[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#183477: libc6's conflict against php4 is pretty much bogus and useless

> -----Original Message-----
> From: GOTO Masanori [mailto:gotom@debian.or.jp] 
> "The versions of php4 in stable and testing (4:4.1.2-X) don't exhibit
> this problem" is brought by glibc's __libc_fork ad-hoc patch.  Broken
> usage of __libc_fork in php4 and wine should be fixed in sarge
> completely.

No, no.  The php4 and wine bugs weren't related in any way, as far as I can
recall.  The versions of php4 in stable were never affected with this bug,
even with the first version of libc6 2.3.1, as the problem seemed to stem
from the way php4 was linked to libssl, and those versions weren't linked
to it at all.

The bug was never "fixed" in libc6, but by Steve Langasek in php4 with the
upload of php4 4:4.2.3-5.  See the bug log for #166414.

> Well, removing the patch about __libc_fork is after releasing sarge,
> and I agree we have to conflict it at the last.  So I change my mind
> that it should keep to conflict.

There's little point in the conflict.  The only php4 versions with the bug
were ones that don't exist anymore.  4:4.2.3-1 through 4:4.2.3-4, to be
precise.  Anything before 4:4.2.3-1 is (and always was) fine, and versions
after 4:4.2.3-4 are also just fine.  Given that one can't find that range
of versions of php4 anywhere in the archive anymore, and anyone who's been
tracking stable has probably had php4 forecefully upgraded for them by the
libc6 conflict exsiting, I really don't see the point a broad conflict that
will break partial upgrades from stable (4:4.1.2-6 in stable -WILL- work
with libc6 2.3.1, even 2.3.1-1)

... Adam

Reply to: