Introduce sys_common::rt::rtprintpanic! to replace sys_common::util functionality

[CDirkx]

This PR introduces a new macro rtprintpanic!, similar to sys_common::util::dumb_print and uses that macro to replace all sys_common::util functionality.

[rust-highfive]

r? @m-ou-se

(rust-highfive has picked a reviewer for you, use r? to override)

[CDirkx]

The name could be something different, rtwrite, rtprint, rtoutput etc.

[m-ou-se]

Maybe something with print in the name would be better. err might suggest it returns from the function or panics or aborts.

[CDirkx]

rtprintpanic!? I think I'll also add a comment about that this writes to the designated "panic output", which is just stderr on most platforms.

[CDirkx]

Is there a particular reason why here in panicking intrinsics::abort is used instead of sys::abort_internal? Otherwise this could also be written with rtabort!.

[m-ou-se]

sys::abort_internal calls libc::abort() (at least on unix). intrinsics::abort does not rely on libc or the platform, but simply results in executing an invalid instruction. I don't know if there's a reason to avoid libc::abort() here. It'd require some digging through the git history to see if this was changed on purpose, or added before the alternative existed, or something like that. :(

[CDirkx]

I'll look into it

[CDirkx]

I looked through the history, these intrinsics::abort were indeed added before sys::abort_internal existed. As far as I can tell these could be replaced, along with a safety comment added to abort_internal that it's implementation is not allowed to panic (else panicking is stuck in a loop). However there are more occurrences of intrinsics::abort in std that in my opinion could also be replaced, so I think I'm leaving it like this for now and will address those together in a future PR.

[CDirkx]

Ready for another look, only changed the name from rterr to rtprintpanic.

[bors]

☔ The latest upstream changes (presumably #81858) made this pull request unmergeable. Please resolve the merge conflicts.

[ijackson]

I have some Opinions about intrinsics::abort vs abort_internal. I have a branch which I am going to turn in an MR shortly, now #81858 has landed. That has some information about the properties of abort_internal. Should be later today.

[ijackson]

See #85377.

[m-ou-se]

@CDirkx Thanks again! This looks good to me. Can you rebase it to include the new abort message in panicking.rs as well?

[CDirkx]

Rebased 👍🏻

[m-ou-se]

Thanks!

@bors r+

[bors]

📌 Commit 4ff5ab5 has been approved by m-ou-se

[bors]

⌛ Testing commit 4ff5ab5 with merge 5ab0f37...

[bors]

☀️ Test successful - checks-actions
Approved by: m-ou-se
Pushing 5ab0f37 to master...