On Fri, 2008-09-12 at 13:43 +0100, Ian Rogers wrote:
As the patch comes from Jikes RVM it should be no surprise we addressed
solving the performance problem in a metacircular manner. I imagine
DaCapo jython is a lot faster on Classpath runtimes now.
It is:
2008-09-10: = DaCapo jython
On Sun, 2008-08-17 at 00:24 +0100, Andrew John Hughes wrote:
Finally add the Scanner implementation from Laszlo.
Finally :-)
- twisti
Hi!
This small changes fix the build on OpenSolaris. It additionally checks
for sys/sockio.h and libsocket.so. Andrew, can you try them on your
Solaris system?
- twisti
---
Index: configure.ac
===
RCS file:
On Mon, 2008-07-07 at 09:40 +0100, Andrew John Hughes wrote:
2008/7/7 Christian Thalinger [EMAIL PROTECTED]:
Hi!
This small changes fix the build on OpenSolaris. It additionally checks
for sys/sockio.h and libsocket.so. Andrew, can you try them on your
Solaris system?
- twisti
-
@@ -1,3 +1,9 @@
+2008-07-01 Christian Thalinger [EMAIL PROTECTED]
+
+ * java/lang/Integer.java (signum): Implemented properly as
+ described in Hacker's Delight Section 2-7, plus Andrew Haley's
+ explanation.
+
2008-07-01 Andrew Haley [EMAIL PROTECTED]
* java
Hi!
This patch renames the JNINativeInterface_ variable in _Jv_JNIEnv to be
compatible with OpenJDK. Without that it's not possible to compile C++
code with both.
OK to commit?
- twisti
---
Index: include/jni.h
===
RCS file:
On Sun, 2008-06-22 at 10:15 -0600, Tom Tromey wrote:
Twisti == Christian Thalinger [EMAIL PROTECTED] writes:
Twisti This patch renames the JNINativeInterface_ variable in
Twisti _Jv_JNIEnv to be compatible with OpenJDK. Without that it's
Twisti not possible to compile C++ code with both
On Fri, 2008-06-06 at 00:52 +0100, Andrew John Hughes wrote:
Are you sure CACAO is using the most up-to-date version of Classpath
given this works with JamVM? If buf is being altered, a new array will
be created inside CPStringBuffer. If a new buffer is not being
created, we should fix the
On Mon, 2008-06-09 at 12:18 +0200, Christian Thalinger wrote:
On Fri, 2008-06-06 at 00:52 +0100, Andrew John Hughes wrote:
Are you sure CACAO is using the most up-to-date version of Classpath
given this works with JamVM? If buf is being altered, a new array will
be created inside
Hi!
I think I found a bug in GNU Classpath, in java/text/ChoiceFormat.java.
The following Mauve test fails with CACAO (but works with JamVM?!?,
maybe a VM interface issue):
$ echo gnu/testlet/java/text/ChoiceFormat/parse | cacao RunnerProcess
On Mon, 2008-04-28 at 20:53 +0100, Andrew John Hughes wrote:
On 28/04/2008, Christian Thalinger [EMAIL PROTECTED] wrote:
Hi!
This patch disables the compilation of dependencies when using the
-compile no option. When the dependency class does not exist it prints
a FAIL message like
Hi!
This patch disables the compilation of dependencies when using the
-compile no option. When the dependency class does not exist it prints
a FAIL message like:
FAIL: java.lang.Character.unicode
One of the dependencies is not compiled.
We need this patch when we cross-compile Mauve for
On Mon, 2008-04-21 at 09:03 +0200, Robert Schuster wrote:
+ public boolean isUp()
+throws SocketException
+ {
+return VMNetworkInterface.isUp(netif.name);
+ }
+ public boolean isPointToPoint()
+throws SocketException
+ {
+return VMNetworkInterface.isUp(netif.name);
On Mon, 2008-04-21 at 11:43 +0100, Andrew Haley wrote:
Why use a short and not an int?
for (int i=MIN_VALUE; i = MAX_VALUE; i++)
would be more idiomatic and faster on many targets.
Same as I suggested. This would also apply to:
--- java/lang/Character.java19 Dec 2006 01:14:23
On Mon, 2008-04-21 at 09:03 +0200, Robert Schuster wrote:
Hi,
people using Jalimo asked for those methods and I found them simple to
implement.
Since I am lousy non-Java hacker I appreciate comments on the code in
java_net_VMNetworkInterface.c.
I wonder if it would be simpler to only have
On Mon, 2008-04-21 at 14:34 +, Andrew John Hughes wrote:
Indeed you did suggest it, but didn't give me a good reason to do it... ;)
True :-)
Anyway, here's the patch to change to an int. I haven't changed the other
two, because it also means introducing a cast in the loop body. If
On Thu, 2008-04-17 at 00:50 +0100, Andrew John Hughes wrote:
twisti, we should also maybe port the CACAO test if it's not VM specific?
No, the testcase is not VM specific. IIRC I also encouraged the student
to write the testcase for Mauve, but that does not work all the time :-)
- twisti
On Tue, 2008-04-15 at 23:28 +0200, Sebastian Mancke wrote:
Hi,
attached is a fix for the method Class.getSimpleName(), together with a
small testcase. The old implementation failed on usual class names, as
well as on inner classes.
Please comment/commit.
This reminds me of this patch:
On Wed, 2008-04-16 at 10:58 +0200, Sebastian Mancke wrote:
I saw this patch and think it broke the behaviour, because of the wrong
usage of fullName.lastIndexOf(., pos): The pos argument, counted from
left makes no sense in this method.
Also, I think, that advancing 'pos', dependent on
On Sun, 2008-03-16 at 20:03 +, Andrew John Hughes wrote:
Done.
ChangeLog:
2008-03-16 Andrew John Hughes [EMAIL PROTECTED]
* vm/reference/java/lang/reflect/VMField.java:
Consistently use 'clazz' as the name of the variable
containing the declaring class.
On Mon, 2008-03-03 at 21:21 +, Andrew John Hughes wrote:
Sorry for the late response. Could we use a consistent class variable
name?
final class VMConstructor
{
+ Class clazz;
+ int slot;
final class VMField
{
+ Class declaringClass;
+ String name;
+ int slot;
final
On Sun, 2008-03-02 at 02:08 +, Andrew John Hughes wrote:
This patch separates out common methods from the reflection VM
interfaces (Field, Constructor, Method) so that they are provided
in Classpath itself, and only native methods now appear in
new VMField, VMMethod and VMConstructor
On Fri, 2008-02-08 at 12:23 +0100, Dalibor Topic wrote:
Yeah, I'll do the latter, as there is no point in having this method in
the VM interface if we can do it (and the methods its implementation
invokes) in Java. Same for Float and the int conversion method.
I completely agree.
- twisti
On Thu, 2007-11-08 at 16:44 +0100, Robert Schuster wrote:
Now if HAVE_STATVFS is not defined the compiler is correct mourning that
'path' and 'type' are not used. I have seen code (in classpath) where
this issue is solved this way:
JNIEXPORT long long
cpio_df (const char *path,
On Tue, 2007-09-18 at 06:47 +0200, Jeroen Frijters wrote:
I believe you are correct that Classpath doesn't currently support the
system assertions boolean. I believe you are also correct in assuming
that noone cares ;-)
Does the Classpath code even have any assertions? If it doesn't, then
On Mon, 2007-09-17 at 12:17 +1000, Robin Garner wrote:
[ping]
Robin Garner wrote:
GCC Bugzilla Bug 32541
On JikesRVM (and possibly other VMs), dacapo xalan performs lots of
small IO using the default character encoding. OutputStreamWriter
currently buffers characters when
On Tue, 2007-09-04 at 10:25 +0200, Christian Thalinger wrote:
On Tue, 2007-08-21 at 15:27 +0200, Gregor Kaufmann wrote:
Small patch to allow setting a different default assertion status for
user and system classes (required to replicate the behavior of the
sun vms).
Hi!
OK
On Tue, 2007-08-21 at 15:27 +0200, Gregor Kaufmann wrote:
Small patch to allow setting a different default assertion status for
user and system classes (required to replicate the behavior of the
sun vms).
Hi!
OK, this email was a bit short without much explanation. We're
currently
Sorry, wrong list.
On Thu, 2007-08-30 at 13:25 +0200, Christian Thalinger wrote:
On Thu, 2007-08-30 at 11:20 +0200, Mark Wielaard wrote:
Hi Christian,
On Thu, 2007-08-30 at 11:04 +0200, Christian Thalinger wrote:
Would that be OK to commit?
2007-08-30 Christian Thalinger
On Fri, 2007-06-29 at 13:53 +0200, Christian Thalinger wrote:
Hi!
