On Fri, 28 May 2004, Stas Bekman wrote:
> Date: Fri, 28 May 2004 11:27:46 -0700
> From: Stas Bekman <[EMAIL PROTECTED]>
> To: Jie Gao <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED]
> Subject: Re: [mp2] 1.99_14: socket problem
>
> Jie Gao wrote:
> >
> >
> > On Wed, 26 May 2004, Stas Bekman wrote:
> >
> >
> >>Date: Wed, 26 May 2004 19:08:40 -0700
> >>From: Stas Bekman <[EMAIL PROTECTED]>
> >>To: Jie Gao <[EMAIL PROTECTED]>
> >>Cc: [EMAIL PROTECTED]
> >>Subject: Re: [mp2] 1.99_14: socket problem
> >>
> >>Jie Gao wrote:
> >>[...]
> >>
> >>>>>>If you've moved to Apache 2.0.49, it now gives you a non-blocking socket. So
> >>>>>>all protocol handlers now must start with:
> >>>>>>
> >>>>>>use APR::Const -compile => 'SO_NONBLOCK';
> >>>>>>$sock->opt_set(APR::SO_NONBLOCK => 0);
> >>>
> >>>
> >>>Is this supported?
> >>>
> >>> $socket->opt_set(APR::SO_KEEPALIVE, 1);
> >>
> >>I haven't tried, but if it's a valid option, why not? Have you tried doing
> >>that and it didn't work? Though I won't advise on doing that for HTTP stuff. I
> >>guess it's OK for your own protocols.
> >
> >
> > That's what I am doing. It's just that the doco did not specify any valid
> > values but "xxxx'.
>
> You mean here, right?
> http://perl.apache.org/docs/2.0/api/APR/Const.html#C__socket_
>
> I haven't used any other constants, but APR::SO_NONBLOCK so I didn't document
> them? Care to send a patch fixing that? It'd be also good to test those before
> documenting. e.g. to figure out how APR::SO_KEEPALIVE behaves just grep the
> apr source include dir for APR_SO_KEEPALIVE, like so:
>
> % cd httpd-2.0/
> % grep -Ir APR_SO_KEEPALIVE srclib/apr/include
> srclib/apr/include/apr_network_io.h:#define APR_SO_KEEPALIVE 2 /**<
> Keepalive */
> srclib/apr/include/apr_network_io.h: * APR_SO_KEEPALIVE -- keep
> connections active
> srclib/apr/include/apr_network_io.h: * APR_SO_KEEPALIVE -- keep
> connections active
>
> looking in that file gives us:
>
> * <PRE>
> * APR_SO_DEBUG -- turn on debugging information
> * APR_SO_KEEPALIVE -- keep connections active
> * APR_SO_LINGER -- lingers on close if data is present
> * APR_SO_NONBLOCK -- Turns blocking on/off for socket
> * APR_SO_REUSEADDR -- The rules used in validating addresses
> * supplied to bind should allow reuse
> * of local addresses.
> * APR_SO_SNDBUF -- Set the SendBufferSize
> * APR_SO_RCVBUF -- Set the ReceiveBufferSize
> * </PRE>
>
> it doesn't document what the possible values are :( But you can guess that
> most accept 0/1 to turn the option off/on, the last two are the size of the
> buffer in bytes and I'm not sure about APR_SO_REUSEADDR.
>
> So now you have almost all the information to at least submit a doc patch
Thanks very much. I'll do when I find some time for it. Right now I'm getting
segmentation faults and am going to debug it today.
>
> > BTW, I am getting this in the log:
> >
> > Undefined subroutine &Apache::Connection::AUTOLOAD called.\n,
> >
> > It seems that some mod_perl module is trying to do autoload a sub that
> > does not exist.
>
> I don't think the mod_perl 2.0 core does. Set:
>
> use Carp;
> $SIG{__WARN__} = \&Carp::cluck;
>
> at the server startup and it'll tell you who called that.
No, it doesn't give me more info. The lines concerned are:
my $content_type = $r->lookup_uri($r->uri)->content_type;
if (defined $content_type) {
print STDERR "content-type is: $content_type\n";
return Apache::OK if $content_type =~ m:^$content_type_to_ignore/:i;
}
But I don't see anything involves autoload here.
Regards,
Jie
--
Report problems: http://perl.apache.org/bugs/
Mail list info: http://perl.apache.org/maillist/modperl.html
List etiquette: http://perl.apache.org/maillist/email-etiquette.html
