FastCGI Accept never returns

[GoogleCodeExporter]

On win32 platform, the Accept() function of jsfastcgi module should return
before the web-server kills the process.

See the related topic on apachelounge forum:
http://www.apachelounge.com/forum/viewtopic.php?t=1845

Original issue reported on code.google.com by sou...@gmail.com on 21 Aug 2007 at 10:29

[GoogleCodeExporter]

FastCGI can be done without libfastcgi.

Original comment by sou...@gmail.com on 5 Jun 2008 at 10:00

• Added labels: Priority-Medium
• Removed labels: Priority-High

[GoogleCodeExporter]

Now, the idea is to stop using libfastcgi and to write the fastcgi layer by 
hand and
to support only TCP connections.
see. http://fastcgi.com/mod_fastcgi/docs/mod_fastcgi.html#FastCgiExternalServer

Note for lighttpd:
http://trac.lighttpd.net/trac/wiki/TracFastCgi
warning: add "check-local" => "disable"

Original comment by sou...@gmail.com on 24 Jul 2008 at 11:33

[GoogleCodeExporter]

This issue is minor, and like I said, FastCGI protocol can be managed in 
JavaScript.

Original comment by sou...@gmail.com on 27 Feb 2009 at 3:58

• Added labels: Priority-Low
• Removed labels: Priority-Medium

[GoogleCodeExporter]

http://svn.nadir-point.com/viewvc/monkeyscript/trunk/libraries/FastCGI.js?revisi
on=44&view=markup
Partial FastCGI support in MonkeyScript legacy. I'm having issues with TCP
unfortunately (It actually has enough support to handle a single simple request,
however for some reason I'm getting a TCP exception and it's stopping the test
request). MonkeyScript is going to to through a full API rewrite, but if anyone
reading wants to experiment and find out what's wrong with FastCGI I'd be happy 
to
help. That'll make release of a small version of MonkeyScript that's actually 
usable
possible after cleaning up the API.

Original comment by nadir.se...@gmail.com on 9 Mar 2009 at 1:45