[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: cannot make gcc report undeclared builtins
From: |
Simon Josefsson |
Subject: |
Re: cannot make gcc report undeclared builtins |
Date: |
Thu, 01 Feb 2024 11:54:16 +0100 |
User-agent: |
Evolution 3.44.4-0ubuntu2 |
ons 2024-01-31 klockan 16:46 -0800 skrev Paul Eggert:
> On 1/31/24 03:23, Simon Josefsson via Bug reports for autoconf wrote:
>
> > https://buildd.debian.org/status/fetch.php?pkg=libidn2&arch=loong64&ver=2.3.7-1&stamp=1706360630&raw=0
>
>
> Here's the crucial part of that log:
>
> > configure:15961: checking for size_t
> > configure:15961: gcc -c -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=.
> > -fstack-protector-strong -Wformat -Werror=format-security -Wdate-
> > time -D_FORTIFY_SOURCE=2 conftest.c >&5
> > In file included from /usr/include/linux/stat.h:5,
> > from /usr/include/loongarch64-linux-
> > gnu/bits/statx.h:31,
> > from /usr/include/loongarch64-linux-
> > gnu/sys/stat.h:465,
> > from conftest.c:72:
> > /usr/include/linux/types.h:5:10: fatal error: asm/types.h: No such
> > file or directory
> > 5 | #include <asm/types.h>
> > | ^~~~~~~~~~~~~
> > compilation terminated.
>
>
> It appears that '#include <sys/stat.h>' doesn't work in some cases,
> and
> I suggest looking into why that is.
Thanks -- given some other information, I now realize it does affect
other packages too. So I'm less certain it has anything to do with
autoconf. The way things fail is surprising though, and a better
failure mode would have been nice; it seems this particular autoconf
test has some other assumptions.
/Simon
signature.asc
Description: This is a digitally signed message part
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: cannot make gcc report undeclared builtins,
Simon Josefsson <=