Re: RFS: php-memcached
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello Sergey,
On 20.11.2011 07:44, Sergey B Kirpichev wrote:
> I'm unsure if it's RC at all. I'll try to fix this that way
> on next upload, but it seems better to
> 1) lower #629896's priority to important
this does not make the problem disappear. I'd agree with Jossip though
who claimed the issue would be release-critical. The fact you render the
package unusable on heavy loaded servers sounds pretty critical, doesn't
it?
> 2) document workarround's (e.g. see Stefan's reply, php-memcached also
> seems to be an option)
Yes, but that's not a solution either. Using fcgi is a completely
different approach which is untypical for Apache at least and
php-memcached uses a different API, right?
Hence, your suggested workarounds would either require to change the
application or the entire web server setup respectively.
> 3) leave real decision on how to fix to upstream
Well, it's not /really/ a big deal. Just replace select() by poll().
That's the lazy solution. Possibly not the most effective since faster
(non-portable) alternatives exist (e.g. libev), but an easy one. Many
devices emulate select() by using poll's kernel interface anyway.
>> The problem occurs to both of your source packages, php-memcached and
>> php-memcache.
>
> I'm unaware. Libmemcached (in testing at least) uses poll, isn't?
Oh, I'm sorry my memories apparently were wrong. php-memcached indeed
doesn't use select() but libmemcached as you said.
- --
with kind regards,
Arno Töll
IRC: daemonkeeper on Freenode/OFTC
GnuPG Key-ID: 0x9D80F36D
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/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=bw6j
-----END PGP SIGNATURE-----
Reply to: