-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 8/29/2008 7:14 PM:
Committed like this. I preferred to do the #ifdef __GNUC__ at configure
time,
because that's more maintainable.
Looks good. But were the changes to lock-tests and tls-tests intentional?
Reuben has chosen to set CPP_PEDANTIC to true, so in order to get rid of the
warning, he needs to make dir-origin == EXTENSION evaluate to false.
This means, specify the directory containing the built gnulib header files
with '-isystem' instead of '-I'.
I could try implementing something
Reuben has chosen to set CPP_PEDANTIC to true, so in order to get rid of the
warning, he needs to make dir-origin == EXTENSION evaluate to false.
This means, specify the directory containing the built gnulib header files
with '-isystem' instead of '-I'.
You can put #pragma GCC
Paolo Bonzini wrote:
You can put #pragma GCC system_header in the gnulib files.
However, this pragma not only affects warnings, it also causes __STDC__ to
evaluate to 0 in such a file, on some platforms (those which define
STDC_0_IN_SYSTEM_HEADERS, namely Solaris and Interix).
Incidentally, the
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 8/25/2008 4:11 AM:
Since -isystem is some burden on the gnulib user (not a big one, but anyway)
I propose to add
#ifdef __GNUC__
# pragma GCC system_header
#endif
to all gnulib headers that use @[EMAIL
I compile my code with -pedantic, because I want it to work with compilers
other than GCC. This means that my compiler output is littered with warnings
about #include_next. How can I stop this? It's a pain to read through; of
course I can grep out the warnings, but that's that's an annoyance
Hello,
Reuben Thomas wrote:
I compile my code with -pedantic, because I want it to work with compilers
other than GCC. This means that my compiler output is littered with warnings
about #include_next.
gnulib is clever enough to use #include_next only with compilers that support it
(i.e. gcc
On Sun, 24 Aug 2008, Bruno Haible wrote:
Hello,
Reuben Thomas wrote:
I compile my code with -pedantic, because I want it to work with compilers
other than GCC. This means that my compiler output is littered with warnings
about #include_next.
gnulib is clever enough to use #include_next only
On Sun, Aug 24, 2008 at 10:52 PM, Reuben Thomas [EMAIL PROTECTED] wrote:
Sounds interesting, but I can't find it. Have you a pointer to where this
comes from? I can't find it in any obvious place.
It was attached to the email to which you were replying.
James.
On Sun, 24 Aug 2008, James Youngman wrote:
On Sun, Aug 24, 2008 at 10:52 PM, Reuben Thomas [EMAIL PROTECTED] wrote:
Sounds interesting, but I can't find it. Have you a pointer to where this
comes from? I can't find it in any obvious place.
It was attached to the email to which you were
Reuben Thomas wrote:
it'd need some way for gnulib to turn it off, and
gnulib would then have to use it.
gnulib cannot avoid the use of #include_next. On non-glibc platforms it
would be possible, by use of #include absolute_system_header_filename,
but with glibc it is not possible, because
On Mon, 25 Aug 2008, Bruno Haible wrote:
Reuben Thomas wrote:
it'd need some way for gnulib to turn it off, and
gnulib would then have to use it.
gnulib cannot avoid the use of #include_next. On non-glibc platforms it
would be possible, by use of #include absolute_system_header_filename,
but
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Reuben Thomas on 8/24/2008 4:47 PM:
gnulib cannot avoid the use of #include_next. On non-glibc platforms it
would be possible, by use of #include absolute_system_header_filename,
but with glibc it is not possible, because glibc itself
Eric Blake wrote:
But obviously, glibc has some way of marking a header file as a system
header, so that the use of extensions such as #include_next do not trigger
gcc -pedantic warnings. Is there a #pragma that glibc uses to do that?
And should gnulib do the same?
The code that emits this
14 matches
Mail list logo