> spelling: it should be "one piece" not "one peice"
Ok
>Very interesting changes. Have you discussed this with upstream?
No
>If not, that's fine; no need to slow down the package update in this
>case. But be sure to eventually.
Ok, i'll send them the patch.
typo? Should this be:
#ifndef i_reserved2
not
#ifndef i_reseved2
Yes
This is an important issue, and you should fix it.
It's not actually all that important, I put the ifndef in to only apply
the hack when nessacery but I don't beleive applying the hack
unnessacerally when using older versions of the headers will actually
break anything. Nevertheless i'll fix it.
Embarasingly this got pointed out when I first submitted the patch but
in the long time between doing that and finally preparing a NMU I fogot
about it :(
I've replaced the NMU in the queue with one that has the typos fixed.
New debdiff attatched.
diff -ur ext3grep-0.10.1/src/ext3.h ext3grep-0.10.1.new/src/ext3.h
--- ext3grep-0.10.1/src/ext3.h 2008-04-09 11:20:31.000000000 +0000
+++ ext3grep-0.10.1.new/src/ext3.h 2011-12-10 07:00:50.000000000 +0000
@@ -24,6 +24,17 @@
#ifndef EXT3_H
#define EXT3_H
+// this trickery needs to happen before ext2_fs is included so
+// bail out if it has already been included by another path
+#ifdef _LINUX_EXT2_FS_H
+ #error please include this file before any other includes of ext2fs/ext2_fs.h
+#endif
+
+// some versions of the ext2 headers call this s_frags_per_group and some
+// call it s_clusters_per_group, define one to the other so our code works
+// with both
+#define s_clusters_per_group s_frags_per_group
+
// Use the header files from e2progs (http://e2fsprogs.sourceforge.net)
// We can use these headers and then everything named ext2 or ext3.
#include <ext2fs/ext2_fs.h> // Definitions of ext2, ext3 and ext4.
@@ -110,6 +121,12 @@
__u32 faddr(void) const { return i_faddr; }
__u16 uid_high(void) const { return i_uid_high; }
__u16 gid_high(void) const { return i_gid_high; }
+#ifndef i_reseved2
+ //i_reseved2 has been split into two fields in recent
+ //versions of the headers, luckilly we can still access
+ //it in one peice through the hurd side of the union
+ #define i_reserved2 osd2.hurd2.h_i_author
+#endif
__u32 reserved2(void) const { return i_reserved2; }
void set_reserved2(__u32 val) { i_reserved2 = val; }