I was sick of renaming java_lang_ClassLoader$NativeLibrary.h all the
time when building openjdk, so I wrote this small patch. I think it's
correct, but I'm not an expert.
I also renamed filename to file, as I think it's
On Tue, 2007-07-31 at 17:32 +0200, Dalibor Topic wrote:
OK to commit.
Commited as:
2007-07-31 Christian Thalinger [EMAIL PROTECTED]
* tools/gnu/classpath/tools/javah/Main.java (writeHeaders):
Renamed filename to file.
* tools/gnu/classpath/tools/javah
On Tue, 2007-07-24 at 07:26 -0700, Tom Tromey wrote:
Twisti == Christian Thalinger [EMAIL PROTECTED] writes:
Twisti @files can include any arguments that are valid, not only
Twisti files. Maybe you know how to get Java getopt to support that.
Ah. I am not sure there is a way from
Hi!
While building OpenJDK b16 I noticed that jar does not support @
arguments. OpenJDK e.g. uses it to import the binary plug files.
This patch adds support for @filelist, where the file filelist may only
contain files and not options, as the spec says.
Should I commit that (incomplete)
On Tue, 2007-07-24 at 06:43 -0700, Tom Tromey wrote:
Twisti == Christian Thalinger [EMAIL PROTECTED] writes:
Twisti This patch adds support for @filelist, where the file filelist
Twisti may only contain files and not options, as the spec says.
Twisti Should I commit that (incomplete
On Wed, 2007-07-04 at 15:41 +0200, Mark Wielaard wrote:
Hi Christian,
On Tue, 2007-07-03 at 21:51 +0200, Christian Thalinger wrote:
I'd like to commit the following patch which does not change anything
for GNU Classpath, but it may break some VMs, as it changes the VM
interface
Hi!
While trying to build OpenJDK b16 yesterday I noticed that one Makefile
uses iso8859-1 as encoding for javac. But GNU Classpath does not
include that name as alias. This patch adds this one alias.
2007-07-23 Christian Thalinger [EMAIL PROTECTED]
* gnu/java/nio/charset/ISO_8859_1
On Mon, 2007-07-23 at 23:34 +0200, Mark Wielaard wrote:
Hi Christian,
On Mon, 2007-07-23 at 19:31 +0200, Christian Thalinger wrote:
While trying to build OpenJDK b16 yesterday I noticed that one Makefile
uses iso8859-1 as encoding for javac. But GNU Classpath does not
include that name
On Sun, 2007-07-01 at 17:53 +0100, Chris Burdess wrote:
Christian Thalinger wrote:
This patch, committed, fixes a bug in the XSL transformer where you have a
with-param node with empty content.
2007-07-01 Chris Burdess [EMAIL PROTECTED]
* gnu/xml/transform
Hi!
Today I noticed that openjdk's current sun.misc.Unsafe implementation
declares unpark with an Object parameter, instead of Thread. This makes
some problems within cacao, as I switched to the Unsafe of openjdk (for
the openjdk library integration).
I'd like to commit the following patch
On Sun, 2007-07-01 at 12:13 +0100, Chris Burdess wrote:
This patch, committed, fixes a bug in the XSL transformer where you have a
with-param node with empty content.
2007-07-01 Chris Burdess [EMAIL PROTECTED]
* gnu/xml/transform/WithParam.java: Handle case where content is
Hi!
I was sick of renaming java_lang_ClassLoader$NativeLibrary.h all the
time when building openjdk, so I wrote this small patch. I think it's
correct, but I'm not an expert.
I also renamed filename to file, as I think it's more appropriate. I
could remove that renaming if not wanted.
Is this
-06-22 Christian Thalinger [EMAIL PROTECTED]
* include/jni.h (JNINativeInterface): Renamed to
JNINativeInterface_, added const to arguments where necessary.
(JNIInvokeInterface): Renamed to JNIInvokeInterface_.
- twisti
Hi!
While doing the cacao port to use openjdk's libraries, I noticed the
function prototypes in sun's jni.h are a bit different than ours. And
also the JNINativeInterface and JNIInvokeInterface struct names. As
cacao does not have it's own jni.h, we'd need these changes.
Any objections?
-
On Thu, 2007-05-10 at 17:00 +0200, Mario Torre wrote:
Il giorno gio, 10/05/2007 alle 13.39 +0100, Chris Burdess ha scritto:
Will the switches just be ignored by 1.4 and earlier versions of
javac, or will they cause an error?
