I just reverted those commits in my tree, and it still fails.
When running mvn package, you get an error like:
Failed to execute goal on project jruby-tests: Could not resolve
dependencies for project org.jruby:jruby-tests:jar:9000.dev: Failed to
collect dependencies at rubygems:rspec:gem:2.14.1
On 3 July 2013 16:31, Charles Oliver Nutter head...@headius.com wrote:
* cext handling?
Kill it! Kill it with fire!
As much as I don't like breaking backward compatibility in a minor
point release, cext is pretty broken as it stands right now, and if
the external project is ready to roll,
On 30 June 2013 00:13, Charles Oliver Nutter head...@headius.com wrote:
For now you can pull later jnr-posix and 'mvn install' to get he snapshot.
Shouldn't the snapshot build be available in Sonatype?
Looks like it is
I just pushed a jnr-posix snapshot, and its in the sonatype snapshots
dir. Not sure if/when those get propagated to maven central.
On 29 June 2013 06:06, Charles Oliver Nutter head...@headius.com wrote:
Oh and it looks like our next goal might be getting snapshot repo set
up, since travis
On 28 June 2013 05:21, kristian m.krist...@web.de wrote:
actually there following native jars missing in the above artifact:
ppc-AIX
sparcv9-SunOS
x86_64-FreeBSD
Ok, good to know. I think we have support for them in jffi, but not
sure why they're not in JRuby repo.
they are in
On 28 June 2013 07:23, kristian m.krist...@web.de wrote:
hi Wayne,
jnr/jffi has no issues :(
other repos in jnr do have issues though.
It should now. Not sure when/how it got turned off.
-
To unsubscribe from this list,
Just thought I'd let people know about a project I've been tinkering
with - XNI - https://github.com/wmeissner/xni
XNI is an attempt to provide a good ruby - C/C++ api for JRuby,
without the problems of the legacy MRI cext interface, nor having to
write extensions directly in JNI.
The basic
:
Major
Reporter:
Wayne Meissner
:
Major
Reporter:
Wayne Meissner
Critical
Reporter:
Project:
JRuby
Priority:
Major
Reporter:
Wayne Meissner
Looks like Charlie pushed a change to github only?
Anyway, I merged that and pushed it back to the jruby repo, so, they
*should* be in sync again.
No need to change your git repo.
On 27 June 2012 14:57, Yoko Harada yoko...@gmail.com wrote:
Hi,
I wonder JRuby source repo moved from jruby.org
Issue Type:
Bug
Assignee:
Wayne Meissner
Wayne Meissner created JRUBY-6669:
-
Summary: failure at test/test_kernel.rb:588 if jruby is in a
directory with mixed-case
Key: JRUBY-6669
URL: https://jira.codehaus.org/browse/JRUBY-6669
Project
Wayne Meissner created JRUBY-6599:
-
Summary: test/test_socket.rb spuriously fails
Key: JRUBY-6599
URL: https://jira.codehaus.org/browse/JRUBY-6599
Project: JRuby
Issue Type: Bug
Wayne Meissner created JRUBY-6584:
-
Summary: File.mtime() is slower than it should be
Key: JRUBY-6584
URL: https://jira.codehaus.org/browse/JRUBY-6584
Project: JRuby
Issue Type: Improvement
Wayne Meissner created JRUBY-6509:
-
Summary: Bogus LDFLAGS on MacOSX
Key: JRUBY-6509
URL: https://jira.codehaus.org/browse/JRUBY-6509
Project: JRuby
Issue Type: Bug
Components: C
Wayne Meissner created JRUBY-6452:
-
Summary: rb_class_of() calls static object method instead of
static long method
Key: JRUBY-6452
URL: https://jira.codehaus.org/browse/JRUBY-6452
Project: JRuby
Wayne Meissner created JRUBY-6451:
-
Summary: memory leak in rb_struct_new
Key: JRUBY-6451
URL: https://jira.codehaus.org/browse/JRUBY-6451
Project: JRuby
Issue Type: Bug
Affects Versions
Wayne Meissner created JRUBY-6439:
-
Summary: cext: Memory leak in String and Array handling
Key: JRUBY-6439
URL: https://jira.codehaus.org/browse/JRUBY-6439
Project: JRuby
Issue Type: Bug
How often does github fall over these days?
At least one major advantage to having github as the primary repo
(with a mirror on EC2 for the CI backup), is the ability to trivally
merge pull requests. It reduces the friction for contributions
significantly.
As for the JIRA - GH issues migration
FFI Structs no longer support bools
---
Key: JRUBY-5828
URL: http://jira.codehaus.org/browse/JRUBY-5828
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.6.2
Reporter: Wayne Meissner
Affects Versions: JRuby 1.6RC2, JRuby 1.6RC1, JRuby 1.5.6
Reporter: Wayne Meissner
Assignee: Wayne Meissner
Priority: Blocker
Fix For: JRuby 1.6RC3
org.jruby.ext.socket.RubyUnixSocket#recvfrom() allocates a fixed size native
buffer of 1024 bytes
On 17 January 2011 17:38, Hiroshi Nakamura nakah...@gmail.com wrote:
Hi,
On Mon, Jan 17, 2011 at 10:48, Shih-gian Lee shih@gmail.com wrote:
I found my own problem. I forgot to include ant.jar in my classpath.
That's what I'm doing every time before/after changing local branches
I believe it has all been relicensed as the jruby tri-license now.
On 26 September 2010 13:29, Hirotsugu Asari asari.r...@gmail.com wrote:
So, COPYING.cext says it is GPL v3. Has there been any discussion on the
implications of this license choice?
I'm mostly curious.
Hiro
On 16 September 2010 09:29, Yoko Harada yoko...@gmail.com wrote:
Probably create org/jruby/util/ArrayUtil and dump the implementation
of copyOf() in there.
Ah, so just a mistake. I got it.
Will someone work on this?
You may as well jump in and fix it yourself, if you have the time.
pty.rb should be converted to use posix_spawn()
---
Key: JRUBY-5078
URL: http://jira.codehaus.org/browse/JRUBY-5078
Project: JRuby
Issue Type: Bug
Reporter: Wayne Meissner
jnr-netdb requires java 6 to run
Key: JRUBY-4902
URL: http://jira.codehaus.org/browse/JRUBY-4902
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.5.1, JRuby 1.5
Reporter: Wayne
Could the weak ref wiring be done lazily, so the overhead is only
incurred when someone sets an ivar on a java object?
e.g. when a java object enters jruby, it gets a new lightweight
wrapper, but when someone does an ivar get/set on it, it looks for the
ivar holder for the java object in the weak
It might work better if you sent this to jruby-dev instead of duby-dev :-)
The only things I would like to see for RC2 are x86 and x86_64 JFFI
stubs for solaris, and the ppc-aix one.
I also need to integrate the powerpc-darwin build using lipo or some
other method.
On 22 April 2010 09:56,
FFI needs api for setting dlopen flags
--
Key: JRUBY-4696
URL: http://jira.codehaus.org/browse/JRUBY-4696
Project: JRuby
Issue Type: New Feature
Affects Versions: JRuby 1.4
Reporter: Wayne
On 9 March 2010 17:31, Vladimir Sizikov vsizi...@gmail.com wrote:
I think Wayne had some thoughts on how to build the launcher so that
it would be compatible with as much Linux distros as possible (by
carefully selecting the toolchain and compiler/lib versions).
The sweet spot I found for
Type: Bug
Affects Versions: JRuby 1.5
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Trivial
Perusing RubyIO.select_static(), it puts sockets that have OP_CONNECT set in
their SelectionKey#readyOps() flags into the read set instead of the write
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.5
The current release of jffi bundled with JRuby has binaries for Darwin (all
arches), i386-Linux, x86_64-Linux, i386-Windows, x86_64-Windows. It might be
useful to have binary builds for more esoteric
On 9 February 2010 18:47, Stephen Bannasch
stephen.banna...@deanbrook.org wrote:
While there are things I don't like about jira on codehaus I think the
collected set of jruby issues and comments are a great resource and having
used a bunch of other systems for issues I think Jira is the best.
Type: Bug
Affects Versions: JRuby 1.4
Reporter: Wayne Meissner
Assignee: Wayne Meissner
Priority: Trivial
Fix For: JRuby 1.5
From [http://github.com/ffi/ffi/issues#issue/27]
The code below works fine in MRI but not JRuby. The problem appears
Versions: JRuby 1.4
Reporter: Wayne Meissner
Assignee: Wayne Meissner
Priority: Minor
Fix For: JRuby 1.5
When using a :bool member in a struct, the size data will be wrong as it is
using a 32bit integer instead of an 8 bit integer for the field.
Fixed
Issue Type: Bug
Affects Versions: JRuby 1.4
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.5
This test fails on a fresh ubuntu 9.10 install, OpenSolaris (and possibly
others).
The relevant line from /etc/hosts is:
127.0.0.1
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.4
Reporter: Wayne Meissner
Assignee: Wayne Meissner
Priority: Trivial
Fix For: JRuby 1.5
java.util.zip.GZIPInputStream is not internally buffered, so every call to read
results
1.4
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.5
It is possible that FFI::IO.native_read can return bogus data for reads 4096
bytes in size. This is due to a bug in
ChannelStream#copyBufferedBytes(ByteBuffer).
Fixed in commit
Solaris 32bit stat struct too large
---
Key: JRUBY-4111
URL: http://jira.codehaus.org/browse/JRUBY-4111
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.4.0RC1
Reporter: Wayne Meissner
jffi in 1.4 needs update to 0.6.0.2
---
Key: JRUBY-4114
URL: http://jira.codehaus.org/browse/JRUBY-4114
Project: JRuby
Issue Type: Bug
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Affects Versions: JRuby 1.4.0RC1
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.4
FFI functions with variadic parameters (e.g. foo(int bar, ...) ) that return a
struct by value, do not get invoked correctly.
Fixed in commit
Issue Type: Bug
Affects Versions: JRuby 1.4.0RC1
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.4
FFI functions which return a struct by value and take non-primitive parameters
(e.g. String, struct, struct-by-value) do
Reporter: Wayne Meissner
Assignee: Vladimir Sizikov
test/test_system_error.rb contains hard-coded MacOS specific values for errno
values. This breaks on anything not macos, and should be wrapped in some form
of platform check.
--
This message is automatically
Environment: OpenSolaris 2009.06
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.4
ant test fails on master with the following error:
[junit] Testsuite: org.jruby.test.ScriptTestSuite
[junit] FAILED 19 expected [[AF_INET, 80
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Minor
Fix For: JRuby 1.4
AbstractMemory (the base for all Pointer ops) is missing get_array_of_uint64()
and put_array_of_uint64() operations
--
This message is automatically generated by JIRA
Issue Type: Bug
Affects Versions: JRuby 1.3.1
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.4
Leaving aside the wisdom or otherwise of trying to read 1G of data in one go,
JRuby fails to load a 1G file even when the jvm
sub-optimal IO#read after IO#getc
-
Key: JRUBY-3721
URL: http://jira.codehaus.org/browse/JRUBY-3721
Project: JRuby
Issue Type: Improvement
Affects Versions: JRuby 1.3RC2
Reporter: Wayne
Affects Versions: JRuby 1.3RC2
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Blocker
Fix For: JRuby 1.3
When a buffer is passed to a function either as a :buffer_out parameter, or
the buffer was allocated via Buffer.new_out
: Improvement
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Minor
Fix For: JRuby 1.4
In the case where a socket is opened, and only readpartial operations are used
on it, or the buffers are flushed, ChannelStream#readpartial will only ever
2009/5/14 Charles Oliver Nutter charles.nut...@sun.com:
Anyone have any objections to disabling id2ref when ObjectSpace is not
enabled? Right now, in order to support this, we have to start tracking a
weak reference to the object whenever you request #object_id. Since most
people don't
Type: Improvement
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Fix For: JRuby 1.x+
Currently, to read data from an IO stream into a FFI memory buffer, the data
must be first read into a ruby string instance, then copied into the FFI buffer.
The number
: Wayne Meissner
Assignee: Wayne Meissner
Priority: Blocker
Fix For: JRuby 1.3
jffi needs an updated release, including a maven release before jruby 1.3 can
be released. The version in the jruby tree currently is a snapshot.
--
This message
: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Blocker
Fix For: JRuby 1.2
It appears that the new FFI backend is not passing string parameters correctly,
as reported in http://kenai.com/jira/browse/RUBY_FFI-5
If you wrap the 'sql' string argument up
2009/2/20 Jochen Theodorou blackd...@gmx.org:
If you distribute your GPL incompatible program without the lib and the user
is required to install that lib to run your program then GPL nethusiast will
most probably still say your program has to be under GPL.
To argue that there could exist a
2009/2/18 Conrad Meyer kon...@tylerc.org:
I was under the impression that non-GPL code could not link statically or
dynamically with GPL libraries, and that this was one of the major differences
between the GPL and LGPL. This is certainly Stallman's opinion (he cites
That correct, with a small
Need FFI platform files generated for x86_64 linux
--
Key: JRUBY-3376
URL: http://jira.codehaus.org/browse/JRUBY-3376
Project: JRuby
Issue Type: Task
Reporter: Wayne Meissner
JRuby needs a new release of JFFI
-
Key: JRUBY-3378
URL: http://jira.codehaus.org/browse/JRUBY-3378
Project: JRuby
Issue Type: Task
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
1.1.7
Reporter: Wayne Meissner
The following bit of code works with ruby 1.8.7 and JRuby 1.1.6, but fails on
JRuby trunk
{noformat}
require tempfile
p Dir.tmpdir
{noformat}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one
8.10 beta, i386
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Minor
The following code at line 131 of test/testTime.rb fails on ubuntu:
ENV['TZ']='Europe/Helsinki'
test_equal +0300 EEST, Time.now.strftime('%z %Z')
(This was introduced
1.1.4
Reporter: Wayne Meissner
Assignee: Wayne Meissner
Priority: Minor
Ruby 1.9 defines __method__ and __callee__ that return the current method.
JRuby currently does not support these.
--
This message is automatically generated by JIRA.
-
If you think
IO#putc flushes on every char
-
Key: JRUBY-3008
URL: http://jira.codehaus.org/browse/JRUBY-3008
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.1.4
Reporter: Wayne Meissner
IO#putc ignores I/O errors
--
Key: JRUBY-3009
URL: http://jira.codehaus.org/browse/JRUBY-3009
Project: JRuby
Issue Type: Bug
Affects Versions: JRuby 1.1.4
Reporter: Wayne Meissner
Assignee
Reporter: Wayne Meissner
Assignee: Wayne Meissner
FFI MemoryPointer instances passed into a native method as a :pointer parameter
are always passed as NULL.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one
That bit should be fixed - it now uses 'msvcrt' as the default lib to
load from.
I also (maybe) fixed a couple of the benchmarks to run on windows -
chmod, getpid, umask - but I don't have windows, so I couldn't test
them.
2008/8/12 Praki [EMAIL PROTECTED]:
Hi,
I am trying to run
Currently its going to be LGPLv3.
Thats mostly so I can easily borrow code from e.g. JNA if/when I need
to to flesh out its capabilities - otherwise I will have to reinvent
the wheel, and I'm lazy, so I don't like having to do extra work
unless I have to. The aim of JFFI was performance vs JNA,
is brand new and there is not much documentation
about it exists yet (there are some samples in JRuby trunk, some
benchmarks, and a couple of libraries using FFI already, like zlib and
syslog, so you could look at them for the reference). Please also note
that the API is not yet finalized.
Wayne
org.jruby.util.RubyInputStream is unused
Key: JRUBY-2732
URL: http://jira.codehaus.org/browse/JRUBY-2732
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
[PATCH] Optimize RubyIO.getline() performance
-
Key: JRUBY-2689
URL: http://jira.codehaus.org/browse/JRUBY-2689
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
[PATCH] Replace StringBuffer with StringBuilder
---
Key: JRUBY-2682
URL: http://jira.codehaus.org/browse/JRUBY-2682
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
[PATCH] fgetpos does not need to flush buffers
--
Key: JRUBY-2663
URL: http://jira.codehaus.org/browse/JRUBY-2663
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
[PATCH] Improve bulk reads in ChannelStream
---
Key: JRUBY-2657
URL: http://jira.codehaus.org/browse/JRUBY-2657
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
: MacOS 10.5.3
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Attachments: bench_io_read.rb
When running a modified version of test/bench/bench_io_read.rb, the RSS of the
java process grows until it hits 500M (the specified limit).
At that point, each iteration
: MacOS 10.5.2
Reporter: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Minor
Attachments: umask.patch
When File.umask is called with no argument, it should return the current umask
without altering it - however, doing so on JRuby sets the umask to 0
Avoid double copying data in ChannelDescriptor#read()
-
Key: JRUBY-2423
URL: http://jira.codehaus.org/browse/JRUBY-2423
Project: JRuby
Issue Type: Improvement
Reporter: Wayne
: Wayne Meissner
Assignee: Thomas E Enebo
Priority: Trivial
org.jruby.util.collections.IdentitySet does not appear to be used outside of
its unit test, so can probably be pruned.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly
Make fields immutable where possible
Key: JRUBY-2417
URL: http://jira.codehaus.org/browse/JRUBY-2417
Project: JRuby
Issue Type: Improvement
Reporter: Wayne Meissner
Assignee
Slowdown is one reason, interfering with code is another.
On unixen, it installs/uninstalls a SIGSEGV and a SIGBUS handler every
time you invoke a native function or do memory copies, basically
anything that might touch native memory.
Some native code (possibly the jvm itself) uses SIGSEGV for
Components: Java Integration
Environment: Solaris
Reporter: Wayne Meissner
Priority: Minor
Kernel::fork() will probably have to use forkall() on Solaris to have the same
semantics as on Linux and MacOS where all threads are copied.
According to http://docs.sun.com/app
80 matches
Mail list logo