Some interesting stuff from PLD

Russ Allbery rra at stanford.edu
Mon May 5 19:24:28 UTC 2008


Julien ÉLIE <julien at trigofacile.com> writes:
> (And thanks, Marco, for your answer too.)
>
>>> -clienttimeout:          600
>>> +clienttimeout:          1800
>>
>> Eh.  I've never had ctlinnd take anywhere near the timeout.
>
> In the man page, it is written that "some newsreaders use the same
> connection for reading and posting and don't deal well with the
> connection timing out while a post is being composed.  If the
> system isn't having a problem with too many long-lived connections,
> it may be a good idea to increase this value to 3600 (an hour)."

Oh, I misread that completely, sorry.  I thought that said ctlinndtimeout.
Yes, increasing the client timeout is somewhat more reasonable, although I
think there's also something in the RFC about that.

> All right.  And I see that there are less files in CURRENT:
>
> AC_CHECK_HEADERS([crypt.h inttypes.h limits.h pam/pam_appl.h stddef.h \
>                   stdint.h string.h sys/bitypes.h sys/filio.h sys/loadavg.h \
>                   sys/time.h unistd.h])

Yeah, I removed a bunch of files that are now checked for automatically at
some point.  More could be removed (stddef, stdint, inttypes, unistd, and
I think string are all part of AC_STDC_HEADERS).

> Some of the missing files are checked elsewhere but for instance
> sys/select.h is not in configure.  I do not know whether it could be a
> problem.

We should definitely be checking for sys/select.h.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.


More information about the inn-workers mailing list