httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Slemko <>
Subject Re: Agenda for 1.2b7
Date Sat, 01 Feb 1997 23:10:19 GMT
On Sat, 1 Feb 1997, Ben Laurie wrote:

> I see no mention of the lingering_close() problem. I propose that, as we had
> already agreed, that lingering_close() is made _not_ the default. I will do the

-1 unless someone shows me why everything I have been spouting for the
past few weeks is wrong.

I was in favor of disabling it untill someone (randy?) said something that
make me start thinking, and Roy followed up a day or so later with
<> in which he describes some of the
reasons why it is needed.  I think the reasons go further than what he
describes in that message, ie. it doesn't only apply to error conditions.  

I'm not going to repost my reasons for thinking this again and again, but
if what I've said in previous messages doesn't make sense plesae bring up
specific examples of what you think is wrong or don't understand.  I could
be crazy.

I _think_ we could manage disabling lingering_close for non-persistent
non-HTTP/1.1 non-keepalive connections if we modify the method of error
responses a bit.  I'm not sure that is worth it. 

> work, given some +1s. I'd still like to know why it didn't happen last time
> round? I'd also still like to know where Randy's pointer copying then freeing
> code is.

support/suexec.c, in clean_env().  Just removing the free() looks good to
me.  It would be nice to get that and perhaps the couple of other small
bugfixes for suexec together soon and put a patch on the web for those
with problems so people can test it. 

View raw message