I don't think they are ignored.
Is there configury in any
On Mon, 2007-04-09 at 21:11 +0200, Christian Thalinger wrote:
On Mon, 2007-04-09 at 19:17 +0200, Christian Thalinger wrote:
Ok, here we go:
input.systemId=file:/home/twisti/cacao/spec/jbb2005/xml/template-document.xml
ids.systemId=jbb-document.dtd
url=jbb-document.dtd
On Mon, 2007-04-16 at 14:16 +0200, Mark Wielaard wrote:
Christian, can you retest against cvs HEAD and let us know whether or
not this solves your regressions?
Yes, works.
- twisti
:-)
Could you please write a changelog entry and fix this little nitpick?
Commited as follows.
- twisti
---
2007-04-12 Christian Thalinger [EMAIL PROTECTED]
* native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkClipboard.c
(clipboard_get_func): Call DeleteLocalRef.
* native/jni/gtk
Hi!
I just found this one in my local working copy. Should I commit it?
- twisti
---
Index: native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkClipboard.c
===
RCS file:
Hi!
I added the AC_C_INLINE check in configure.ac, since some compilers
don't understand the inline keyword.
This one is commited and should go on the release branch.
- twisti
---
2007-04-11 Christian Thalinger [EMAIL PROTECTED]
* configure.ac: Added AC_C_INLINE.
Index
Hi!
MIPSpro warned about that. Commited.
- twisti
---
2007-04-11 Christian Thalinger [EMAIL PROTECTED]
* native/jni/native-lib/cpproc.c (cpproc_forkAndExec): Fixed
warning on some compilers, e.g. MIPSpro.
Index: native/jni/native-lib/cpproc.c
Hi!
IRIX had a problem with including sys/fcntl.h directly. Commited.
- twisti
---
2007-04-11 Christian Thalinger [EMAIL PROTECTED]
* native/jni/java-nio/javanio.c: Include fcntl.h instead of
sys/fcntl.h.
Index: native/jni/java-nio/javanio.c
Hi!
This patch checks for the number of arguments of the gethostbyname_r
library call. The six- and five-argument case is implemented, the
three-argument case is missing, but I added an #error.
Ok to commit?
- twisti
---
2007-04-11 Christian Thalinger [EMAIL PROTECTED]
* LICENSE
On Sat, 2007-04-07 at 19:14 +0200, Mark Wielaard wrote:
Hi Christian,
On Sat, 2007-04-07 at 17:03 +0200, Christian Thalinger wrote:
This patch broke SPECjbb2005
(http://developer.classpath.org/pipermail/classpath/2007-March/001917.html).
Could you provide us with the input
On Mon, 2007-04-09 at 19:17 +0200, Christian Thalinger wrote:
Ok, here we go:
input.systemId=file:/home/twisti/cacao/spec/jbb2005/xml/template-document.xml
ids.systemId=jbb-document.dtd
url=jbb-document.dtd
This is the output before Chris' patch:
input.systemId=file:/home/twisti/cacao/spec
On Tue, 2007-02-06 at 10:13 +, Chris Burdess wrote:
I committed this patch which does two things: fixes a problem whereby
Xerces did not support an asynchronous DOM LS parser, and works
around a problem with the URLs specified in XML system IDs.
Note, this is only a workaround and
On Wed, 2007-04-04 at 13:36 +0200, Christian Thalinger wrote:
Yeah, tomorrow :-) Anyways, here it is. Commit?
Commited as follows.
2007-04-05 Christian Thalinger [EMAIL PROTECTED]
* native/jni/java-nio/gnu_java_nio_charset_iconv_IconvDecoder.c,
native/jni/java-nio
On Thu, 2007-04-05 at 14:21 +0200, Christian Thalinger wrote:
On Wed, 2007-04-04 at 13:36 +0200, Christian Thalinger wrote:
Yeah, tomorrow :-) Anyways, here it is. Commit?
Commited as follows.
Had another one of this type pending. Commited.
- twisti
---
2007-04-05 Christian Thalinger
Hi!
I reverted the reverted patch because it's required. See also:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22800.
This should finally fix all double-arm problems.
- twisti
---
2007-04-05 Christian Thalinger [EMAIL PROTECTED]
PR classpath/22800:
* native/jni/java-lang
On Fri, 2007-03-30 at 01:36 +0200, Christian Thalinger wrote:
On Thu, 2007-03-29 at 16:35 -0600, Tom Tromey wrote:
We discussed on irc but I thought the answer should be in the list
archives too.
