On Fri, Aug 10, 2001 at 03:28:59PM +0530, Nikunj A. Dadhania wrote:
> here is the patch
Thanks. A correction:
> diff -ruN stat-2.5.orig/stat.c hurd/stat/stat-2.5.orig/stat.c
> --- stat-2.5.orig/stat.c Thu May 10 13:25:29 2001
> +++ hurd/stat/stat-2.5.orig/stat.c Fri Aug 10 14:48:27 2001
> @@ -23,11 +23,11 @@
> #ifdef __USE_FILE_OFFSET64
> printf("%s %x %x %lu %lx %lld %lld %lld %ld %lld %lld\n",
You need a change here, too ^^^^^.
> #else
> - printf("%s %x %x %d %x %ld %ld %ld %d %ld %ld\n",
> + printf("%s %Lx %Lx %d %x %d %d %d %d %d %d\n",
No need to print the thing twice, once is enough.
fsid does not change with USE_FILE_OFFSET64 (it's not related to files at
all).
> #endif
> filename,
> - statfsbuf.f_fsid.__val[0],
> - statfsbuf.f_fsid.__val[1],
> + statfsbuf.f_fsid,
> + statfsbuf.f_fsid,
And only one of these two.
> statfsbuf.f_namelen,
> statfsbuf.f_type,
> statfsbuf.f_blocks,
> @@ -42,78 +42,162 @@
>
> printf(" File: \"%s\"\n", filename);
> #ifdef __USE_FILE_OFFSET64
> - printf(" ID: %-8x %-8x Namelen: %-7ld Type: ",
> statfsbuf.f_fsid.__val[0], statfsbuf.f_fsid.__val[1], statfsbuf.f_namelen);
> + printf(" ID: %-16Lx %-16Lx Namelen: %-7ld Type: ", statfsbuf.f_fsid,
> statfsbuf.f_fsid, statfsbuf.f_namelen);
> #else
> - printf(" ID: %-8x %-8x Namelen: %-7d Type: ",
> statfsbuf.f_fsid.__val[0], statfsbuf.f_fsid.__val[1], statfsbuf.f_namelen);
> + printf(" ID: %-16Lx %-16Lx Namelen: %-7d Type: ", statfsbuf.f_fsid,
> statfsbuf.f_fsid, statfsbuf.f_namelen);
> #endif
Same here, only one time each. The 16 was only a guess, it is needed to make
the
column align. I suggest a test run to find the real value (17 = 8 + 1 + 8?)
> - switch (statfsbuf.f_type) {
> - case AFFS_SUPER_MAGIC:
Why did you reindent the whole thing?
> + case MINIX_SUPER_MAGIC:
> + printf("MINIX\n");
> + case MINIX_SUPER_MAGIC2:
Here is a bug in the upstream code? The entry misses a break;
> #ifdef __USE_FILE_OFFSET64
> printf("UNKNOWN (0x%lx)\n", statfsbuf.f_type);
> #else
> @@ -124,8 +208,8 @@
> printf("Blocks: Total: %-10lld Free: %-10lld Available: %-10lld Size:
> %ld\n", statfsbuf.f_blocks, statfsbuf.f_bfree, statfsbuf.f_bavail,
> statfsbuf.f_bsize);
> printf("Inodes: Total: %-10lld Free: %-10lld\n", statfsbuf.f_files,
> statfsbuf.f_ffree);
> #else
> - printf("Blocks: Total: %-10ld Free: %-10ld Available: %-10ld Size:
> %d\n", statfsbuf.f_blocks, statfsbuf.f_bfree, statfsbuf.f_bavail,
> statfsbuf.f_bsize);
> - printf("Inodes: Total: %-10ld Free: %-10ld\n", statfsbuf.f_files,
> statfsbuf.f_ffree);
> + printf("Blocks: Total: %-10d Free: %-10d Available: %-10d Size: %d\n",
> statfsbuf.f_blocks, statfsbuf.f_bfree, statfsbuf.f_bavail, statfsbuf.f_bsize);
> + printf("Inodes: Total: %-10d Free: %-10d\n", statfsbuf.f_files,
> statfsbuf.f_ffree);
> #endif
Why did you touch this code? long is correct for these.
> @@ -145,14 +229,14 @@
> }
>
> if (terse != 0) {
> - printf("%s %u %u %x %d %d %x %d %d %x %x %d %d %d %d\n",
> + printf("%s %u %u %x %d %d %Lx %d %d %x %x %d %d %d %d\n",
> filename,
> (unsigned int)statbuf.st_size,
> (unsigned int)statbuf.st_blocks,
> statbuf.st_mode,
> statbuf.st_uid,
> statbuf.st_gid,
> - (int)statbuf.st_dev,
> + statbuf.st_dev,
> - printf("Device: %xh/%dd\tInode: %-10d Links: %-5d",
> (int)statbuf.st_dev, (int)statbuf.st_dev,
> - (int)statbuf.st_ino, (int)statbuf.st_nlink);
> + printf("Device: %Lxh/%dd\tInode: %-10d Links: %-5d",
> statbuf.st_dev,(int)statbuf.st_dev,
> + statbuf.st_ino, (int)statbuf.st_nlink);
Ayupp, we have a problem here. st_dev is dev_t which is int on Linux, but
on the Hurd, it is the same as fsid (long). But why did you remove the cast
to int of statbuf.st_ino?
It is possible to determine the size of the field in C. When you do that
and set a variable st_dev_len to 4/8 and then switch on that, do you think
this would be a good solution? Otherwise, you can also just do
#ifndef (__GNU__)
Original code
#else
Your version
#endif
Which is far simpler ;)
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' Debian http://www.debian.org [EMAIL PROTECTED]
Marcus Brinkmann GNU http://www.gnu.org [EMAIL PROTECTED]
[EMAIL PROTECTED]
http://www.marcus-brinkmann.de
