[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: FAIL: autopoint-3
From: |
Bruno Haible |
Subject: |
Re: FAIL: autopoint-3 |
Date: |
Thu, 20 Jun 2024 23:19:41 +0200 |
Hi Santiago,
> > I would try to apply the commits
> > 5eecb370ed02eb44f9e0dd7a648d961b8ccd54a1
> > and
> > f23b7847990cb31b59729517314e5d0fb649f64e
> > from the GNU gettext git repository.
>
> Ok, I added those two patches.
>
> It still fails, but now, using the file-based backend of sbuild
> to build the package, there is a thing which I missed before:
You didn't miss it before. It was not shown, due to a typo in file name.
Fixed through the commit 5eecb370ed02eb44f9e0dd7a648d961b8ccd54a1 that you
applied.
> + autoconf
> + test 0 = 0
> + touch ltmain.sh
> + rm -rf autom4te.cache
> + automake -a -c
> + test 0 = 0
> + : /bin/bash
> + /bin/bash ./configure
> + test 0 = 0
> + make
> + test 2 = 0
> + cat autopoint.err
> gcc: error: unrecognized command-line option
> ‘-<A8><A8><A8><A8><A8><A8><A8><A8>^Q^PDHAVE_STDIO_H=1’
> make[9]: *** [Makefile:506: hello.o] Error 1
> make[8]: *** [Makefile:542: all-recursive] Error 1
>
>
> Why would gcc be receiving such strange string with funny chars
> ending by HAVE_STDIO_H=1 as a command line?
I've never seen something like that. Such characters occur
a) when interactively typing in some editor,
b) when raw binary data is exfiltrated, for example due to a memory bug.
I would suggest to search for a bug in 'sbuild', by using clang's ASAN and
UBSAN sanitizers:
CC="clang
-fsanitize=address,undefined,signed-integer-overflow,shift,integer-divide-by-zero";
CXX="clang++
-fsanitize=address,undefined,signed-integer-overflow,shift,integer-divide-by-zero";
export CC CXX
CFLAGS="-O0 -fno-omit-frame-pointer -ggdb"; CXXFLAGS="-O0
-fno-omit-frame-pointer -ggdb"; export CFLAGS CXXFLAGS
ASAN_OPTIONS="detect_leaks=0 abort_on_error=1 allocator_may_return_null=1"
export ASAN_OPTIONS
Bruno
- FAIL: autopoint-3, Santiago Vila, 2024/06/18
- Re: FAIL: autopoint-3, Bruno Haible, 2024/06/19
- Re: FAIL: autopoint-3, Santiago Vila, 2024/06/20
- Re: FAIL: autopoint-3, Santiago Vila, 2024/06/21
- Re: FAIL: autopoint-3, Bruno Haible, 2024/06/21
- Re: FAIL: autopoint-3, Santiago Vila, 2024/06/22
- Re: FAIL: autopoint-3, Bruno Haible, 2024/06/22