Annotation the argument with UNUSED. This is defined in config.h.
We already use this all
On Thu, 2007-04-05 at 00:57 +0200, Mark Wielaard wrote:
Christian, this hopefully solves the issue you reported on irc with
MegaMek. Could you test?
Yes, this fixed the problem.
- twisti
Hi!
This patch fixes -Werror compilation with !HAVE_ICONV. Daney pointed me
to the (void)-solution. I'm not sure if that is the best solution, but
it works.
Comments?
- twisti
---
Index: native/jni/java-nio/gnu_java_nio_charset_iconv_IconvDecoder.c
On Thu, 2007-03-29 at 16:35 -0600, Tom Tromey wrote:
We discussed on irc but I thought the answer should be in the list
archives too.
Annotation the argument with UNUSED. This is defined in config.h.
We already use this all over; gcc's __attribute__ ((unused)) really
means might or might
On Wed, 2007-03-28 at 11:48 +0200, Mark Wielaard wrote:
Hi,
Andrew and Christian pointed out that the way we fiddle with umask is 1)
not-thread-safe and 2) unnecessary. So this patch fixes that up. It also
removes the linking against libclasspathnative for the nio library since
it doesn't
On Thu, 2007-02-15 at 23:46 +, Andrew John Hughes wrote:
I'm committing Edwin's patch for PR28652. According
to Mark, this is minor enough not to warrant an assignment
just yet.
Changelog:
2006-10-14 Edwin Steiner [EMAIL PROTECTED]
PR classpath/28652:
*
+++ ChangeLog 15 Dec 2006 14:53:13 -
@@ -1,3 +1,7 @@
+2006-12-15 Christian Thalinger [EMAIL PROTECTED]
+
+ * autogen.sh: Check for all 2.6x autoconf versions.
+
2006-12-14 David Daney [EMAIL PROTECTED]
* gnu/java/net/protocol/http/HTTPURLConnection.java: Clean
Index: autogen.sh
On Fri, Dec 15, 2006 at 11:18:40AM -0700, Tom Tromey wrote:
Twisti == Christian Thalinger [EMAIL PROTECTED] writes:
Twisti Ok to commit? I guess so...
Yes, thanks.
Commited.
- twisti
--- ChangeLog 15 Dec 2006 21:48:39 - 1.8991
+++ ChangeLog 15 Dec 2006 22:05:37 -
@@ -1,3 +1,14 @@
+2006-12-15 Christian Thalinger [EMAIL PROTECTED]
+
+ * m4/acinclude.m4 (CLASSPATH_FIND_JAVAC): Added support for Sun's
+ javac.
+ (CLASSPATH_WITH_JAVAC): Added
On Sat, Oct 21, 2006 at 09:01:01AM +0200, Christian Thalinger wrote:
On Sat, 2006-10-21 at 00:45 +0100, Robert Lougher wrote:
Hi,
I tried building the lastest distcheck on Mac OS X last night (first
time in a while). The build trivially failed in two files:
native/jni/java-net
-04 Christian Thalinger [EMAIL PROTECTED]
* native/jni/java-nio/javanio.c (cpnio_read, cpnio_readv)
(cpnio_write, cpnio_writev, cpnio_socket, cpnio_connect)
(cpnio_accept, cpnio_sendto, cpnio_recvfrom, cpnio_fcntl)
(cpnio_select): Use CPNIO_EXPORT.
* native
Hi!
Current head did not build on darwin. This patch fixes it. I'll commit
it later, when I'm in linux.
- twisti
2006-09-29 Christian Thalinger [EMAIL PROTECTED]
* native/jni/java-net/gnu_java_net_VMPlainSocketImpl.c
(config-int.h): Added.
* native/jni/java-nio
This patch removes the ZERO !LEFT_JUSTIFY check, as it's ok to print
leading zeros with right justification.
- twisti
2006-09-20 Christian Thalinger [EMAIL PROTECTED]
* java/util/Formatter.java (basicIntegralConversion): Removed
check for ZERO !LEFT_JUSTIFY.
Index: java
On Thu, Sep 14, 2006 at 09:47:37AM +0200, Mark Wielaard wrote:
Hi Andrew and Christian,
By now you two probably know mprec.c better than others. Could you take
a look at this patch suggestion? Might this be related to the recent
miscompilation observed by Andrew? Do we have testcases for
On Wed, Sep 13, 2006 at 11:34:16AM +0200, Mark Wielaard wrote:
Hi Christian,
On Tue, 2006-09-12 at 16:57 +0200, Christian Thalinger wrote:
I really think we should include that patch and revert my old one. I've
tested it and it works. Ok to commit?
2006-10-12 Christian Thalinger
On Thu, Sep 14, 2006 at 11:18:34AM -0700, Casey Marshall wrote:
Does cacao support GetDirectBufferAddress properly? There may be a bug
there.
Hmm, it unlikely, because OpenGL stuff works and that uses directu
buffers a lot. But I'll try the patch.
- twisti
On Fri, Sep 15, 2006 at 12:20:33AM +0200, Christian Thalinger wrote:
On Thu, Sep 14, 2006 at 11:18:34AM -0700, Casey Marshall wrote:
Does cacao support GetDirectBufferAddress properly? There may be a bug
there.
Hmm, it unlikely, because OpenGL stuff works and that uses directu
buffers
Hi!
I really think we should include that patch and revert my old one. I've
tested it and it works. Ok to commit?
- twisti
2006-10-12 Christian Thalinger [EMAIL PROTECTED]
Fixes PR22800
* native/fdlibm/mprec.h (Storeinc): Define correctly for LE
architectures (like
On Thu, 2006-06-29 at 10:35 +0100, Andrew John Hughes wrote:
Based on your comments, it seems you agree with my original intuition
of making this a native VM call (by default) in the majority of
cases, but
efficiency would seem to be fairly VM specific.
I suppose I was aiming on
Hi!
Edwin mentioned that CACAO isn't listed in the README. So, here it
goes.
@mark: maybe you can backport this to the release branch.
TWISTI
2006-05-15 Christian Thalinger [EMAIL PROTECTED]
* README: Added CACAO to list of VMs.
Index: README
On Wed, Mar 22, 2006 at 11:23:50AM +0100, Christian Thalinger wrote:
On Tue, 2006-03-14 at 16:56 +0100, Michael Koch wrote:
Together with a mauve testcase that proves this I'm fine with it. :-)
(as I told you on IRC)
Ok, I'm unable to write a mauve testcase for that. The problem
On Thu, 2006-04-06 at 21:05 +0200, Mark Wielaard wrote:
My apologies. This seems to have been some strange issue with my local
setup. Everything works fine now again. Cacao is great!
:-)
TWISTI
On Tue, 2006-04-04 at 11:37 -0600, Tom Tromey wrote:
Actually I've been curious to know why these don't simply use
getModifiers (or whatever the equivalent is on VMClass).
I was meaning to make that change...
Tom, are you working on that or...?
TWISTI
On Tue, 2006-04-04 at 10:53 +0200, Christian Thalinger wrote:
On Mon, 2006-04-03 at 21:05 +0100, Andrew John Hughes wrote:
* vm/reference/java/lang/VMClass.java:
(isAnonymousClass(Class?)): New native method.
(isLocalClass(Class?)): New native method.
(isMemberClass(Class
On Sat, 2006-03-25 at 01:05 +0900, Ito Kazumitsu wrote:
ChangeLog:
2006-03-24 Ito Kazumitsu [EMAIL PROTECTED]
* gnu/regexp/CharIndexed.java(setLastMatch, getLastMatch, getAnchor):
New methods.
* gnu/regexp/CharIndexedCharArray.java(setLastMatch, getLastMatch,
On Tue, 2006-03-14 at 16:56 +0100, Michael Koch wrote:
Together with a mauve testcase that proves this I'm fine with it. :-)
(as I told you on IRC)
Ok, I'm unable to write a mauve testcase for that. The problem is:
$ cat test.java
import java.util.*;
public class test {
public static
On Fri, Mar 10, 2006 at 03:38:02PM +0100, Christian Thalinger wrote:
Hi!
We had a problem with a mauve testlet which turned out to be related to
the default locale. CACAO sets the user.country property instead of
the user.region per default. Some websearch
(http://www.kaffe.org/pipermail
On Mon, 2006-03-06 at 16:44 +0100, Robert Schuster wrote:
Hi,
I wanted to be able to know the VM I am running the Swing Demo on. Here is a
small patch that makes that possible.
