You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@apr.apache.org by Branko Čibej <br...@xbc.nu> on 2003/02/14 00:21:56 UTC
Re: cvs commit: apr/include/arch/unix apr_arch_file_io.h
trawick@apache.org wrote:
>trawick 2003/02/13 05:24:48
>
> Modified: file_io/unix filepath.c filepath_util.c
> include/arch/unix apr_arch_file_io.h
> Log:
> get the prototypes for these functions in the right place
> so they actually do some good
>
> > filepath_util.c:70: warning: no previous prototype for \
> `apr_filepath_list_split_impl'
> > filepath_util.c:107: warning: no previous prototype for \
> `apr_filepath_list_merge_impl'
>
Excuse me? your compiler is broken. A function definition is a
prototype, so you don't need an extra prototype in this case. But I
agree that the prototypse should live in a better place.
> 1.2 +9 -0 apr/include/arch/unix/apr_arch_file_io.h
>
> Index: apr_arch_file_io.h
> ===================================================================
> RCS file: /home/cvs/apr/include/arch/unix/apr_arch_file_io.h,v
> retrieving revision 1.1
> retrieving revision 1.2
> diff -u -r1.1 -r1.2
> --- apr_arch_file_io.h 7 Jan 2003 00:52:54 -0000 1.1
> +++ apr_arch_file_io.h 13 Feb 2003 13:24:48 -0000 1.2
> @@ -153,6 +153,15 @@
>
> mode_t apr_unix_perms2mode(apr_fileperms_t perms);
> apr_fileperms_t apr_unix_mode2perms(mode_t mode);
> +apr_status_t apr_filepath_list_split_impl(apr_array_header_t **pathelts,
> + const char *liststr,
> + char separator,
> + apr_pool_t *p);
> +apr_status_t apr_filepath_list_merge_impl(char **liststr,
> + apr_array_header_t *pathelts,
> + char separator,
> + apr_pool_t *p);
> +
>
>
This file is not used on Windows. Upon reflection, I think apr_private.h
wouldbe the right place for these prototypes. And
file_io/win32/filepath.c should be changed, too, not just its Unix
counterpart.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
Re: cvs commit: apr/include/arch/unix apr_arch_file_io.h
Posted by Branko Čibej <br...@xbc.nu>.
Branko Čibej wrote:
>Branko Čibej wrote:
>
>
>
>>I'll move those protos to apr_private.h now.
>>
>>
>>
>>
>Hm, no I won't... somehow, I was under the impression that apr_private.h
>wasn't arch-specific. Duh.
>
>Would anyone object if I added a file include/arch/apr_private_common.h
>for private stuff that's not architecture-specific? I could include that
>file from the various apr_private.h instances. I suspect there are more
>things that could go into a common private header.
>
>
Since there were no objections, I went ahead and made this change. Every
apr_private.h now includes this platform-independent private header. For
now, apr_private_common.h only contains the apr_filepath_list_split_impl
and apr_filepath_list_merge_impl prototypes, but more stuff can probably
move there over time.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
Re: cvs commit: apr/include/arch/unix apr_arch_file_io.h
Posted by Branko Čibej <br...@xbc.nu>.
Branko Čibej wrote:
>I'll move those protos to apr_private.h now.
>
>
Hm, no I won't... somehow, I was under the impression that apr_private.h
wasn't arch-specific. Duh.
Would anyone object if I added a file include/arch/apr_private_common.h
for private stuff that's not architecture-specific? I could include that
file from the various apr_private.h instances. I suspect there are more
things that could go into a common private header.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
Re: cvs commit: apr/include/arch/unix apr_arch_file_io.h
Posted by Branko Čibej <br...@xbc.nu>.
Cliff Woolley wrote:
>On Fri, 14 Feb 2003, [UTF-8] Branko Čibej wrote:
>
>
>
>>Excuse me? your compiler is broken. A function definition is a
>>prototype, so you don't need an extra prototype in this case. But I
>>agree that the prototypse should live in a better place.
>>
>>
>
>gcc -Wstrict-prototypes causes this. We look for these on purpose. :)
>
>
Oh, I'm not saying it's good style, only that it's valid C. :-)
I'll move those protos to apr_private.h now.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
Re: cvs commit: apr/include/arch/unix apr_arch_file_io.h
Posted by Cliff Woolley <jw...@virginia.edu>.
On Fri, 14 Feb 2003, [UTF-8] Branko Čibej wrote:
> Excuse me? your compiler is broken. A function definition is a
> prototype, so you don't need an extra prototype in this case. But I
> agree that the prototypse should live in a better place.
gcc -Wstrict-prototypes causes this. We look for these on purpose. :)
--Cliff