Index
> Main > flat assembler 1.58Goto page 1, 2 Next |
| Author |
|
|
Tomasz Grysztar 12 Feb 2005, 13:32
Now available in the Download section.
This is probably the last cleanup release before switching to the 1.6x line (with x86-64 support). I've changed the packaging, so from now only one package for Win32 is available, containing both the command line version and IDE. I consider also adding some NSIS (or some other similar tool) based installer for Win32 version in future. The includes for Win32 version have been updated in a way, which allows to include them with any method - with the %FASMINC% variable, by specifying the full path or by doing just INCLUDE "WIN32AX.INC" with the INCLUDE variable properly set up (which is now recommended by fasm's docs). Also documentation has been extended a bit, few more examples for some directives has been added. If you think some directive still needs more explanation or examples in official docs, please let me know. Generally all the existing features of fasm are mentioned in the docs, however some of the info may be hidden too well for anyone but me. Last edited by Tomasz Grysztar on 12 Mar 2005, 18:50; edited 4 times in total |
|||
12 Feb 2005, 13:32 |
|
|
Tomasz Grysztar 12 Feb 2005, 13:43
Well, some people (beginners) asked for this. It would also set up the environment variables automatically, etc.
|
|||
|
12 Feb 2005, 13:43 |
|
|
f0dder 12 Feb 2005, 14:07
True, it would be nice for beginners. I hope you'll keep distributing .zip releases though, for something like fasm having to install and then move/remove files would be annoying
 |
|||
|
12 Feb 2005, 14:07 |
|
|
Embrance 12 Feb 2005, 17:37
the v1.6 branch will support all current stuff,wont it?
|
|||
|
12 Feb 2005, 17:37 |
|
|
Tomasz Grysztar 12 Feb 2005, 17:45
Of course.
|
|||
|
12 Feb 2005, 17:45 |
|
|
bubach 12 Feb 2005, 18:10
How much will the 64-bit support add to the assembler size?
If you don't want to guess, whats the size of your current build? |
|||
|
12 Feb 2005, 18:10 |
|
|
Tomasz Grysztar 12 Feb 2005, 18:37
Go here to get the latest pre-release:
http://board.flatassembler.net/topic.php?t=3000 |
|||
|
12 Feb 2005, 18:37 |
|
|
madmatt 12 Feb 2005, 20:31
Quote: If you think some directive still needs more explanation or examples in official docs, please let me know. Hi, everyone This is more a windows question, but how about more explanations on using the 'rescource.inc' and 'if.inc' macros, this would be nice especially for beginners, or even assembly pros, but are new to windows/fasm assembly programming. And I'll ask my next question in the windows section. MadMatt |
|||
|
12 Feb 2005, 20:31 |
|
|
bubach 12 Feb 2005, 20:52
Ok..
|
|||
|
12 Feb 2005, 20:52 |
|
|
vid 14 Feb 2005, 01:01
just wondering - how does modularity of fasm look now. I would really like it if you could change destination processor just by including other "x86.inc" and recompiling FASM.
|
|||
|
14 Feb 2005, 01:01 |
|
|
beppe85 14 Feb 2005, 02:14
vid wrote: just wondering - how does modularity of fasm look now. I would really like it if you could change destination processor just by including other "x86.inc" and recompiling FASM. I had to modify a few lines in PARSER.INC, ASSEMBLE.INC and EXPRESSI.INC for MIPS, but only the last was really needed to support a larger register size. I guess I could let others untouched if I don't wanted to bother with some intrinsic x86 stuff. For this x86-64, I saw there's modifications in (almost) every file, but I guess was not structural ones. _________________ "I assemble, therefore I am" If you got some spare time, visit my blog: http://www.beppe.theblog.com.br/ and sign my guestmap |
|||
|
14 Feb 2005, 02:14 |
|
|
Tomasz Grysztar 14 Feb 2005, 05:52
The latest version (called 1.59.0 now) can be recompiled to either support x86-64 or nor just by using either the X86_64.INC or X86.INC include file.
|
|||
|
14 Feb 2005, 05:52 |
|
|
MCD 14 Feb 2005, 09:07
I got 2 suggestions: couldn't you add this cursor resource macro for the regular Fasm Win32 archiv? This snippet should be in the "/INCLUDE/MACRO/RESOURCE.INC" file:
Code: macro cursor group,[label,cursor_file] { common local count count = 0 forward local data,size,position label dd RVA data,size+4,0,0 virtual at 0 file cursor_file:6,16 load size dword from 8 load position dword from 12 end virtual data file cursor_file:10,4 file cursor_file:position,size count = count+1 common local header align 4 group dd RVA header,6+count*14,0,0 header dw 0,2,count forward local width,height virtual at 0 file cursor_file:6,2 load width byte from 0 load height byte from 1 end virtual dw width,width+height,1,1 dd size+4 dw label#.resid common align 4 } Not my code. It was posted here on the Fasm board some time ago, but I don't remember the author. Another thing is, I moved some more variables/data from the interface files "FASM.ASM"/"FASM.INC" to the "VARIABLE.INC" file: Code: ; flat assembler core variables ; Copyright (c) 1999-2005, Tomasz Grysztar. ; All rights reserved. ; Initialized variables: _copyright dm 'Copyright (c) 1999-2005 Tomasz Grysztar',nl _logo dm 'Flat Assembler version ',VERSION_STRING,nl _usage dm 'usage: fasm [-m#] [-p#] source [output]',nl _passes_suffix dm ' passes, ' _seconds_suffix dm ' seconds, ' _bytes_suffix dm ' bytes.',nl _error_prefix dm 'error: ' _error_suffix dm '.',nl _line_number_start dm ' [' _line_data_start dm ':',nl align 4 _include db 'INCLUDE',0 locals_counter db 7,'0000000' ; Uninitialized variables which have to be set up by interface: memory_start dd ? memory_end dd ? additional_memory dd ? additional_memory_end dd ? input_file dd ? output_file dd ? passes_limit dw ? ; Uninitialized internal core variables: current_pass dw ? include_paths dd ? free_additional_memory dd ? source_start dd ? code_start dd ? code_size dd ? real_code_size dd ? written_size dd ? headers_size dd ? current_line dd ? macros_list dd ? macro_constants dd ? macro_line dd ? macro_block dd ? macro_block_line dd ? macro_block_line_number dd ? macro_embed_level dd ? struc_name dd ? current_locals_prefix dd ? fix_constants_count dd ? equ_constants_count dd ? anonymous_reverse dd ? anonymous_forward dd ? labels_list dd ? label_hash dd ? label_leaf dd ? hash_tree dd ? org_origin dq ? org_registers dd ? org_start dd ? org_symbol dd ? undefined_data_start dd ? undefined_data_end dd ? counter dd ? counter_limit dd ? error_line dd ? error dd ? display_buffer dd ? structures_buffer dd ? number_start dd ? current_offset dd ? value dq ? fp_value rd 8 adjustment dq ? symbol_identifier dd ? address_symbol dd ? address_high dd ? format_flags dd ? symbols_stream dd ? number_of_relocations dd ? number_of_sections dd ? stub_size dd ? stub_file dd ? current_section dd ? machine dw ? subsystem dw ? subsystem_version dd ? image_base dd ? resource_data dd ? resource_size dd ? next_pass_needed db ? macro_status db ? parenthesis_stack db ? output_format db ? code_type db ? reloc_labels db ? times_working db ? virtual_data db ? fp_sign db ? fp_format db ? value_size db ? size_override db ? value_undefined db ? value_type db ? address_size db ? compare_type db ? base_code db ? extended_code db ? postbyte_register db ? segment_register db ? operand_size db ? mmx_size db ? imm_sized db ? jump_type db ? operand_prefix db ? characters rb 100h converted rb 100h Maybe someone should then remove the leading _ from some of these variables, thus "_usage" will become "usage". Note that it's crucial here that the interfaces define their appropriate "dm" macro (define message, $-terminated for DOS, 0-terminated for all others) and "nl" (new line, 0Dh,0Ah for DOS/Win32, 0Ah for Linux/Unix) equate. Another, only superficial rearrangement would be to melt the "_copyright" and "_usage" variables into 1, named "_info" variable, which will be displayed only if no/incorrect parameters were passed at program start. This: Code: _copyright dm 'Copyright (c) 1999-2005 Tomasz Grysztar',nl _logo dm 'Flat Assembler version ',VERSION_STRING,nl _usage dm 'usage: fasm [-m#] [-p#] source [output]',nl would become Code: _logo dm 'Flat Assembler version ',VERSION_STRING,nl _info dm 'Copyright (c) 1999-2005 Tomasz Grysztar',nl,\ nl,\ 'fasm [-m#] [-p#] source [output]',nl Note that I also added the memory and passes switches for command line display with [-m#] [-p#]. _________________ MCD - the inevitable return of the Mad Computer Doggy -||__/ .|+-~ .|| || |
|||
|
14 Feb 2005, 09:07 |
|
|
Tomasz Grysztar 14 Feb 2005, 10:00
These variables (in fact: constants) are not common for all interfaces, therefore should not be included in core.
The "-m#" and "-p#" parameters can be specified anywhere in the command line (before, after, or even between the source and the output) - do you think listing them this way is the best form? |
|||
|
14 Feb 2005, 10:00 |
|
|
MCD 14 Feb 2005, 10:11
Privalov wrote: The "-m#" and "-p#" parameters can be specified anywhere in the command line (before, after, or even between the source and the output) - do you think listing them this way is the best form? Well, it's not perfect, but how would you manage this else? I find something like "[-memoryKb] [-passes]" a bit long. Privalov wrote: These variables (in fact: constants) are not common for all interfaces, therefore should not be included in core. hm, what about the cursor resource macro? I find it rather handy. _________________ MCD - the inevitable return of the Mad Computer Doggy -||__/ .|+-~ .|| || Last edited by MCD on 14 Feb 2005, 10:49; edited 1 time in total |
|||
|
14 Feb 2005, 10:11 |
|
|
Tomasz Grysztar 14 Feb 2005, 10:33
Indeed, Win32 includes need many thing to be added, I will update them when finished with other things.
|
|||
|
14 Feb 2005, 10:33 |
|
|
Tomasz Grysztar 16 Feb 2005, 08:12
Done.
|
|||
|
16 Feb 2005, 08:12 |
|
|
FlashBurn 19 Feb 2005, 11:20
I have some problems with the new version I hadn´t with the old!
Things like "mov ds, eax" don´t work anymore, if I´m in 32bit modus! Also "SYSCALL" seems to be reserved label, why?! |
|||
|
19 Feb 2005, 11:20 |
|
|
Tomasz Grysztar 19 Feb 2005, 11:38
As I wrote in WHATSNEW.TXT, there was "more strict checking of operands for few instructions" added - "mov ds,eax" is in fact invalid instruction, since such MOV always uses 16-bit size (use "mov ds,ax" instead, there will be no any prefix added). However "mov eax,ds" is still allowed, because in this exceptional case the all 32 bits of target are affected (the value is zero-extended).
"syscall" and "sysret" are the new instructions (introduced by both AMD64 and Intel EM64T). |
|||
|
19 Feb 2005, 11:38 |
|
| Goto page 1, 2 Next < Last Thread | Next Thread > |
Forum Rules:
|