From: Jean-Pierre A. <jea...@wa...> - 2009-12-09 20:50:07
|
Hi Dominique, > Message du 09/12/09 21:02 > De : "Dominique Leuenberger" > A : "Jean-Pierre André" > Copie à : ntf...@li... > Objet : Re: [ntfs-3g-devel] compiling ntfs-3g with -Werror > > > Thank you very much for this patch. I'll work with this one. > > I got a question though regarding this snippet: > > On Wed, 2009-12-09 at 18:49 +0100, Jean-Pierre André wrote: > > @@ -67,8 +70,8 @@ > > #else > > #define USESTUBS 0 /* direct calls to API, based on following > > definitions */ > > #define ENVNTFS3G "NTFS3G" > > -#define LIBFILE64 "/lib64/libntfs-3g.so.4921" > > -#define LIBFILE "/lib/libntfs-3g.so.4921" > > +#define LIBFILE64 "/lib64/libntfs-3g.so.491" > > +#define LIBFILE "/lib/libntfs-3g.so.491" > > Is it correct that the soNUM is being decreased? Especially interesting > for me in this regard is that the 2009.11 release actually ships with > libntfs-3g.so.71, so I'm not sure what exactly this number refers to. You need not bother about that. This is skipped in your situation, only needed for stand-alone execution. > > -#define VERSION "1.1.2" > +#define VERSION "1.1.3" > > This part should not be a problem, as it's just the version of this > file. Well, I increment the version for any change. Useful when bug reports come back (note that this tool can be used out of ntfs-3g context). > The remainings of this patch are clear too. > > OTher than that I'm happy that this can be solved together with upstream > development. Thank you very much for this!! > > Besides this, I still carry for the moment a patch against fusermount.c, > which. This tool I assume is not used on Windows, and as such the patch > might be suitable for you: This looks good (well, acceptable) to me, but this file comes from the fuse projet, and I am not a reference for this. > > Index: ntfs-3g-2009.11.14/libfuse-lite/fusermount.c > =================================================================== > --- ntfs-3g-2009.11.14.orig/libfuse-lite/fusermount.c > +++ ntfs-3g-2009.11.14/libfuse-lite/fusermount.c > @@ -588,6 +588,7 @@ static int mount_fuse(const char *mnt, c > &source, &mnt_opts); > > if (currdir_fd != -1) { > + __attribute__((unused))int ignored_fchdir_status = > fchdir(currdir_fd); > close(currdir_fd); > } > > > As with all the others, of course, I think it would be good to have it > in a way so everybody can profit. > Regards Jean-Pierre |