Add at_quick_exit() / quick_exit() #10
Conversation
|
What are the plans for C11 adoption upstream? Would it make sense to send our code there (and then move it downstream)? |
To my understanding upstream doesn't actively pursue C11 support atm, but is open to adding parts of it when requested. I do actually plan to submit this upstream, too, but figured that merging it here would be the more quick way on the way to libc++ - rebasing to upstream if/when merged there could then happen together with a clean-up of the dlmalloc-changes. |
functions/stdlib/at_quick_exit.c
Outdated
| static void checkhandler( void ) | ||
| { | ||
| int i; | ||
| for ( i = 0; i < 31; ++i ) |
There was a problem hiding this comment.
Is there any reason why there are flags[0 to 31], but only flags[0 to 30] are being tested?
There was a problem hiding this comment.
The testcase is almost an exact copy of the one from at_exit.c, and I assume it was done this way by mistake. I changed the loop here and it seems to still work fine.
There was a problem hiding this comment.
I had thought so; please also fix / report issue upstream (before resolving conversation / merging PR - so we don't forget about it)
| #ifndef REGTEST | ||
|
|
||
| /* TODO - "...except that a function is called after any previously registered | ||
| functions that had already been called at the time it was registered." |
There was a problem hiding this comment.
Context for this quote would help
There was a problem hiding this comment.
It's from the C standard (copied from exit.c), from the section where quick_exit and exit are explained. I added a small note.
There was a problem hiding this comment.
What is there left TODO then? Is this about the case where functions are registered from within the callbacks? I'm still not sure what is meant by that comment.
There was a problem hiding this comment.
If I may? I added that comment to my implementation of exit.c to remind myself that I am not 100% sure I understood the meaning of the standard here, and that I want (TODO) to go back and double-check I did not miss something subtle.
The comment just got copied to quick_exit.c, as the implementations are virtually identical.
| EXIT_SUCCESS and EXIT_FAILURE above.) | ||
| quick_exit() does not return. | ||
| */ | ||
| _PDCLIB_PUBLIC void quick_exit( int status ); |
There was a problem hiding this comment.
nit: It's weird how none of the functions listed here seem to use _PDCLIB_NORETURN; we can probably change this in a follow-up across the codebase / upstream.
There was a problem hiding this comment.
I noticed as well and plan to mention this upstream.
There was a problem hiding this comment.
Would be good to have this reported before resolving this conversation / finishing this PR, because we'll probably just forget about it otherwise.
|
The submitted functionality has been implemented upstream. I'm closing this in favor of an update-PR I plan to submit soon, so we can benefit from the additional improvements made upstream. |
…anks to sam-itt for pointing it out (GitHub issue #10). git-svn-id: svn://rootdirectory.ddns.net/pdclib/trunk@863 bcf39385-58cc-4174-9fcf-14f50f90dd47
This implements
at_quick_exit()andquick_exit(). Both are C11 functions that are required by libc++.No "xbox"-prefix for this commit since it's a generic implementation.