severity 441339 serious
reassign 441339
libc6-i386,lib32ncurses5,ia32-libs,fakeroot,ia32-libs-gtk,lib32asound2,lib32gcc1,lib32z1,fglrx-driver
thanks
sean finney <[EMAIL PROTECTED]> writes:
> On Monday 10 September 2007 01:57:07 pm Goswin von Brederlow wrote:
>> Bdale Garbee <[EMAIL PROTECTED]> writes:
>> > If you *really* want to file a release-critical bug about this
>> > somewhere, file it instead against dpkg for lack of multiarch support
>> > sufficient to allow obsoleting ia32-libs ...
>
> if you *really* want to get your point across, don't forget to cc the
> submitter...
From: Goswin von Brederlow <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
I did. Well, I TOed you and CCed the bug.
> but in any event, "lack of $feature to allow obsoleting $package" is not
> really justification for allowing $package to remain in violation of policy.
I wasn't talking about full multiarch. Just being able to link.
> also, i don't see how that rationale warrants lowering the severity
> to "wishlist". afaik "important" is the designated severity for policy
> violations that are not deemed to be rc. however i'm not going to get
> into a severity pingpong over it.
The other problem and reason for /emul/ia32-linux is
--- from linux-2.6.20.18/include/asm-ia64/namei.h ---
#define EMUL_PREFIX_LINUX_IA32 "/emul/ia32-linux/"
static inline char *
__emul_prefix (void)
{
switch (current->personality) {
case PER_LINUX32:
return EMUL_PREFIX_LINUX_IA32;
default:
return NULL;
}
}
The kernel has some special case handling for for /emul/ia32-linux
when the personality is set to 32bit that afaik does some implicit
chroot magic.
So at least on ia64 we can't just abandon /emul/ia32-linux. The best
we could do is link it somewhere else. But then what would we win? And
don't forget the trouble of changing a directory into a symlink.
I agree with bdale: "The important part to me is the wontfix tag"
And yet another small point. I think the glibc has more weight
deciding where to put the 32bit libs on amd64 than ia32-libs has. Get
that fixed and ia32-libs will follow (at least on amd64). IA64 might
never get around the kernel hack.
MfG
Goswin
PS: I reassigned it to all involved packages I remeber. They all have
to move or none.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
