On Thu, Mar 31, 2005 at 05:56:40PM -0800, Steve Langasek wrote:
> > On Wed, Mar 30, 2005 at 06:06:28PM +0200, Michael Vistein wrote:
> > > Package: e2fsprogs
> > > Version: 1.37-1
> > > Severity: grave
> > > Justification: renders package unusable
> 
> > This is a problem only on IA64, and only breaks mke2fs, and not the
> > other e2fsprogs programs.  The problem seems to be the malloc() goes
> > wonky if we don't include stdlib.h.  The following patch fixes the
> > problem.  I'll upload a fixed set of sources very shortly.
> 
> Does "only breaks mke2fs" mean it doesn't affect mkfs.ext3?  The path seems
> to suggest that it affects the library, which would seem to imply mkfs.ext3
> is also affected; which would also seem to break the installer in the
> default configuration.  That sounds RC to me.

mkfs.ext3 and mke2fs are the same program (they are hard-linked
together).

> Though hopefully it doesn't affect the version in testing at least, given
> that no one has reported the bug before now...

No, this is new to 1.37-1, where the function of taking an OS type and
converting it to a string was moved to the e2p library as a helper
function.  Only on IA-64 is it the case that malloc() will return an
invalid pointer and core dump if you don't #include stdlib.h (probably
because ints are 32 bits and pointers are 64 bits on IA-64), which is
why I didn't notice it during my own development and testing.  

So as long as you don't build an installer based on the version of
e2fsprogs 1.37-1, you're fine.  

I don't think this quite rises to the level of "grave" --- the rest of
programs in e2fsprogs wwork just fine, so it doesn't render the entire
e2fsprogs package "unusuable or mostly so", since only people who need
to create new filesystems are going to be hit by this.

Granted I'd be in much bigger hurry if this were for a more popular
architecture than the Itanic --- but fear not, I will get a fixed
release out by next week.

                                                - Ted


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to