In the case of (2), would it be possible to force a thread attempting to
use a feature in a library that was in the process of loading to block
until the library had finished? That seems like it would solve the only
potential problem with (2).
I agree that (3) has all kinds of potential deadlo
That looks to me like a pathing issue, which mirrors a similar issue
that we've had reported on the Glassfish user list. I'd guess offhand
that that's where this problem's root is.
Brandon Hauff (JIRA) wrote:
Glassfish Gem 0.9.0 does not work with JRuby 1.1.5
--
Charles Oliver Nutter wrote:
2. I posted a sample of what static compilation to a "real live Java
class" might look like here: http://gist.github.com/37000. The idea is
that you'd decorate the class with a little extra Java information,
and then a compiler could create this Java class that expo
Charles Oliver Nutter wrote:
So, thoughts?
- Charlie
First thought: Very cool.
Second thought: Runtime method definitions? _Very_ cool.
It will be great to see what kind of interesting new things this opens
up for us. It might also fit in well with the annotation-based classes
that were sug
In the Java, you're making an anonymous class that defines your
GrizzlyAdapter, which you aren't doing in the Ruby. I'm not sure how/if
JRuby lets you define anonymous classes, but if you made a separate Ruby
class that extended GrizzlyAdapter (and implemented service()), you
should be able to
while I was working on Glassfish Gem issue 24787
(http://rubyforge.org/tracker/index.php?func=detail&aid=24787&group_id=5450&atid=21080),
I think I've narrowed it down to an issue with changing the load_path
for embedded instances of Jruby.
In the first (working) case, of mongrel + Jruby,
RAI
bring in config/environment.rb, and our rack handler, then call
call(env) on whatever rails hands to our rack handler. We tried to stay
as hands-off with the Ruby codes as possible, to try to avoid weird
things like this.
Charles Oliver Nutter wrote:
Charles Oliver Nutter wrote:
Jacob Kessler
I've tried with both boot.rb and environment.rb, the behavior seems
identical. Environment.rb requires boot.rb, so identical behavior would
make sense.
Charles Oliver Nutter wrote:
Jacob Kessler wrote:
I'm a bit fuzzy on gem vs. stand-alone startup, but I don't think
that t
What error is it giving?
If obj.getShape actually is returning an object with a getRadius method,
the code should work as written. If obj.getShape doesn't return an
object with a getRadius method, it will end up calling the
method_missing method, which by default throws an error. I think, anyw
1.1.6
Environment: Ubuntu 8.04, Glassfish V3 trunk, Rails 2.1.1, Jruby
1.1.6RC1
Reporter: Jacob Kessler
Assignee: Thomas E Enebo
Using Jruby 1.1.5, I'm able to get a fairly consistent 80 ops/second per
glassfish instance
Using Jruby 1.1.6RC1, I'm able to
: JRuby 1.1.7
Environment: Jruby 8302, Glassfish trunk
Reporter: Jacob Kessler
Attachments: JrubyLeakAfter, JrubyLeakBefore
Jruby seems to be leaking a weakreference for each processed request from
Glassfish. This appears to be related to the RC2 hardReference bug, but
idea
Key: JRUBY-
URL: http://jira.codehaus.org/browse/JRUBY-
Project: JRuby
Issue Type: Bug
Reporter: Jacob Kessler
Assignee: Thomas E Enebo
--
This message is automatically generated by JIRA.
-
If
ation
Affects Versions: JRuby 1.1.6
Environment: Jruby 1.1.7 b8810, Java 1.6.0_07 (Hotspot 10.0-b23)
Reporter: Jacob Kessler
a call to Ruby.tearDown() from within Java doesn't seem to properly unregister
the Mbeans that Jruby uses, resulting in a memory leak.
--
This
: Ubuntu 8.0.4, Jruby 1.1.7 r8810, jruby-openssl 0.3,
net-ssh 2.0.8
Reporter: Jacob Kessler
Issue 2747 (Net::SSH io padding error) seems to have resurface in some version
of net-ssh between 2.0.4 and 2.0.8. version 2.0.4 works fine, version 2.0.8
throws the same error as in issue 2747
Affects Versions: JRuby 1.1.6
Environment: JRuby 1.1.6
Reporter: Jacob Kessler
When the JRuby OpenSSL gem is loaded inside another environment (such as an
application server), it does not properly unregister itself with the security
manager on shutdown, causing a per-deploy
15 matches
Mail list logo
