On Fri, Oct 03, 2008 at 09:43:19PM -0400, Joe Malicki wrote:
> This faked-link2.diff (patch to fakeroot 1.5.10) passes tests and holds up
> through
> reasonable testing.
>
> It should fix all cases noted in this bug, which were probably all caused by
> ldconfig.
Forward-ported to 1.10 so I can
- "Clint Adams" <[EMAIL PROTECTED]> wrote:
> On Fri, Oct 03, 2008 at 07:11:08PM -0400, Joe Malicki wrote:
> > faked.diff was thoroughly tested. When I posted I noticed that it
> didn't cover
> > all of the cases and added faked-link.diff, and meant to mention
> that I hadn't
> > tried it yet
On Fri, Oct 03, 2008 at 07:11:08PM -0400, Joe Malicki wrote:
> faked.diff was thoroughly tested. When I posted I noticed that it didn't
> cover
> all of the cases and added faked-link.diff, and meant to mention that I hadn't
> tried it yet (I am testing it now, with the {).
Okay, I've uploaded t
Ugh, should never have posted faked-link.diff without trying it, it's
completely busted.
Respinning and fixing.
faked.diff is well tested.
- "Clint Adams" <[EMAIL PROTECTED]> wrote:
> On Fri, Oct 03, 2008 at 06:22:44PM -0400, Joe Malicki wrote:
> > +if ((buf->st.mode&S_IFMT) != (st->mod
Err, thanks.
I got distracted and didn't finish and screwed up the original message as well.
faked.diff was thoroughly tested. When I posted I noticed that it didn't cover
all of the cases and added faked-link.diff, and meant to mention that I hadn't
tried it yet (I am testing it now, with the {
On Fri, Oct 03, 2008 at 06:22:44PM -0400, Joe Malicki wrote:
> +if ((buf->st.mode&S_IFMT) != (st->mode&S_IFMT) &&
> +((buf->st.mode&S_IFMT) != S_IFREG || (st->mode&(S_ISBLK|S_ISCHR
Missing { here?
> +
> + fprintf(stderr,"FAKEROOT: chmod mode=%lo incompatible with "
>
> The first patch (faked-link.diff) fixes it in all cases except when
> the *new* inode type is a regular file.
> The second fixes it for more cases, wherever this sort of scheme is
> possible (i.e. everywhere but inodes that are
> regular files in our database but device nodes on disk).
>
Patch
- "Joe Malicki" <[EMAIL PROTECTED]> wrote:
> We've root caused this bug being related to statically linked binaries
> (usually ldconfig) which don't let fakeroot intercept their syscalls.
>
> When ldconfig is run and modifies /etc/ld.so.cache, it creates a new
> file and replaces it. Howeve
We've root caused this bug being related to statically linked binaries (usually
ldconfig) which don't let fakeroot intercept their syscalls.
When ldconfig is run and modifies /etc/ld.so.cache, it creates a new file and
replaces it. However, the old ld.so.cache never gets purged from the fakeroo
9 matches
Mail list logo
