31

Correctly build in a separate object directory again.
Derive all references to the version number from the version file.
New standard macros __NASM_SUBMINOR__ and __NASM_VER__ macros.
Lots of Makefile updates and bug fixes.
New %ifmacro directive to test for multiline macros.
Documentation updates.
Fixes for 16-bit OBJ format output.
Changed the NASM environment variable to NASMENV.

C.3.10. Version 0.98.30

Changed doc files a lot: completely removed old READMExx and Wishlist files, incorporating all information in CHANGES and TODO.
I waited a long time to rename zoutieee.c to (original) outieee.c
moved all output modules to output/ subdirectory.
Added 'make strip' target to strip debug info from nasm & ndisasm.
Added INSTALL file with installation instructions.
Added –v option description to nasm man.
Added dist makefile target to produce source distributions.
16-bit support for ELF output format (GNU extension, but useful.)

C.3.11. Version 0.98.28

Fastcooked this for Debian's Woody release: Frank applied the INCBIN bug patch to 0.98.25alt and called it 0.98.28 to not confuse poor little apt-get.

C.3.12. Version 0.98.26

Reorganised files even better from 0.98.25alt

C.3.13. Version 0.98.25alt

Prettified the source tree. Moved files to more reasonable places.
Added findleak.pl script to misc/ directory.
Attempted to fix doc.

C.3.14. Version 0.98.25

Line continuation character \.
Docs inadvertently reverted – "dos packaging".

C.3.15. Version 0.98.24p1

FIXME: Someone, document this please.

C.3.16. Version 0.98.24

Documentation – Ndisasm doc added to Nasm.doc.

C.3.17. Version 0.98.23

Attempted to remove rdoff version1
Lino Mastrodomenico's patches to preproc.c (%$$ bug?).

C.3.18. Version 0.98.22

Update rdoff2 – attempt to remove v1.

C.3.19. Version 0.98.21

Optimization fixes.

C.3.20. Version 0.98.20

Optimization fixes.

C.3.21. Version 0.98.19

H. J. Lu's patch back out.

C.3.22. Version 0.98.18

Added ".rdata" to "-f win32".

C.3.23. Version 0.98.17

H. J. Lu's "bogus elf" patch. (Red Hat problem?)

C.3.24. Version 0.98.16

Fix whitespace before "[section ..." bug.

C.3.25. Version 0.98.15

Rdoff changes (?).
Fix fixes to memory leaks.

C.3.26. Version 0.98.14

Fix memory leaks.

C.3.27. Version 0.98.13

There was no 0.98.13

C.3.28. Version 0.98.12

Update optimization (new function of "-O1")
Changes to test/bintest.asm (?).

C.3.29. Version 0.98.11

Optimization changes.
Ndisasm fixed.

C.3.30. Version 0.98.10

There was no 0.98.10

C.3.31. Version 0.98.09

Add multiple sections support to "-f bin".
Changed GLOBAL_TEMP_BASE in outelf.c from 6 to 15.
Add "-v" as an alias to the "-r" switch.
Remove "#ifdef" from Tasm compatibility options.
Remove redundant size-overrides on "mov ds, ex", etc.
Fixes to SSE2, other insns.dat (?).
Enable uppercase "I" and "P" switches.
Case insinsitive "seg" and "wrt".
Update install.sh (?).
Allocate tokens in blocks.
Improve "invalid effective address" messages.

C.3.32. Version 0.98.08

Add "%strlen" and "%substr" macro operators
Fixed broken c16.mac.
Unterminated string error reported.
Fixed bugs as per 0.98bf

C.3.33. Version 0.98.09b with John Coffman patches released 28-Oct-2001

Changes from 0.98.07 release to 98.09b as of 28-Oct-2001

More closely compatible with 0.98 when –O0 is implied or specified. Not strictly identical, since backward branches in range of short offsets are recognized, and signed byte values with no explicit size specification will be assembled as a single byte.
More forgiving with the PUSH instruction. 0.98 requires a size to be specified always. 0.98.09b will imply the size from the current BITS setting (16 or 32).
Changed definition of the optimization flag:
-O0 strict two-pass assembly, JMP and Jcc are
handled more like 0.98, except that back-
ward JMPs are short, if possible.
-O1 strict two-pass assembly, but forward
branches are assembled with code guaranteed
to reach; may produce larger code than
-O0, but will produce successful assembly
more often if branch offset sizes are not
specified.
-O2 multi-pass optimization, minimize branch
offsets; also will minimize signed immed-
iate bytes, overriding size specification.
-O3 like -O2, but more passes taken, if needed

C.3.34. Version 0.98.07 released 01/28/01

Added Stepane Denis' SSE2 instructions to a *working* version of the code – some earlier versions were based on broken code – sorry 'bout that. version "0.98.07"
Cosmetic modifications to nasm.c, nasm.h, AUTHORS, MODIFIED

C.3.35. Version 0.98.06f released 01/18/01

Add "metalbrain"s jecxz bug fix in insns.dat
Alter nasmdoc.src to match – version "0.98.06f"