Wanted to try that coool feature :-), but unfortunately:
java.lang.IllegalArgumentException: position: 2,
On Sat, 2006-03-04 at 21:53 +0100, Mark Wielaard wrote:
This should solve the issue Christian was seeing.
Yes, looks good. Thanks!
TWISTI
also be applied to the release
branch (if not too late).
Commited.
TWISTI
2006-03-06 Christian Thalinger [EMAIL PROTECTED]
* configure.ac: Check for FREETYPE2. This is a reverted patch and
is required on Darwin.
* native/jni/gtk-peer/Makefile.am (AM_LDFLAGS): Added
On Fri, 2006-02-17 at 11:30 +0100, Mark Wielaard wrote:
It was removed by this patch:
http://lists.gnu.org/archive/html/classpath/2005-11/msg00065.html
For which you said it worked for you:
http://lists.gnu.org/archive/html/classpath/2005-11/msg00066.html
That was my fault, as i only tested
On Mon, 2005-11-07 at 14:31 +0100, Mark Wielaard wrote:
Hi,
As discussed on the main list the checks for freetype2 are not necessary
and might break on (really) old gtk+ installations. This patch removes
them.
2005-11-07 Mark Wielaard [EMAIL PROTECTED]
* configure.ac: Don't
On Mon, 2006-02-13 at 21:40 +0100, Christian Thalinger wrote:
Haven't recompiled the classpath for a long time on darwin, but this
check is really necessary for darwin. What was the exact reason to
remove this?
As Mark suggested on irc, here is the linking error message:
ld: .libs
On Thu, Jan 26, 2006 at 08:17:53AM -0600, Archie Cobbs wrote:
Could instead you just have a configure test for byte order?
Not sure if __IEEE_BYTES_LITTLE_ENDIAN is going to cover all
possible cases.
Of course, since you've got a patch and I don't feel free
to ignore me :-)
Originally the
Hi!
Not sure if it's related to this commit, but mips-irix does not work
anymore:
cc-1020 cc: ERROR File = javanet.c, Line = 1062
The identifier MSG_NOSIGNAL is undefined.
TARGET_NATIVE_NETWORK_SOCKET_SEND (fd, p + offset, len, bytes_sent);
^
cc-1020 cc: ERROR File =
On Mon, 2006-01-16 at 20:45 +0100, Mark Wielaard wrote:
Seems something is completely broken in the new native layer, I cannot
even get it to compile (Roman, did you try the --enable-Werror configure
flag?). And since the macros are such a pain to debug I have not been
able to really
On Tue, 2006-01-17 at 12:43 +0100, Mark Wielaard wrote:
The old target layer didn't define any symbols itself so it didn't need
to be linked into any other (jni) libraries. We might want to handle it
like we do with jcl (native/jni/classpath) for now. But I haven't yet
looked at the new target
, hint...)
Yeah, sorry, i was a bit overeager :-) Here the corrected patch _with_
changelog entry.
TWISTI
2006-01-17 Christian Thalinger [EMAIL PROTECTED]
* configure.ac: Set TARGET.
* native/Makefile.am, native/jni/classpath/Makefile.am,
native/jni/gtk-peer/Makefile.am
On Tue, 2006-01-17 at 12:39 -0700, Tom Tromey wrote:
Twisti --- native/jni/classpath/Makefile.am 24 Oct 2005 22:16:10 -
1.6
[...]
Twisti AM_LDFLAGS = @CLASSPATH_MODULE@
This setting causes libclasspath to be built as a loadable module.
I don't think that is what we want. I
On Thu, 2006-01-12 at 21:37 +0100, Guilhem Lavaux wrote:
To define isnan in any cases. Looking at IRC logs it looks like that we
met some trouble on darwin5. Here is a new patch which should make
everyone happy.
Tell me if it's ok for you.
Yeah, that one makes me happy. Commit!
TWISTI
) can be something different.
Commit?
TWISTI
2005-01-12 Christian Thalinger [EMAIL PROTECTED]
* resource/Makefile.am: Install
logging.properties into $(prefix)/lib.
* resource/Makefile.am (securitydir): Changed to
$(prefix)/lib/security.
Index: resource
On Tue, Jan 10, 2006 at 04:47:41PM +0100, Mark Wielaard wrote:
* liIf the system property codejava.util.logging.config.file/code
* is not set, however, the contents of the URL
* {gnu.classpath.home.url}/logging.properties are passed to
* [EMAIL PROTECTED]
1 - 100 of 142 matches
Mail list logo
