Re: Potentially serious problem with kernel-headers...
>>Not tried this. I can't reproduce the problem locally. zif? Dale?
First off, serious thanks for doing this for me.
>Running kernel 2.0.34:
>
>Running cdparanoia compiled with 2.0.33 headers: crash
>Running cdparanoia compiled with 2.0.34 headers: success
As expected, OK...
>Running kernel 2.0.33:
>
>Running cdparanoia compiled with 2.0.33 headers: success
>Running cdparanoia compiled with 2.0.34 headers: success
>
>So the answer appears to be that the new binary *does* run under the old
>kernels.
This makes me think some important struct got bigger (think about it).
I'm going to say about a 15% probability of it... I'm going to read
all the kernel diffs tonight.
This is *still* a serious problem. Think of it: You have to update
every compiled binary on your machine to go from 2.0.33->2.0.34 safely
(including libc). This would explain the huge volume of complaints
I've gotten from various forums of 'things randomly segfaulting in
2.0.34 with glibc when 2.0.33 works fine'. The Linux Emacs people at
MIT have been scratching their heads over this all month.)
I'm going to try to nail this down for certain tonight. If someone
did indeed come up with a non-compatible interface change, there is
gonna be some pennance to pay.
Grrr. This is exactly the kind of thing that will eventually drive me
to NetBSD...
Monty
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: