Re: [PATCH v3 1/6] common-main.o: move non-trace2 exit() behavior out of trace2.c

> Change the exit() wrapper added in ee4512ed481 (trace2: create new
> combined trace facility, 2019-02-22) so that we'll split up the trace2
> logging concerns from wanting to wrap the "exit()" function itself for
> other purposes.
> 
> This makes more sense structurally, as we won't seem to conflate
> non-trace2 behavior with the trace2 code. I'd previously added an
> explanation for this in 368b5843158 (common-main.c: call exit(), don't
> return, 2021-12-07), that comment is being adjusted here.
> 
> Now the only thing we'll do if we're not using trace2 is to truncate
> the "code" argument to the lowest 8 bits.
> 
> We only need to do that truncation on non-POSIX systems, but in
> ee4512ed481 that "if defined(__MINGW32__)" code added in
> 47e3de0e796 (MinGW: truncate exit()'s argument to lowest 8 bits,
> 2009-07-05) was made to run everywhere. It might be good for clarify
> to narrow that down by an "ifdef" again, but I'm not certain that in
> the interim we haven't had some other non-POSIX systems rely the
> behavior. On a POSIX system taking the lowest 8 bits is implicit, see
> exit(3)[1] and wait(2)[2]. Let's leave a comment about that instead.
> 
> 1. https://man7.org/linux/man-pages/man3/exit.3.html
> 2. https://man7.org/linux/man-pages/man2/wait.2.html
> 
> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> ---
>  common-main.c     | 20 ++++++++++++++++----
>  git-compat-util.h |  4 ++--
>  trace2.c          |  8 ++------
>  trace2.h          |  8 +-------
>  4 files changed, 21 insertions(+), 19 deletions(-)