In article <3ngu42$e87@newshost.lanl.gov>,
Rob Hartill <hartill@lanl.gov> wrote:
>franl@centerline.com (Francis Litterio) wrote:
>>I'm running NCSA's HTTPD and the error_log file is showing lots and
>>lots of entries like this:
>
>> [Fri Mar 10 19:14:42 1995] killing CGI process 26264
>
>>As best as I can tell, each of these errors is logged at or about the
>>same time the access_log shows someone making a request via a
>>clickable imagemap.
>>
>>Has anyone else seen this?
>
>yes
>
>> What can I do to fix it?
The messages are caused by kill_children every time the imagemap redirect a
URL. I just commented out the kill_children in redirect code in
http_script.c. I didn't see any side-effect...
>upgrade, e.g. to Apache. http://www.apache.org/apache/
I tried apache (0.62 beta) last weekend and immediately found a
bug/misfeature also related to redirect (not in knowbugs). The apache home
page suggests we post questions/bug reports regarding apache to
comp.infosystems.www.servers.unix -- it doesn't seem to exist on our nntp
server... So here it goes:
Everything seems to be fine with scripts in scriptaliased directory, eg,
/cgi-bin/, however if the script is outside cgi-bin, eg, index.cgi, the
bug surfaces when the index.cgi just does a redirection by printing a
location header: it will prepend content-type: application/x-httpd-cgi
to screw up every browser in sight. A workaround is to print a redundant
content-type: text/html header besides the location header. NCSA httpd
1.3R doesn't have this quirk...
__Luke
Rob Hartill <hartill@lanl.gov> wrote:
>franl@centerline.com (Francis Litterio) wrote:
>>I'm running NCSA's HTTPD and the error_log file is showing lots and
>>lots of entries like this:
>
>> [Fri Mar 10 19:14:42 1995] killing CGI process 26264
>
>>As best as I can tell, each of these errors is logged at or about the
>>same time the access_log shows someone making a request via a
>>clickable imagemap.
>>
>>Has anyone else seen this?
>
>yes
>
>> What can I do to fix it?
The messages are caused by kill_children every time the imagemap redirect a
URL. I just commented out the kill_children in redirect code in
http_script.c. I didn't see any side-effect...
>upgrade, e.g. to Apache. http://www.apache.org/apache/
I tried apache (0.62 beta) last weekend and immediately found a
bug/misfeature also related to redirect (not in knowbugs). The apache home
page suggests we post questions/bug reports regarding apache to
comp.infosystems.www.servers.unix -- it doesn't seem to exist on our nntp
server... So here it goes:
Everything seems to be fine with scripts in scriptaliased directory, eg,
/cgi-bin/, however if the script is outside cgi-bin, eg, index.cgi, the
bug surfaces when the index.cgi just does a redirection by printing a
location header: it will prepend content-type: application/x-httpd-cgi
to screw up every browser in sight. A workaround is to print a redundant
content-type: text/html header besides the location header. NCSA httpd
1.3R doesn't have this quirk...
__Luke