Activity (RSS)

Applied in http://github.com/macournoyer/th...

thx Wincent!

Attaching patch.

aaarg nooo

what are the names of those processes? ruby, swifitply or "Thin test server" ?

are there any phantom ruby processes? I used to have that problem all the time, but I haven't compiled lately.

I got all specs running on both Mac OS 10.5 and an Ubuntu VM.

But there's a couple phantom process left after running the specs on Ubuntu, I'll try to fix tha...

Only 1 spec still fails now on my new Ubuntu VM.

Can you run the specs again and see if it's better for you now?

Applied http://github.com/macournoyer/th...

Also I loaded cgi_multipart_eof_fix if Ruby 1.8.5, reopen if you have a better solution

thx for the patch!

http://github.com/macournoyer/th...

I moved all perf specs under spec/perf and exclude them from rake spec. Run through rake spec:perf now.

thx for the suggestion Wincent!

I fixed a couple on master. I still get 4 failures on my Ubuntu VM.

Most seem to be due to Swiftiply not beeing installed. I'll make those conditional.

Also the "Thin::Server app builder should work with Rack url mapper" failu...

Ok, I mentioned in ticket #50 that some specs were failing for me on RHEL. Just to confirm that it's not a machine-local issue I tried to reproduce on Mac OS X...

hey Wincent if you would like to help w/ this I can give you a github invite so you can fork the Thin repo. It's much better then patch ping-pong.

Just send m...

Sure all specs pass on my machine, I'm a big TDD/BDD fan ;)

benchmark_unit is supposed to be machine independent but I guess it's not 100% correct.

I like th...

I've been trying the Thin 0.7.0 specs out on RHEL 5.1 and I'm seeing what I would consider to be a bunch of spurious failures. For example:

12)

'Thin::Server...

I see no other way then looking at the Ruby version number (apply of < 1.8.6).

I don't think it you cause any trouble if we require the gem even if the Ruby v...

nice job Wincent, I'll check that out and maybe require the cgi_multipart_eof_fix if needed.

Ruby 1.8.5 compatibility came up in this thread over at Google groups:

http://groups.google.com/group/t...

Here is a patch making Thin buildable on 1.8.5:

h...

Full request was stored in @data, that was causing the very high memory usage. It is not released as soon as the parsing is complete. Mem usage is now lower th...

i'm uploading a 350Mb linux iso and with mongrel my process stays at 20Mb with thin it grows to 370Mb

reported by emobius on irc channel

Provide an adapter to make only some urls use deferrable.

Something usable like this in a rackup file:

use Thin::Defer, '/my/url'
run YouAdapter.new

And a...

Include Merb (0.9) and Ramaze (maybe other) Rack adapter and select correct one based on dir structure or something else.

So it can be as easy as w/ Rails: ju...

we need EM to release for jruby before we can do anything.

Maybe related to this thread - can the following be done for thin (to have

[port or socket/number of requests currently queued/being processed concurrently/nu...

yay that's an awesome idea Kevin.

Should be possible even w/ an adapter.

I would MUCH rather see output like this: http://www.igvita.com/2008/02/19....

Perhaps we should make this a config option somehow?

I''m using Ragel 5.24, which I installed from source.

This is clearly doable and since the parser doesn't change that much.

But I wonder how EM performs on j...

I'm still trying to get Ruby allowed at work, and JRuby is probably the runtime I'd have to use. Count me a +1 vote.

EM (trunk) has a java version, Rack seems...

Long requests (2-3 sec) might block the event loop.

We could use EventMachine Deferrable to process some request in threads. But this impact speed. So we need...

just committed a fix for it, still lots of failing specs but they seem to be windows-specific weirdness

That won't fix it, as the 'require' happens at load time and will fail if 'daemons' gem isn't loaded.

I've got the code to fix it, but there are 30 or 40 othe...

Committed in http://github.com/macournoyer/th...

Let me know if it's rly fixed.

I think adding:

include Daemonizable unless Thin.win?

in Thin::Server, will fix this

I was not able to reproduce

Was it fixed w/ http://groups.google.com/group/t... ?

If not, plz reopen the ticket.

http://www.ruby-forum.com/topic/...

i installed thin thin 0.6.4 (x86-mswin32-60) and eventmachine 0.8.1

(i386-mswin32)

i went to my script folder under rai...

example file now support unix and tcp sockets

god leaks mem so I rly don't think it's a good idea to integrate more w/ thin.

Will be solved by next EM gem 0.11.0, should be out soon ...

I tailed the log file to get the output

Yes, we could have an options to do a simple rolling restart, i.e. making sure only one thin server is down at any time. Would you like a patch for that?

You mean something like this http://blog.carlmercier.com/2007... ?

Unfortunately, a colleague from work tells me that it's likely to suffer from memory leaks, even if you kick of the garbage collector manually. So restarting i...

That definitely require some testing before including. But it looks very interesting.

thx for the patch Alex