RedGhost, you MUST use match instead. Check my previous post to see why.

Because the point is not including different files, the point is always parsing a specific file before the main file.
Short story: I designed a macro language with nasm. I used it just like

where "clcomp" was an alias:

It went well, and I made a lot of code.
Unfortunatelly nasm lacks some feature I needed, which fasm has, but fasm lacks command line input, so I can't use neither of them. Here we are. I made a request for command line include, I got many stupid answers for a stupid example and reasons why not. Thanks, but that won't help me out, and I don't want to mess up my CL files with includes at begining. End of story.

Besides, all common compiler has command line input, wonder why...

ps1.: you don't have to use command line include if you don't want to.
ps2.: I started this thread beacuse I thought there are others missing this feature.
ps3.: I patched my version

can't help you, i just want to react on this:

Most languages are overbloated and badly designed. FASM is unlike these, it's minimalistic and most clear language in it's category. Wonder why... maybe it has something to do with not including every feature someone would use once a especially when problem can be solved other way.

Didn't source in your language have any header? Is it that pain to have one more line there. In your case external input even is not needed, there are problems which require external input much more than saving one line per source.

before posting i tested win32.inc had xor eax, eax and linux.inc had 'mov dax, 4' <-- error, when WIN32 = 1 xor eax, eax was compiled and when LINUX = 1 an error was given and if neither were defined it would print the display message assemble nothing and give no error

how exactly could the wrong macro be used if only one file is included and parsed?

i find it hard to believe that the many people who replied, all of whom were against commandline includes are stupid and you are infact the only intelligent one, infact i assume it may be quite the opposite but there is no reason for flamming or an argument

edit: gah my first double post nooooo

Just because you prefer FASM doesn't mean you should ignore NASM -P (or whatever) if it does what you want. If you need certain "features" of FASM and/or NASM, then you are heading towards HLL and not pure assembly anyways. If you look at FASM's source, it does not use every advanced feature there is. I personally avoid lots of the high level stuff, but that's out of unfamiliarity and distaste for such things (and also so my wimpy code is easier to port to other assemblers, if needed).

FASM was written with different goals in mind than NASM, so if you want FASM to do something, you're best off modifying it yourself. Privalov has better things to do than to cater to every feature request. Believe me, I've had the urge to request a few things myself, but I have restrained myself so as not to be lame. Maybe some people should fork FASM (like XEmacs, YASM, Kolibri, etc).

For instance, I don't like the way FASM is able to overwrite the input file with the output, but this can be avoided (if necessary) by a simple .BAT file. See? No reason to nag Privalov or to modify FASM itself when an outside tool can do the job.

Sometimes, if you want something done right, you've got to do it yourself. Besides, doing stuff by yourself is the best way to learn.



P.S. I like Tomasz's batch hack above, it's cool. I do stuff like that all the time.

P.P.S. Post number 100, woot!

If you use "if" instead of "match" both files gets included and parsed and one of them gets assembled. I'm think that the better way is using match as UCM does (2nd post in this thread), is faster and prevents undesirable interactions between both include files.