C.3.36. Version 0.98.06e released 01/09/01

Removed the "outforms.h" file – it appears to be someone's old backup of "outform.h". version "0.98.06e"
fbk – finally added the fix for the "multiple %includes bug", known since 7/27/99 – reported originally (?) and sent to us by Austin Lunnen – he reports that John Fine had a fix within the day. Here it is...
Nelson Rush resigns from the group. Big thanks to Nelson for his leadership and enthusiasm in getting these changes incorporated into Nasm!
fbk – [list +], [list –] directives – ineptly implemented, should be re-written or removed, perhaps.
Brian Raiter / fbk – "elfso bug" fix – applied to aoutb format as well – testing might be desirable...
James Seter – –postfix, –prefix command line switches.
Yuri Zaporozhets – rdoff utility changes.

C.3.37. Version 0.98p1

GAS-like palign (Panos Minos)
FIXME: Someone, fill this in with details

C.3.38. Version 0.98bf (bug-fixed)

Fixed – elf and aoutb bug – shared libraries – multiple "%include" bug in "-f obj" – jcxz, jecxz bug – unrecognized option bug in ndisasm

C.3.39. Version 0.98.03 with John Coffman's changes released 27-Jul-2000

Added signed byte optimizations for the 0x81/0x83 class of instructions: ADC, ADD, AND, CMP, OR, SBB, SUB, XOR: when used as 'ADD reg16,imm' or 'ADD reg32,imm.' Also optimization of signed byte form of 'PUSH imm' and 'IMUL reg,imm'/'IMUL reg,reg,imm.' No size specification is needed.
Added multi-pass JMP and Jcc offset optimization. Offsets on forward references will preferentially use the short form, without the need to code a specific size (short or near) for the branch. Added instructions for 'Jcc label' to use the form 'Jnotcc $+3/JMP label', in cases where a short offset is out of bounds. If compiling for a 386 or higher CPU, then the 386 form of Jcc will be used instead.
This feature is controlled by a new command-line switch: "O", (upper case letter O). "-O0" reverts the assembler to no extra optimization passes, "-O1" allows up to 5 extra passes, and "-O2"(default), allows up to 10 extra optimization passes.
Added a new directive: 'cpu XXX', where XXX is any of: 8086, 186, 286, 386, 486, 586, pentium, 686, PPro, P2, P3 or Katmai. All are case insensitive. All instructions will be selected only if they apply to the selected cpu or lower. Corrected a couple of bugs in cpu-dependence in 'insns.dat'.
Added to 'standard.mac', the "use16" and "use32" forms of the "bits 16/32" directive. This is nothing new, just conforms to a lot of other assemblers. (minor)
Changed label allocation from 320/32 (10000 labels @ 200K+) to 32/37 (1000 labels); makes running under DOS much easier. Since additional label space is allocated dynamically, this should have no effect on large programs with lots of labels. The 37 is a prime, believed to be better for hashing. (minor)

C.3.40. Version 0.98.03

"Integrated patchfile 0.98-0.98.01. I call this version 0.98.03 for historical reasons: 0.98.02 was trashed." ––John Coffman johninsd@san.rr.com, 27-Jul-2000

Kendall Bennett's SciTech MGL changes
Note that you must define "TASM_COMPAT" at compile-time to get the Tasm Ideal Mode compatibility.
All changes can be compiled in and out using the TASM_COMPAT macros, and when compiled without TASM_COMPAT defined we get the exact same binary as the unmodified 0.98 sources.
standard.mac, macros.c: Added macros to ignore TASM directives before first include
nasm.h: Added extern declaration for tasm_compatible_mode
nasm.c: Added global variable tasm_compatible_mode
Added command line switch for TASM compatible mode (-t)
Changed version command line to reflect when compiled with TASM additions
Added response file processing to allow all arguments on a single line (response file is @resp rather than –@resp for NASM format).
labels.c: Changes islocal() macro to support TASM style @@local labels.
Added islocalchar() macro to support TASM style @@local labels.
parser.c: Added support for TASM style memory references (ie: mov [DWORD eax],10 rather than the NASM style mov DWORD [eax],10).
preproc.c: Added new directives, %arg, %local,%stacksize to directives table
Added support for TASM style directives without a leading % symbol.
Integrated a block of changes from Andrew Zabolotny bit@eltech.ru:
A new keyword %xdefine and its case-insensitive counterpart%ixdefine. They work almost the same way as%define and %idefine but expand the definition immediately, not on the invocation. Something like a cross between%define and %assign. The "x" suffix stands for "eXpand", so "xdefine" can be deciphered as "expand-and-define". Thus you can do things like this:
%assign ofs 0
%macro arg 1
%xdefine %1 dword [esp+ofs]
%assign ofs ofs+4
%endmacro
Changed the place where the expansion of %$name macros are expanded. Now they are converted into ..@ctxnum.name form when detokenizing, so there are no quirks as before when using %$name arguments to macros, in macros etc. For example:
%macro abc 1
%define %1 hello
%endm
abc %$here
%$here
Now last line will be expanded into "hello" as expected. This also allows for lots of goodies, a good example are extended "proc" macros included in this archive.
Added a check for "cstk" in smacro_defined() before calling get_ctx() – this allows for things like:
%ifdef %$abc
%endif
to work without warnings even in no context.
Added a check for "cstk" in %if*ctx and %elif*ctx directives – this allows to use %ifctx without excessive warnings. If there is no active context, %ifctx goes through "false" branch.
Removed "user error: " prefix with %error directive: it just clobbers the output and has absolutely no functionality. Besides, this allows to write macros that does not differ from built-in functions in any way.
Added expansion of string that is output by %errordirective. Now you can do things like:
%define hello(x) Hello, x!
%define %$name andy
%error "hello(%$name)"
Same happened with %include directive.
Now all directives that expect an identifier will try to expand and concatenate everything without whitespaces in between before usage. For example, with "unfixed" nasm the commands
%define %$abc hello
%define __%$abc goodbye
__%$abc
would produce "incorrect" output: last line will expand to
hello goodbyehello
Not quite what you expected, eh? :-) The answer is that preprocessor treats the %define construct as if it would be
%define __ %$abc goodbye
(note the white space between __ and %$abc). After my "fix" it will "correctly" expand into
goodbye
as expected. Note that I use quotes around words "correct", "incorrect" etc because this is rather a feature not a bug; however current behaviour is more logical (and allows more advanced macro usage :-).

