Pure FB Runtime Library (in progress)

[counting_pine]

Yeah, 'for(;;)' is a common idiom for an "endless" loop. Presumably the loop body has a 'break' statement somewhere.

[Imortis]

... Presumably the loop body has a 'break' statement somewhere.

Correct. I was pretty sure that I could just replace it, but wanted to ask someone a bit more knowledgeable.

[Imortis]

signals.c requires a c header called signal.h I am not sure if that can be converted directly or if there should be a different code to replace that.

Any one have some insight on this one?

Also: crt/getopt.bi does not seem to exist in our crt headers, but it is called in crt/win32/unistd.bi. The linux and DOS versions of that file have a whole lot more content in them. Is this a problem with the unistd.bi file for win32 or with the getopt.bi file not existing?

[Imortis]

Also, as a sanity check:

Code: Select all

*--dst

in c is equivilant to

Code: Select all

dst[-1]

in FB right?

[St_W]

Also, as a sanity check:

Code: Select all

*--dst

in c is equivilant to

Code: Select all

dst[-1]

in FB right?

It also decrements dst.

[Imortis]

dst is a ptr. Is it decrementing the pointer or the value? I thought it was the ptr.

[St_W]

dst is a ptr. Is it decrementing the pointer or the value? I thought it was the ptr.

The pointer ist decremented first, then the (decremented) pointer is dereferenced.

dst[-1] does not update the pointer.

[Imortis]

Ah! So it would be better to replace it with this:

Code: Select all

dst -= 1
*dst = blah blah blah 

[St_W]

Ah! So it would be better to replace it with this:

Code: Select all

dst -= 1
*dst = blah blah blah 

Yes.

btw there are nice webservices to test all kinds of code, even including C/C++ or FreeBasic. For the discussed code e.g. try this: https://repl.it/repls/WorriedFrenchAurochs

[Imortis]

https://repl.it/repls/WorriedFrenchAurochs

I did not know this was a thing. Thanks for the link it will be helpful, indeed.

[Imortis]

I just hit a pretty big milestone:

Every non-OS specific file had been converted! I also got to work on the Darwin files but his a bit of a snag. They include a header for Mach-O stuff, which is mac only. It looks like they only use one function, but I am unsure how much of the headers are required to make that one function work. If anyone wants to take a look, please do.
https://github.com/ImortisInglorian/fbr ... ter/darwin

In the mean time I am going to tackle the other OS-specific files. DOS is up next.

[PaulSquires]

Great work Imortis, you can be sure that the whole community is rooting for you and great success for this project. It will help move FB forward immensely. Awesome!

[Imortis]

So, there is a file called cpudetect.s It is assembly passed to GCC. Could this be passed directly to GAS or even converted into a bas file using asm...end asm blocks? I tried to find info about what syntax the .s files use, but I could not find anything concrete from a quick google search.

[marcov]

Afaik gcc only processes .S files with a preprocessor so that they become .s files. So if it has #ifdefs, it might need preprocessing

[St_W]

So, there is a file called cpudetect.s It is assembly passed to GCC. Could this be passed directly to GAS or even converted into a bas file using asm...end asm blocks? I tried to find info about what syntax the .s files use, but I could not find anything concrete from a quick google search.

Just create a BAS file with inline assembly, if needed. As long as the function signature and calling convention matches everything should be fine.