IRC channel logs

<Guest74>Hi, quick question about stage0-posix; are the M2-Planet variants produced in Phases 5 and 15 supposed to be functionally identical, with the only difference being debuggability, or are there additional differences?

<stikonas>Guest74: Hi

<stikonas>functionally output should be the same

<stikonas>well, mostly

<stikonas>they use different C libraries

<stikonas>theoretically that shouldn't introduce any functional differences

<stikonas>it does mean performance differences

<stikonas>also on stage0-uefi, that 2nd C library is functionally better, so there are some functional differences in UEFI (e.g. posix paths are only supported by the 2nd build)

<Guest74>Hmm. I am asking since it seems to me like `#if !__M2__` behaves incorrectly under the Phase 5 one

<stikonas>hmm, that should work I think

<stikonas>could be bug, but not one that I'm aware of

<stikonas>do you have a reproducer?

<Guest74>I mean, just save something like this to a file and run `M2-Planet -f foo.c`

<Guest74>```

<Guest74>#if !__M2__

<Guest74>int main() {}

<Guest74>#endif

<Guest74>```

<stikonas>hmm, yes, I can see the difference

<stikonas>at the moment it is not clear to me why that happens

<stikonas>Guest74: so __M2__ is correctly expanded

<stikonas>so I think the bug is in #if processing

<stikonas>just __M2__ is replaced with 42 in both M2 and M2-Planet

<stikonas>Guest74: oh, maybe ! is not working

<stikonas>Guest74: anyway, I don't really understand it at the moment. If you want, you could try to inserts prints in various places in the sourcecode

<stikonas>and try to track the difference

<stikonas>that's probably the only viable way to debug M2...

<stikonas>since in gdb you just see pure disassembly, very hard to follow

IRC channel logs

2025-06-11.log