Thread (1 message) 1 message, 1 author, 2025-07-11

Re: [PATCH RESEND] sane-ctype: fix compiler error on Amazon Linux 2

From: Junio C Hamano <hidden>
Date: 2025-07-11 15:11:27

Patrick Steinhardt [off-list ref] writes:
On Thu, Jul 10, 2025 at 02:01:15PM -0700, Junio C Hamano wrote:
quoted
Patrick Steinhardt [off-list ref] writes:
quoted
quoted
quoted
This error bisect back to 75a044f748 (git-compat-util.h: split out
POSIX-emulating bits, 2025-02-18), where lots of bits got split out of
"git-compat-util.h" into a new "compat/posix.h" header.
So there was a subtle change in the inclusion order or something,
i.e. we used to include <sane-ctype.h> a lot earlier as part of the
<git-compat-util.h> that tightly controls inclusion order exactly to
avoid this kind of problems, but that has eroded recently?

In any case, it means that our definitions in <sane-ctype.h> will be
overwritten if we allow system-supplied <ctype.h> included, and this
is an obvious workaround, likely to remain correct as long as their
<ctype.h> is not so broken to allow multiple inclusion.
Yup, exactly. Toon figured out that this is actually neither glibc nor
GCC that causes this -- it's OpenSSL that transitively pulls in
<ctype.h>. Later versions of OpenSSL seem to not do that anymore.

I can reroll and update the commit message, but am a bit hesitant given
that I wasn't able to send out this mail in the first place due to
whatever reason :/
I did not see much need to update the proposed log message; I was
merely trying to make sure I understand what motivation and
assumption are behind the change.

It does make me curious what in your original message got it
truncated.  If you received the message you are responding to in
full, that means there is not something funky in the diff output
;-), but then I do not know what it is in your original message,
either.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help