Z
Zed Shaw
Hi folks,
I just put a pre-release of Mongrel up for people to try. You can
install it with:
gem install mongrel --source=http://mongrel.rubyforge.org/releases/
And try out your application and confirm the stuff works.
USE RFUZZ AND RUBY
If you find a bug, I'd appreciate it if you can go grab RFuzz (or use
net/http) and work up a little bit of Ruby that demonstrates the bug.
If you do this then I'll include it in the RFuzz test suite I'm doing
and the bug will never show it's ugly head again.
Same goes for problems with command line options. Just work up a little
ruby that starts mongrel the way you start it, tests for the error, then
stops mongrel.
CHANGE LOG
Knocked out a bunch of bugs people reported so far:
* Makes Mongrel act like WEBrick re: REQUEST_URI, PATH_INFO,
SCRIPT_INFO, and QUERY_STRING. This should fix redirects, but in order
to make it work I've added REQUEST_PATH as well which is technically
PATH_INFO+SCRIPT_INFO.
* Mongrel now includes *all* of the requested URI even if Mongrel
doesn't have a clue what it should do with it. Let's hope we don't see
tons of requests for proxy crap (like * and full hostnames) to Rails
apps. Could get interesting.
* Expanded some of the size limits that Mongrel enforces to accommodate
the new REQUEST_URI changes. REQUEST_URI can now be 2k. If it's over
2k then you need to use a FORM POST you dumbass.
* REQUEST_PATH can only be 1k. (only? man).
* Fixes the PID files and log files confusion. Mongrel now does
absolutely not transforming of your -P and -l settings (pid_file,
log_file in the configs). This means that it should work as expected,
and that everyone should use absolute paths if they're going to set it,
otherwise leave it to log/mongrel.log and log/mongrel.pid.
* DirHandler directory listings are more consistent thanks to a patch
from carmen whats-your-name (no relation to Carmen San Diego).
* Removed the extraneous ; in the C code that was choking the Sun C
compilers. Is there someone at Sun I can kick to death over this?
* Removes junk from the packaged gem such as the .so and test/coverage
files.
* Extra checks on the --prefix option to make sure they get it right.
It's a *path* prefix folks. That means it starts with a / but don't end
it in a /.
* SENDFILE IS BANISHED! If you have it installed Mongrel ignores it.
It causes way too many instability problems and isn't really worth the
trouble. Since I can't trust that everyone will remove it, I've just
ripped that code out.
* Dan Kubb fixed up the conditional responses to not use
Unmodified-Since.
* Fixes up the IO loop so that it is more exact in it's processing and
error handling.
* The start of a full RFC compliance test using RFuzz. This test suite
basically breaks down the RFC, fires up Rails or other web apps, and
then makes sure they can follow the RFC. So far I've got the
conditional response stuff going, a few lame duck things about rails
tested, and the layout of sections 3 and 4 set. Contributors are
welcome on this.
Enjoy and report those bugs. The sooner I get 0.3.13.4 out and happy
the sooner I can get 0.4 done.
I just put a pre-release of Mongrel up for people to try. You can
install it with:
gem install mongrel --source=http://mongrel.rubyforge.org/releases/
And try out your application and confirm the stuff works.
USE RFUZZ AND RUBY
If you find a bug, I'd appreciate it if you can go grab RFuzz (or use
net/http) and work up a little bit of Ruby that demonstrates the bug.
If you do this then I'll include it in the RFuzz test suite I'm doing
and the bug will never show it's ugly head again.
Same goes for problems with command line options. Just work up a little
ruby that starts mongrel the way you start it, tests for the error, then
stops mongrel.
CHANGE LOG
Knocked out a bunch of bugs people reported so far:
* Makes Mongrel act like WEBrick re: REQUEST_URI, PATH_INFO,
SCRIPT_INFO, and QUERY_STRING. This should fix redirects, but in order
to make it work I've added REQUEST_PATH as well which is technically
PATH_INFO+SCRIPT_INFO.
* Mongrel now includes *all* of the requested URI even if Mongrel
doesn't have a clue what it should do with it. Let's hope we don't see
tons of requests for proxy crap (like * and full hostnames) to Rails
apps. Could get interesting.
* Expanded some of the size limits that Mongrel enforces to accommodate
the new REQUEST_URI changes. REQUEST_URI can now be 2k. If it's over
2k then you need to use a FORM POST you dumbass.
* REQUEST_PATH can only be 1k. (only? man).
* Fixes the PID files and log files confusion. Mongrel now does
absolutely not transforming of your -P and -l settings (pid_file,
log_file in the configs). This means that it should work as expected,
and that everyone should use absolute paths if they're going to set it,
otherwise leave it to log/mongrel.log and log/mongrel.pid.
* DirHandler directory listings are more consistent thanks to a patch
from carmen whats-your-name (no relation to Carmen San Diego).
* Removed the extraneous ; in the C code that was choking the Sun C
compilers. Is there someone at Sun I can kick to death over this?
* Removes junk from the packaged gem such as the .so and test/coverage
files.
* Extra checks on the --prefix option to make sure they get it right.
It's a *path* prefix folks. That means it starts with a / but don't end
it in a /.
* SENDFILE IS BANISHED! If you have it installed Mongrel ignores it.
It causes way too many instability problems and isn't really worth the
trouble. Since I can't trust that everyone will remove it, I've just
ripped that code out.
* Dan Kubb fixed up the conditional responses to not use
Unmodified-Since.
* Fixes up the IO loop so that it is more exact in it's processing and
error handling.
* The start of a full RFC compliance test using RFuzz. This test suite
basically breaks down the RFC, fires up Rails or other web apps, and
then makes sure they can follow the RFC. So far I've got the
conditional response stuff going, a few lame duck things about rails
tested, and the layout of sections 3 and 4 set. Contributors are
welcome on this.
Enjoy and report those bugs. The sooner I get 0.3.13.4 out and happy
the sooner I can get 0.4 done.
