* Time to use the new DSO sorting as default?
@ 2021-10-21 19:29 Adhemerval Zanella
2021-10-21 19:35 ` Florian Weimer
0 siblings, 1 reply; 3+ messages in thread
From: Adhemerval Zanella @ 2021-10-21 19:29 UTC (permalink / raw)
To: Libc-alpha, Florian Weimer, Chung-Lin Tang, Carlos O'Donell,
H.J. Lu, Joseph Myers
Now that we have the new DSO sorting algorithm (thanks Chung-Ling Tang), the
question is if and when make it the default. I don't tested it extensively
besides running the testsuite on multiple different architectures, but the
testsuite now checks a lot of more permutations and allows us to craft and
check regression quick.
Carlos and Florian, do you think we should enable it as default so you can
pickup on Fedora and check if something breaks?
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Time to use the new DSO sorting as default?
2021-10-21 19:29 Time to use the new DSO sorting as default? Adhemerval Zanella
@ 2021-10-21 19:35 ` Florian Weimer
2021-10-21 19:54 ` Adhemerval Zanella
0 siblings, 1 reply; 3+ messages in thread
From: Florian Weimer @ 2021-10-21 19:35 UTC (permalink / raw)
To: Adhemerval Zanella
Cc: Libc-alpha, Chung-Lin Tang, Carlos O'Donell, H.J. Lu, Joseph Myers
* Adhemerval Zanella:
> Now that we have the new DSO sorting algorithm (thanks Chung-Ling Tang), the
> question is if and when make it the default. I don't tested it extensively
> besides running the testsuite on multiple different architectures, but the
> testsuite now checks a lot of more permutations and allows us to craft and
> check regression quick.
>
> Carlos and Florian, do you think we should enable it as default so you can
> pickup on Fedora and check if something breaks?
I think we should switch the default for glibc 2.35, yes.
(I'm not even sure if we need both implementations, TBH.)
Thanks,
Florian
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Time to use the new DSO sorting as default?
2021-10-21 19:35 ` Florian Weimer
@ 2021-10-21 19:54 ` Adhemerval Zanella
0 siblings, 0 replies; 3+ messages in thread
From: Adhemerval Zanella @ 2021-10-21 19:54 UTC (permalink / raw)
To: Florian Weimer
Cc: Libc-alpha, Chung-Lin Tang, Carlos O'Donell, H.J. Lu, Joseph Myers
On 21/10/2021 16:35, Florian Weimer wrote:
> * Adhemerval Zanella:
>
>> Now that we have the new DSO sorting algorithm (thanks Chung-Ling Tang), the
>> question is if and when make it the default. I don't tested it extensively
>> besides running the testsuite on multiple different architectures, but the
>> testsuite now checks a lot of more permutations and allows us to craft and
>> check regression quick.
>>
>> Carlos and Florian, do you think we should enable it as default so you can
>> pickup on Fedora and check if something breaks?
>
> I think we should switch the default for glibc 2.35, yes.
>
> (I'm not even sure if we need both implementations, TBH.)
I chatted briefly with Carlos and he started to spin a rawhide build/test
and I think once we have some confidence that it is ok and can make the
flip and remove the old implementation. I agree that there is no much
gain in keeping an implementation with worse expectations if the newer
on provides the same functionality.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-10-21 19:54 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-21 19:29 Time to use the new DSO sorting as default? Adhemerval Zanella
2021-10-21 19:35 ` Florian Weimer
2021-10-21 19:54 ` Adhemerval Zanella
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).