Same change was applied to:%push,%macro,%imacro,%define,%idefine,%xdefine,%ixdefine,%assign,%iassign,%undef

A new directive [WARNING {+|-}warning-id] have been added. It works only if the assembly phase is enabled (i.e. it doesn't work with nasm –e).
A new warning type: macro-selfref. By default this warning is disabled; when enabled NASM warns when a macro self-references itself; for example the following source:
[WARNING macro-selfref]
%macro push 1-*
%rep %0
push %1
%rotate 1
%endrep
%endmacro
push eax,ebx,ecx

will produce a warning, but if we remove the first line we won't see it anymore (which is The Right Thing To Do {tm} IMHO since C preprocessor eats such constructs without warnings at all).

Added a "error" routine to preprocessor which always will set ERR_PASS1 bit in severity_code. This removes annoying repeated errors on first and second passes from preprocessor.
Added the %+ operator in single-line macros for concatenating two identifiers. Usage example:
%define _myfunc _otherfunc
%define cextern(x) _ %+ x
cextern (myfunc)

After first expansion, third line will become "_myfunc". After this expansion is performed again so it becomes "_otherunc".

Now if preprocessor is in a non-emitting state, no warning or error will be emitted. Example:
%if 1
mov eax,ebx
%else
put anything you want between these two brackets,
even macro-parameter references %1 or local
labels %$zz or macro-local labels %%zz - no
warning will be emitted.
%endif
Context-local variables on expansion as a last resort are looked up in outer contexts. For example, the following piece:
%push outer
%define %$a [esp]
%push inner
%$a
%pop
%pop

will expand correctly the fourth line to [esp]; if we'll define another %$a inside the "inner" context, it will take precedence over outer definition. However, this modification has been applied only to expand_smacro and not to smacro_define: as a consequence expansion looks in outer contexts, but %ifdef won't look in outer contexts.
This behaviour is needed because we don't want nested contexts to act on already defined local macros. Example:
%define %$arg1 [esp+4]
test eax,eax
if nz
mov eax,%$arg1
endif
In this example the "if" mmacro enters into the "if" context, so %$arg1 is not valid anymore inside "if". Of course it could be worked around by using explicitly %$$arg1 but this is ugly IMHO.

Fixed memory leak in %undef. The origline wasn't freed before exiting on success.
Fixed trap in preprocessor when line expanded to empty set of tokens. This happens, for example, in the following case:
#define SOMETHING
SOMETHING

C.3.41. Version 0.98

All changes since NASM 0.98p3 have been produced by H. Peter Anvin hpa@zytor.com.

The documentation comment delimiter is
Allow EQU definitions to refer to external labels; reported by Pedro Gimeno.
Re-enable support for RDOFF v1; reported by Pedro Gimeno.
Updated License file per OK from Simon and Julian.

C.3.42. Version 0.98p9

Update documentation (although the instruction set reference will have to wait; I don't want to hold up the 0.98 release for it.)
Verified that the NASM implementation of the PEXTRW and PMOVMSKB instructions is correct. The encoding differs from what the Intel manuals document, but the Pentium III behaviour matches NASM, not the Intel manuals.
Fix handling of implicit sizes in PSHUFW and PINSRW, reported by Stefan Hoffmeister.
Resurrect the –s option, which was removed when changing the diagnostic output to stdout.

