Re: [PATCH] Hurd port

[Andrew Clausen]

On Sun, Jun 17, 2001 at 10:44:03PM -0500, Neal H Walfield wrote:
> This corrects (against 1.5.2-pre1) a bunch of problems with the Hurd
> port.  I have used this code and tested it on a file making partitions
> and filesystems and it _seems_ to work correctly.  I even used to code
> to make a partition and a filesystem on my hard drive; I held my breath,
> but it worked!  I do, however, warn you that I have not verified all of
> the code -- I think that Andrew changed some of the semantics between
> this and the last version.
> 
> Having said that, I had to use this hack:

Can you please attach patches, gzipped?  It stops them from getting
mangled.  (So, could you resend?  Thanks)

I get =20 and all kinds of stuff throughout :-(

(Mutt seems to be able to filter them to display, but doesn't give
an option to save the filtered email!)

> Index: libparted/Makefile.am
> ===================================================================
> RCS file: /home/neal/cvs/parted/libparted/Makefile.am,v
> retrieving revision 1.1.1.1
> diff -u -r1.1.1.1 Makefile.am
> --- libparted/Makefile.am     2001/06/17 14:12:56     1.1.1.1
> +++ libparted/Makefile.am     2001/06/18 03:27:49
> @@ -18,8 +18,6 @@
>                       exception.c             \
>                       filesys.c               \
>                       geom.c                  \
> -                     llseek.c                \
> -                     llseek.h                \
>                       libparted.c             \
>                       natmath.c               \
>                       disk.c                  \
> @@ -29,10 +27,14 @@
>                       disk_mac.c              \
>                       disk_pc98.c             \
>                       disk_sun.c
> +                     # llseek.c              \
> +                     # llseek.h              \ 

I guess llseek.[ch] should go into device_linux.c

>  
>  EXTRA_libparted_la_SOURCES = device_linux.c device_gnu.c
>  
>  libparted_la_LIBADD   =      @UUID_LIBS@                     \
> +                     -lstore                         \
> +                     -lshouldbeinlibc                \
>                       fs_ext2/libext2.la              \
>                       fs_fat/libfat.la                \
>                       fs_hfs/libhfs.la                \

Some autoconfusion required...  We should probably have @OS_LIBS@,
for OS-specific libs, rather than globally adding to LIB, so as
to avoid humongous command lines...

> This is because _GNU_SOURCE already defines __USE_GNU.  Maybe this can
> be autoconfized.

Autoconfused :P

Well, it never hurts to #define _GNU_SOURCE, right?  What does it
officially mean?

Thanks,
Andrew Clausen