I am having a problem using the RootDirectory chroot
with mathopd.
When I try to run mathopd with this option set I get:
mathopd: Invalid user name.
I've looked in the mail archives but no luck. The docs
deal with php and so on, but not chroot! Unless the
mail archives simply don't go
Hi.
A new beta has been released at www.mathopd.org. Get it while it's hot.
The new beta now supports - drumroll - the `Range' header, which means
that browsers should now be able to resume a previously interrupted
download, whereas previously they would have to start from scratch.
I have only
Does Mathopd support something like a .access file in the same directory
to check for permissions?
Not at the moment.
Cheers
Michiel
Maybe this has something to do with the error. I have php5 and the path to
the php file seems to be correct. PHP installed also some other files like
phpize, phpextdist. I removed them. I don't know if they cause the
problem.
Right. I had not tested mathopd with php5. PHP is known to break
Sat Dec 27 22:15:52 2003 [20889] accept: Software caused connection abort
Sat Dec 27 22:15:53 2003 [20889] accept: Software caused connection abort
... same ... untill my scripts restart mathopd.
Try the attached patch. Note: this is totally untested: I am away from my
Hi.
I just put out 1.5p2. This is a tiny update with respect to p1. If you
already run p1 without problems there is no reason to upgrade.
The p2 version will print a warning if NumConnections is set too high.
This should be useful for troubleshooting if you run into 'too many open
files' errors.
Can I start mathopd from xinetd?
No. Running from inetd is not very efficient, because you would have to
re-read the configuration file for every request.
Cheers
Michiel
The most recent MathOpd doesn't seem to support HTTP 0.9 Simple-Request
messaqes. Is it optional now, or completely ripped out?
It's been ripped out. HTTP/0.9 was obsolete even at the time I started
development on Mathopd (1995).
Hi. Sorry about the delay.
On Mon, Oct 25, 2004 at 07:38:23PM +0800, Xuefer wrote:
yes! i can't use ExtraHeaders! why?
with status Not Modified mathopd is NOT sending ExtraHeaders
this make IE send request again and again, get NO cache-control header :(
the status bar showing downloading xx
On Sun, Dec 12, 2004 at 02:08:52AM +0800, Albert wrote:
I was just wondering, if there is a list of features or items of
pursuit for the 1.6.x version of mathopd vs. the 1.5.x.
The wish-list at the moment is
- FastCGI
- SSL
- kqueue/epoll/etc support
- .htaccess-like thingies
None of these
Hi. The machine that hosts the mathopd web site and this mailing lists
will be taken down at the end of the year. I am trying to migrate the site
somewhere else, but I don't think I will be able to set up a new mailing
list, at least not for a while.
So I think what will happen is this: this
Hi. I am pleased to inform you that the mathopd mailing list is back up.
Many, many thanks to Anton Berezin for helping me out.
Cheers
Michiel
Eek. The dir_cgi.c on the mathopd site is rather old. It should use
SCRIPT_NAME, rather than PATH_INFO. I have replaced it. The new version
should work ok.
Thanks, it works wonderfully now! Question, though: Why should it use
SCRIPT_NAME instead of PATH_INFO? Or, I guess, what was the purpose of
[...]
it's important to close requests asap.
I've looked through the documentation but i'm not sure i have seen a
reference to such a feature, does it exist ? Is this done using Timeout, or
Wait ?
There is no knob in mathopd to turn off keepalives. The server will
kill connections that are
On Fri, 25 Feb 2005, Martin Johnson wrote:
[...]
But, as a better solution, I guess it must also be possible somehow to
send arbitrarily-long CGI script output through Mathopd, in the style of
NCSA httpd's non-parsed-header CGI interface. For NCSA http you just
had to name your CGI script nph-
Hi. I just released mathopd 1.5p5 and 1.6b6.
The new version fixes a problem in internal_dump() that could be exploited to
append to arbitrary files writable by the server's UID. Thanks to Carsten Eiram
for pointing this out.
Internal dumps will now create files in /tmp with unpredictable
Hi. It has come to my attention that mathopd 1.5 with the AIO patch on the
website does not work quite well. I have fiddled a bit with aio over the
weekend and have now prepared a new patch, this time for 1.6.
Cheers
Michiel
NumConnections=512 nfiles=1024
warning: NumConnections is too high
Why do I get this message ? Can i expect problems from this ?
Hi. The nfiles number (1024) indicates how many file descriptors your
mathopd process can use. Mathopd uses two descriptors per connection, plus
a few extras for
Hi.
This is a PHP problem, not something to do with IndexNames.
But if I fetch the http://.co.uk/ with a web browser, I get the
message:
No input file specified.
It appears you are using the CLI version of PHP instead of the CGI
version. Try looking for php-cgi somewhere under
I just ran into mathopd via a benchmark page where
mathopd and zeus were the top 2 performing web
servers.
Nice to hear :-) url of that page ?
My guess is that the original poster was referring to Jef Poskanzer's Web
Server Comparisons page (google for thttpd :). But that is a bit outdated.
For CGI programs that want to use ETag values (e.g. MD5 hashes) to
avoid re-sending dynamically-generated content that hasn't changed since
the last time the client requested it, it's useful to have access to the
HTTP header If-None-Match.
This HTTP header isn't passed in the environment from
I would think that the mattoclf utility is sort-of in the public
domain. So you can use that utility any way you see fit. I would add a
notice to that file but it seems a bit wasteful to have a file where the
license takes up more space then the code. :)
Cheers
Michiel
Hi. Sorry about the delay.
On Tue, 23 Jan 2007, Germain Leutwyler wrote:
Please I have one question regarding mathopd error logging: I get regularly
in the error.log file the two following lines:
Mon Jan 22 11:10:51 2007 [12970] child process 28869 exited with status 6
Status 6 means that
I've checked all my PERL scripts for a wrong or missing #!/usr/bin/perl: all
are OK !
My guess is that you have something like this in your config file
Control {
Specials {
CGI { pl }
}
}
for this to work every perl script needs to be executable. That's a bit of
a
Hi. I have released a new stable version of mathopd. This contains some
bugfixes, plus one feature that has been merged from 1.6.
1) On FreeBSD there was a rather embarrassing bug when sendfile was used,
which manifests itself in bogus 'premature end of file' messages in the
error log. This
Hi. It was brought to my attention that the distfile for mathopd-1.5p6
contained CVS directories, which does not make any sense.
So I have decided to re-generate mathopd-1.5p6.tar.gz without the CVS
directories.
Sorry about this. Next time I will be more careful. Remind me to not
do anything just
On Thu, 20 Mar 2008, Dmitry wrote:
Hello!
Sometimes my server statistics looks like this:
Uptime: 1243972 seconds
Active connections: -61 out of 500
Max simultaneous connections since last dump: -36
Forked child processes: 3034661
Exited child processes: 3034661
I'am worrying about minus
The only thing i am missing in mathopd is a forwarding proxy function.
I need one directory been forwarded to another webserver and the answer
routed back. e.g. thisdomain.com/fwd should be forwarded to
internal-ip/fwd/app.
What is the best solution/add on to realize this behavior with mathopd?
Rabindra Paikaray wrote:
Hi All,
I want to know is mathopd supports file upload to server or not?
Recently I was trying to upload the file , but it was failing .
I don't know is it require to configure properly or it does not support
at all.
Mathopd by itself does not support file uploading,
Dmitry wrote:
Hello,
I need to debug mathopd process because it segfaults every 2 day. The
pattern is always the same. Something like mathopd[24068]: segfault at
a17b000 ip 005be8ac sp bfd6098c error 4 in libc-2.7.so[54d000+153000]
I am sure it is not the problem of mathopd itself, but
Dmitry wrote:
Michiel Boland wrote:
Hi. Core dumps are only enabled if you specify a directory in the
config file to dump core in. You can do this using the CoreDirectory
directive. Make sure the user you specify in the config file can
create files in this directory.
Michiel,
I forgot
Hi. I just released a security update for mathopd. (mathopd 1.5p7)
The problem:
If you use the * construct in your config, as in
Control {
Alias /
Location /var/www/*
}
then the * will be substituted with the value of the Host header that was
supplied by the client. However this
Hi. The 1.5p7 release contains yet another embarassing bug that causes a crash
when a HTTP/1.0 request without a Host header is made.
There is a new stable and a new beta. If you are running 1.5p7 by any chance
please upgrade immediately.
Sorry about this. I should have known better than to
Hi. Just a short note to indicate that this list is still alive. :)
I will be migrating this list from majordomo to mailman somewhere during this
month. There will be another message when the migration is complete.
Cheers
Michiel
Hi. If you see this message then that means the migration of the mathopd mailing
list to mailman has been completed.
Welcome to the 21st century. :)
Cheers
Michiel
On 11/12/2012 20:25, David J. Weller-Fahy wrote:
* Michiel Boland mich...@boland.org [2012-11-12 12:12 -0500]:
Hi. If you see this message then that means the migration of the
mathopd mailing list to mailman has been completed.
Welcome to the 21st century. :)
Oooo... it's so shiny
Hi. I am going to shut down the mathopd mailing list. There has been zero
traffic for the last few years, contrasted to about 1 spam messages that I
have to deal with.
Cheers
Michiel
37 matches
Mail list logo