C.3.43. Version 0.98p8

Fix for "DB" when NASM is running on a bigendian machine.
Invoke insns.pl once for each output script, making Makefile.in legal for "make –j".
Improve the Unix configure-based makefiles to make package creation easier.
Included an RPM .spec file for building RPM (RedHat Package Manager) packages on Linux or Unix systems.
Fix Makefile dependency problems.
Change src/rdsrc.pl to include sectioning information in info output; required for install-info to work.
Updated the RDOFF distribution to version 2 from Jules; minor massaging to make it compile in my environment.
Split doc files that can be built by anyone with a Perl interpreter off into a separate archive.
"Dress rehearsal" release!

C.3.44. Version 0.98p7

Fixed opcodes with a third byte-sized immediate argument to not complain if given "byte" on the immediate.
Allow %undef to remove single-line macros with arguments. This matches the behaviour of #undef in the C preprocessor.
Allow –d, –u, –i and –p to be specified as –D, –U, –I and –P for compatibility with most C compilers and preprocessors. This allows Makefile options to be shared between cc and nasm, for example.
Minor cleanups.
Went through the list of Katmai instructions and hopefully fixed the (rather few) mistakes in it.
(Hopefully) fixed a number of disassembler bugs related to ambiguous instructions (disambiguated by –p) and SSE instructions with REP.
Fix for bug reported by Mark Junger: "call dword 0x12345678" should work and may add an OSP (affected CALL, JMP, Jcc).
Fix for environments when "stderr" isn't a compile-time constant.

C.3.45. Version 0.98p6

Took officially over coordination of the 0.98 release; so drop the p3.x notation. Skipped p4 and p5 to avoid confusion with John Fine's J4 and J5 releases.
Update the documentation; however, it still doesn't include documentation for the various new instructions. I somehow wonder if it makes sense to have an instruction set reference in the assembler manual when Intel et al have PDF versions of their manuals online.
Recognize "idt" or "centaur" for the –p option to ndisasm.
Changed error messages back to stderr where they belong, but add an –E option to redirect them elsewhere (the DOS shell cannot redirect stderr.)
–M option to generate Makefile dependencies (based on code from Alex Verstak.)
%undef preprocessor directive, and –u option, that undefines a single-line macro.
OS/2 Makefile (Mkfiles/Makefile.os2) for Borland under OS/2; from Chuck Crayne.
Various minor bugfixes (reported by): – Dangling %sin preproc.c (Martin Junker)
THERE ARE KNOWN BUGS IN SSE AND THE OTHER KATMAI INSTRUCTIONS. I am on a trip and didn't bring the Katmai instruction reference, so I can't work on them right now.
Updated the License file per agreement with Simon and Jules to include a GPL distribution clause.

C.3.46. Version 0.98p3.7

(Hopefully) fixed the canned Makefiles to include the outrdf2 and zoutieee modules.
Renamed changes.asm to changed.asm.

C.3.47. Version 0.98p3.6

Fixed a bunch of instructions that were added in 0.98p3.5 which had memory operands, and the address-size prefix was missing from the instruction pattern.

C.3.48. Version 0.98p3.5

Merged in changes from John S. Fine's 0.98-J5 release. John's based 0.98-J5 on my 0.98p3.3 release; this merges the changes.
Expanded the instructions flag field to a long so we can fit more flags; mark SSE (KNI) and AMD or Katmai-specific instructions as such.
Fix the "PRIV" flag on a bunch of instructions, and create new "PROT" flag for protected-mode-only instructions (orthogonal to if the instruction is privileged!) and new "SMM" flag for SMM-only instructions.
Added AMD-only SYSCALL and SYSRET instructions.
Make SSE actually work, and add new Katmai MMX instructions.
Added a –p (preferred vendor) option to ndisasm so that it can distinguish e.g. Cyrix opcodes also used in SSE. For example:
ndisasm -p cyrix aliased.bin
00000000 670F514310 paddsiw mm0,[ebx+0x10]
00000005 670F514320 paddsiw mm0,[ebx+0x20]
ndisasm -p intel aliased.bin
00000000 670F514310 sqrtps xmm0,[ebx+0x10]
00000005 670F514320 sqrtps xmm0,[ebx+0x20]
Added a bunch of Cyrix-specific instructions.

C.3.49. Version 0.98p3.4

Made at least an attempt to modify all the additional Makefiles (in the Mkfiles directory). I can't test it, but this was the best I could do.
DOS DJGPP+"Opus Make" Makefile from John S. Fine.
changes.asm changes from John S. Fine.

C.3.50. Version 0.98p3.3

Patch from Conan Brink to allow nesting of %rep directives.
If we're going to allow INT01 as an alias for INT1/ICEBP (one of Jules 0.98p3 changes), then we should allow INT03 as an alias for INT3 as well.
Updated changes.asm to include the latest changes.
Tried to clean up the s that had snuck in from a DOS/Windows environment into my Unix environment, and try to make sure than DOS/Windows users get them back.
We would silently generate broken tools if insns.dat wasn't sorted properly. Change insns.pl so that the order doesn't matter.
Fix bug in insns.pl (introduced by me) which would cause conditional instructions to have an extra "cc" in disassembly, e.g. "jnz" disassembled as "jccnz".

C.3.51. Version 0.98p3.2

Merged in John S. Fine's changes from his 0.98-J4 prerelease; see http://www.csoft.net/cz/johnfine/
Changed previous "spotless" Makefile target (appropriate for distribution) to "distclean", and added "cleaner" target which is same as "clean" except deletes files generated by Perl scripts; "spotless" is union.
Removed BASIC programs from distribution. Get a Perl interpreter instead (see below.)
Calling this "pre-release 3.2" rather than "p3-hpa2" because of John's contributions.
Actually link in the IEEE output format (zoutieee.c); fix a bunch of compiler warnings in that file. Note I don't know what IEEE output is supposed to look like, so these changes were made "blind".

C.3.52. Version 0.98p3-hpa

Merged nasm098p3.zip with nasm-0.97.tar.gz to create a fully buildable version for Unix systems (Makefile.in updates, etc.)
Changed insns.pl to create the instruction tables in nasm.h and names.c, so that a new instruction can be added by adding it *only* to insns.dat.
Added the following new instructions: SYSENTER, SYSEXIT, FXSAVE, FXRSTOR, UD1, UD2 (the latter two are two opcodes that Intel guarantee will never be used; one of them is documented as UD2 in Intel documentation, the other one just as "Undefined Opcode" –– calling it UD1 seemed to make sense.)
MAX_SYMBOL was defined to be 9, but LOADALL286 and LOADALL386 are 10 characters long. Now MAX_SYMBOL is derived from insns.dat.
A note on the BASIC programs included: forget them. insns.bas is already out of date. Get yourself a Perl interpreter for your platform of choice athttp://www.cpan.org/ports/index.html.

C.3.53. Version 0.98 pre-release 3

added response file support, improved command line handling, new layout help screen
fixed limit checking bug, 'OUT byte nn, reg' bug, and a couple of rdoff related bugs, updated Wishlist; 0.98 Prerelease 3.

C.3.54. Version 0.98 pre-release 2

fixed bug in outcoff.c to do with truncating section names longer than 8 characters, referencing beyond end of string; 0.98 pre-release 2

C.3.55. Version 0.98 pre-release 1

Fixed a bug whereby STRUC didn't work at all in RDF.
Fixed a problem with group specification in PUBDEFs in OBJ.
Improved ease of adding new output formats. Contribution due to Fox Cutter.
Fixed a bug in relocations in the `bin' format: was showing up when a relocatable reference crossed an 8192-byte boundary in any output section.
Fixed a bug in local labels: local-label lookups were inconsistent between passes one and two if an EQU occurred between the definition of a global label and the subsequent use of a local label local to that global.
Fixed a seg-fault in the preprocessor (again) which happened when you use a blank line as the first line of a multi-line macro definition and then defined a label on the same line as a call to that macro.
Fixed a stale-pointer bug in the handling of the NASM environment variable. Thanks to Thomas McWilliams.
ELF had a hard limit on the number of sections which caused segfaults when transgressed. Fixed.
Added ability for ndisasm to read from stdin by using `-' as the filename.
ndisasm wasn't outputting the TO keyword. Fixed.
Fixed error cascade on bogus expression in %if – an error in evaluation was causing the entire %if to be discarded, thus creating trouble later when the %else or%endif was encountered.
Forward reference tracking was instruction-granular not operand- granular, which was causing 286-specific code to be generated needlessly on code of the form `shr word [forwardref],1'. Thanks to Jim Hague for sending a patch.
All messages now appear on stdout, as sending them to stderr serves no useful purpose other than to make redirection difficult.
Fixed the problem with EQUs pointing to an external symbol – this now generates an error message.
Allowed multiple size prefixes to an operand, of which only the first is taken into account.
Incorporated John Fine's changes, including fixes of a large number of preprocessor bugs, some small problems in OBJ, and a reworking of label handling to define labels before their line is assembled, rather than after.
Reformatted a lot of the source code to be more readable. Included 'coding.txt' as a guideline for how to format code for contributors.
Stopped nested %reps causing a panic – they now cause a slightly more friendly error message instead.
Fixed floating point constant problems (patch by Pedro Gimeno)
Fixed the return value of insn_size() not being checked for –1, indicating an error.
Incorporated 3Dnow! instructions.
Fixed the 'mov eax, eax + ebx' bug.
Fixed the GLOBAL EQU bug in ELF. Released developers release 3.
Incorporated John Fine's command line parsing changes
Incorporated David Lindauer's OMF debug support
Made changes for LCC 4.0 support (__NASM_CDecl__, removed register size specification warning when sizes agree).

C.4. NASM 0.90-0.97

Revisions before 0.98.

C.4.1. Version 0.97 released December 1997

This was entirely a bug-fix release to 0.96, which seems to have got cursed. Silly me.
Fixed stupid mistake in OBJ which caused `MOV EAX,' to fail. Caused by an error in the `MOV EAX,' support.
ndisasm hung at EOF when compiled with lcc on Linux because lcc on Linux somehow breaks feof(). ndisasm now does not rely on feof().
A heading in the documentation was missing due to a markup error in the indexing. Fixed.
Fixed failure to update all pointers on realloc() within extended- operand code in parser.c. Was causing wrong behaviour and seg faults on lines such as `dd 0.0,0.0,0.0,0.0,...'
Fixed a subtle preprocessor bug whereby invoking one multi-line macro on the first line of the expansion of another, when the second had been invoked with a label defined before it, didn't expand the inner macro.
Added internal.doc back in to the distribution archives – it was missing in 0.96 *blush*
Fixed bug causing 0.96 to be unable to assemble its own test files, specifically objtest.asm. *blush again*
Fixed seg-faults and bogus error messages caused by mismatching%rep and %endrep within multi-line macro definitions.
Fixed a problem with buffer overrun in OBJ, which was causing corruption at ends of long PUBDEF records.
Separated DOS archives into main-program and documentation to reduce download size.

C.4.2. Version 0.96 released November 1997

Fixed a bug whereby, if `nasm sourcefile' would cause a filename collision warning and put output into `nasm.out', then `nasm sourcefile –o outputfile' still gave the warning even though the `-o' was honoured. Fixed name pollution under Digital UNIX: one of its header files defined R_SP, which broke the enum in nasm.h.
Fixed minor instruction table problems: FUCOM and FUCOMP didn't have two-operand forms; NDISASM didn't recognise the longer register forms of PUSH and POP (eg FF F3 for PUSH BX); TEST mem,imm32 was flagged as undocumented; the 32-bit forms of CMOV had 16-bit operand size prefixes; `AAD imm' and `AAM imm' are no longer flagged as undocumented because the Intel Architecture reference documents them.
Fixed a problem with the local-label mechanism, whereby strange types of symbol (EQUs, auto-defined OBJ segment base symbols) interfered with the `previous global label' value and screwed up local labels.
Fixed a bug whereby the stub preprocessor didn't communicate with the listing file generator, so that the –a and –l options in conjunction would produce a useless listing file.
Merged `os2' object file format back into `obj', after discovering that `obj' _also_ shouldn't have a link pass separator in a module containing a non-trivial MODEND. Flat segments are now declared using the FLAT attribute. `os2' is no longer a valid object format name: use `obj'.
Removed the fixed-size temporary storage in the evaluator. Very very long expressions (like `mov ax,1+1+1+1+...' for two hundred 1s or so) should now no longer crash NASM.
Fixed a bug involving segfaults on disassembly of MMX instructions, by changing the meaning of one of the operand-type flags in nasm.h. This may cause other apparently unrelated MMX problems; it needs to be tested thoroughly.
Fixed some buffer overrun problems with large OBJ output files. Thanks to DJ Delorie for the bug report and fix.
Made preprocess-only mode actually listen to the %linemarkers as it prints them, so that it can report errors more sanely.
Re-designed the evaluator to keep more sensible track of expressions involving forward references: can now cope with previously-nightmare situations such as:
mov ax,foo | bar
foo equ 1
bar equ 2
Added the ALIGN and ALIGNB standard macros.
Added PIC support in ELF: use of WRT to obtain the four extra relocation types needed.
Added the ability for output file formats to define their own extensions to the GLOBAL, COMMON and EXTERN directives.
Implemented common-variable alignment, and global-symbol type and size declarations, in ELF.
Implemented NEAR and FAR keywords for common variables, plus far-common element size specification, in OBJ.
Added a feature whereby EXTERNs and COMMONs in OBJ can be given a default WRT specification (either a segment or a group).
Transformed the Unix NASM archive into an auto-configuring package.
Added a sanity-check for people applying SEG to things which are already segment bases: this previously went unnoticed by the SEG processing and caused OBJ-driver panics later.
Added the ability, in OBJ format, to deal with `MOV EAX,' type references: OBJ doesn't directly support dword-size segment base fixups, but as long as the low two bytes of the constant term are zero, a word-size fixup can be generated instead and it will work.
Added the ability to specify sections' alignment requirements in Win32 object files and pure binary files.
Added preprocess-time expression evaluation: the %assign(and %iassign) directive and the bare %if (and%elif) conditional. Added relational operators to the evaluator, for use only in %if constructs: the standard relationals = < > <= >= <> (and C-like synonyms == and !=) plus low-precedence logical operators &&, ^^ and ||.
Added a preprocessor repeat construct: %rep /%exitrep / %endrep.
Added the __FILE__ and __LINE__ standard macros.
Added a sanity check for number constants being greater than 0xFFFFFFFF. The warning can be disabled.
Added the %0 token whereby a variadic multi-line macro can tell how many parameters it's been given in a specific invocation.
Added %rotate, allowing multi-line macro parameters to be cycled.
Added the `*' option for the maximum parameter count on multi-line macros, allowing them to take arbitrarily many parameters.
Added the ability for the user-level forms of EXTERN, GLOBAL and COMMON to take more than one argument.
Added the IMPORT and EXPORT directives in OBJ format, to deal with Windows DLLs.
Added some more preprocessor %if constructs:%ifidn / %ifidni (exact textual identity), and%ifid / %ifnum / %ifstr (token type testing).
Added the ability to distinguish SHL AX,1 (the 8086 version) from SHL AX,BYTE 1 (the 286-and-upwards version whose constant happens to be 1).
Added NetBSD/FreeBSD/OpenBSD's variant of a.out format, complete with PIC shared library features.
Changed NASM's idiosyncratic handling of FCLEX, FDISI, FENI, FINIT, FSAVE, FSTCW, FSTENV, and FSTSW to bring it into line with the otherwise accepted standard. The previous behaviour, though it was a deliberate feature, was a deliberate feature based on a misunderstanding. Apologies for the inconvenience.
Improved the flexibility of ABSOLUTE: you can now give it an expression rather than being restricted to a constant, and it can take relocatable arguments as well.
Added the ability for a variable to be declared as EXTERN multiple times, and the subsequent definitions are just ignored.
We now allow instruction prefixes (CS, DS, LOCK, REPZ etc) to be alone on a line (without a following instruction).
Improved sanity checks on whether the arguments to EXTERN, GLOBAL and COMMON are valid identifiers.
Added misc/exebin.mac to allow direct generation of .EXE files by hacking up an EXE header using DB and DW; also added test/binexe.asm to demonstrate the use of this. Thanks to Yann Guidon for contributing the EXE header code.
ndisasm forgot to check whether the input file had been successfully opened. Now it does. Doh!
Added the Cyrix extensions to the MMX instruction set.
Added a hinting mechanism to allow [EAX+EBX] and [EBX+EAX] to be assembled differently. This is important since [ESI+EBP] and [EBP+ESI] have different default base segment registers.
Added support for the PharLap OMF extension for 4096-byte segment alignment.

