Do some other assemblers have standard. Because i want to look at those, steal interesting parts, and make fun of badly designed or writen ones

only one i found yet was HLA... at first peek it seems well designed and not-that-well written

Seems Randall had VERY similar ideas about standard library design as me.

masm has m32lib

thx

after 10 seconds of lookin at it... little weird, i wouldn't use that... :s

yeah, m32lib is ill-designed (or it has no design at all)

has no design, looks like just a set of routines, AND is also badly written

His primary goal was to get the library out asap and think about writing it better later or leaving that job to others. It's going to end up on sourceforge after some consideration is put into making it portable across more OSes.

hm... you can't say his library isn't clear...

I think there is no standard in assembly.
Here you can find various libraries and code snippets.
http://www.programmersheaven.com/zone5/cat17/index.htm
Also my Os has its library, and 100% of octasm programmers (octavio) think it is very readable

gimme your lib

Octavio wrote:

en Espanol?

Stay away from m32lib; it indeed has no design but is a hodgepodge of - often - bad code. And there's a bunch of bugs in there, some of which won't be fixed.

i saw. in about 10 seconds...

NO just a few comments in spanish or spaninglish, but also few comments in english
I admit that is chaotic and hard to read (for others) but has a lot of code
and makes development easier.
But assembly programmers are not limited to assembly libraries, it is easy
to use a library written in another programing language since assembly supports all calling conventions.On my Os i use the ZLIB wich is written in
C ,there is no need to translate everthing to assembly.

Translating just for the sake of translating is silly and a waste of time. There might be a point if you're going to improve something, and improve it substantially... but most times, efforts can be spent better elsewhere.

like writing library that returns error ALWAYS in CF and thus prevents you from hell error checking in C-ish libraries (libc, win32 api, ...).

Error in CF isn't that hot an idea, IMHO. Often you don't need a boolean fail/success, but a completion code - carry can't give you that. Of course you can then add a "GetLastError", but that means even more function calls and whatnot. If you're doing internal code, sure, CF can be nifty enough. If you're writing libraries, stick to the standard ABI - makes it easier for more people to benefit from the work.

what is ABI?

if you mean something like windows api returns error - no. sometimes error = 0, sometimes 1, sometimes non 0, sometimes -1, sometimes 7FFFFFFFh, sometimes error code is returned directly. most times GetLastError is also set, but not always... ke?

i want to FASM users benefit from my work, not as many people as possible. Too much generality causes less power.

ABI = Application Binary Interace. In the case of x86-32 this generally means preserve {EBX,ESI,EDI,EBP} as well as flags, {EAX,ECX,EDX} are trashable, and return values go in EAX or EDX:EAX (or whatever for floating-point).

Back when I did much (and almost exclusively, for a while) assembly programming many years ago, I bitched at why compilers didn't use CF, and used it myself. But today it's irrelevant optimization-wise (and sorta was back then anyway), and since this thread is titled "Standard libraries of other assemblers", IMHO following the standard ABI is most important.