Matthew Burgess wrote:
> On Mon, 24 Jan 2011 20:06:39 -0800, Bryan Kadzban
> <[email protected]> wrote:
>
>> I'm not sure if that's the best setup; we'll have to make sure at each
>> glibc release (until the bug is fixed) that no new private-futex tests
>> are added. (That the sed is still equivalent to the patch.) It'd be
>> nice if the sed could take into account the context of each line, but I
>> believe that's rather difficult with sed.
>>
>> Anyone else have a preference?
>
> Given the relative complexity of Andrew's sed expression, and your concerns
> above,
> I'd rather we go with the patch.
Here are more specific, but a little clearer seds:
sed -i '213 s/PRIVATE_FUTEX/FUTEX_CLOCK_REALTIME/' \
nptl/sysdeps/unix/sysv/linux/x86_64/pthread_rwlock_timedrdlock.S
sed -i '195,210 s/PRIVATE_FUTEX/FUTEX_CLOCK_REALTIME/'
nptl/sysdeps/unix/sysv/linux/x86_64/pthread_rwlock_timedwrlock.S
or
sed -i '195,213 s/PRIVATE_FUTEX/FUTEX_CLOCK_REALTIME/' \
nptl/sysdeps/unix/sysv/linux/x86_64/pthread_rwlock_timed{rd,wr}lock.S
Actually, using one of these provides an educational example of how to
specify an address range for a sed.
The above would need to be checked for any new release, but we need to
do that anyway for a patch and checking/changing these seds would be easier.
-- Bruce
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