C.4.3. Version 0.95 released July 1997

Fixed yet another ELF bug. This one manifested if the user relied on the default segment, and attempted to define global symbols without first explicitly declaring the target segment.
Added makefiles (for NASM and the RDF tools) to build Win32 console apps under Symantec C++. Donated by Mark Junker.
Added `macros.bas' and `insns.bas', QBasic versions of the Perl scripts that convert `standard.mac' to `macros.c' and convert `insns.dat' to `insnsa.c' and `insnsd.c'. Also thanks to Mark Junker.
Changed the diassembled forms of the conditional instructions so that JB is now emitted as JC, and other similar changes. Suggested list by Ulrich Doewich.
Added `@' to the list of valid characters to begin an identifier with.
Documentary changes, notably the addition of the `Common Problems' section in nasm.doc.
Fixed a bug relating to 32-bit PC-relative fixups in OBJ.
Fixed a bug in perm_copy() in labels.c which was causing exceptions in cleanup_labels() on some systems.
Positivity sanity check in TIMES argument changed from a warning to an error following a further complaint.
Changed the acceptable limits on byte and word operands to allow things like `~10111001b' to work.
Fixed a major problem in the preprocessor which caused seg-faults if macro definitions contained blank lines or comment-only lines.
Fixed inadequate error checking on the commas separating the arguments to `db', `dw' etc.
Fixed a crippling bug in the handling of macros with operand counts defined with a `+' modifier.
Fixed a bug whereby object file formats which stored the input file name in the output file (such as OBJ and COFF) weren't doing so correctly when the output file name was specified on the command line.
Removed [INC] and [INCLUDE] support for good, since they were obsolete anyway.
Fixed a bug in OBJ which caused all fixups to be output in 16-bit (old-format) FIXUPP records, rather than putting the 32-bit ones in FIXUPP32 (new-format) records.
Added, tentatively, OS/2 object file support (as a minor variant on OBJ).
Updates to Fox Cutter's Borland C makefile, Makefile.bc2.
Removed a spurious second fclose() on the output file.
Added the `-s' command line option to redirect all messages which would go to stderr (errors, help text) to stdout instead.
Added the `-w' command line option to selectively suppress some classes of assembly warning messages.
Added the `-p' pre-include and `-d' pre-define command-line options.
Added an include file search path: the `-i' command line option.
Fixed a silly little preprocessor bug whereby starting a line with a `%!' environment-variable reference caused an `unknown directive' error.
Added the long-awaited listing file support: the `-l' command line option.
Fixed a problem with OBJ format whereby, in the absence of any explicit segment definition, non-global symbols declared in the implicit default segment generated spurious EXTDEF records in the output.
Added the NASM environment variable.
From this version forward, Win32 console-mode binaries will be included in the DOS distribution in addition to the 16-bit binaries. Added Makefile.vc for this purpose.
Added `return 0;' to test/objlink.c to prevent compiler warnings.
Added the __NASM_MAJOR__ and __NASM_MINOR__ standard defines.
Added an alternative memory-reference syntax in which prefixing an operand with `&' is equivalent to enclosing it in square brackets, at the request of Fox Cutter.
Errors in pass two now cause the program to return a non-zero error code, which they didn't before.
Fixed the single-line macro cycle detection, which didn't work at all on macros with no parameters (caused an infinite loop). Also changed the behaviour of single-line macro cycle detection to work like cpp, so that macros like `extrn' as given in the documentation can be implemented.
Fixed the implementation of WRT, which was too restrictive in that you couldn't do `mov ax,[di+abc wrt dgroup]' because (di+abc) wasn't a relocatable reference.

C.4.4. Version 0.94 released April 1997

Major item: added the macro processor.
Added undocumented instructions SMI, IBTS, XBTS and LOADALL286. Also reorganised CMPXCHG instruction into early-486 and Pentium forms. Thanks to Thobias Jones for the information.
Fixed two more stupid bugs in ELF, which were causing `ld' to continue to seg-fault in a lot of non-trivial cases.
Fixed a seg-fault in the label manager.
Stopped FBLD and FBSTP from _requiring_ the TWORD keyword, which is the only option for BCD loads/stores in any case.
Ensured FLDCW, FSTCW and FSTSW can cope with the WORD keyword, if anyone bothers to provide it. Previously they complained unless no keyword at all was present.
Some forms of FDIV/FDIVR and FSUB/FSUBR were still inverted: a vestige of a bug that I thought had been fixed in 0.92. This was fixed, hopefully for good this time...
Another minor phase error (insofar as a phase error can _ever_ be minor) fixed, this one occurring in code of the form
rol ax,forward_reference
forward_reference equ 1
The number supplied to TIMES is now sanity-checked for positivity, and also may be greater than 64K (which previously didn't work on 16-bit systems).
Added Watcom C makefiles, and misc/pmw.bat, donated by Dominik Behr.
Added the INCBIN pseudo-opcode.
Due to the advent of the preprocessor, the [INCLUDE] and [INC] directives have become obsolete. They are still supported in this version, with a warning, but won't be in the next.
Fixed a bug in OBJ format, which caused incorrect object records to be output when absolute labels were made global.
Updates to RDOFF subdirectory, and changes to outrdf.c.

C.4.5. Version 0.93 released January 1997

This release went out in a great hurry after semi-crippling bugs were found in 0.92.

Really did fix the stack overflows this time. *blush*
Had problems with EA instruction sizes changing between passes, when an offset contained a forward reference and so 4 bytes were allocated for the offset in pass one; by pass two the symbol had been defined and happened to be a small absolute value, so only 1 byte got allocated, causing instruction size mismatch between passes and hence incorrect address calculations. Fixed.
Stupid bug in the revised ELF section generation fixed (associated string-table section for .symtab was hard-coded as 7, even when this didn't fit with the real section table). Was causing `ld' to seg-fault under Linux.
Included a new Borland C makefile, Makefile.bc2, donated by Fox Cutter lmb@comtch.iea.com.

C.4.6. Version 0.92 released January 1997

The FDIVP/FDIVRP and FSUBP/FSUBRP pairs had been inverted: this was fixed. This also affected the LCC driver.
Fixed a bug regarding 32-bit effective addresses of the form[other_register+ESP].
Documentary changes, notably documentation of the fact that Borland Win32 compilers use `obj' rather than `win32' object format.
Fixed the COMENT record in OBJ files, which was formatted incorrectly.
Fixed a bug causing segfaults in large RDF files.
OBJ format now strips initial periods from segment and group definitions, in order to avoid complications with the local label syntax.
Fixed a bug in disassembling far calls and jumps in NDISASM.
Added support for user-defined sections in COFF and ELF files.
Compiled the DOS binaries with a sensible amount of stack, to prevent stack overflows on any arithmetic expression containing parentheses.
Fixed a bug in handling of files that do not terminate in a newline.

C.4.7. Version 0.91 released November 1996

Loads of bug fixes.
Support for RDF added.
Support for DBG debugging format added.
Support for 32-bit extensions to Microsoft OBJ format added.
Revised for Borland C: some variable names changed, makefile added.
LCC support revised to actually work.
JMP/CALL NEAR/FAR notation added.
`a16', `o16', `a32' and `o32' prefixes added.
Range checking on short jumps implemented.
MMX instruction support added.
Negative floating point constant support added.
Memory handling improved to bypass 64K barrier under DOS.
$ prefix to force treatment of reserved words as identifiers added.
Default-size mechanism for object formats added.
Compile-time configurability added.
#, @, ~ and c{?} are now valid characters in labels.
-e and -k options in NDISASM added.

C.4.8. Version 0.90 released October 1996

First release version. First support for object file output. Other changes from previous version (0.3x) too numerous to document.