On Mon, 2005-06-27 at 17:03 -0400, Bryce McKinlay wrote:
> Please post ChangeLog entries whenever you post a patch. I've made some
> comments below.
ChangeLog's were posted with the original patches, and their all the
same "* filename.java: New file." I assumed such a trivial thing would
be far
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Wow,
what a huge patch.
My nitpicks:
- - each author gets a separate @author tag (instead of using commas)
- - null and field names in (doc-) comments should be enclosed by
- - Instead of comments like "Returns true if largeModel is set" try to
expla
On Mon, 2005-06-27 at 21:18 +0200, Pierrick Brihaye wrote:
> Ziga Mahkovec a écrit :
> > Following up my previous patch[1], this patch enables the use of
> > BufferedImages for component double buffering. This results in much
> > better performance and surprisingly even fixes the Pango text render
Keith Seitz wrote:
I believe the attached patch addresses concerns over temporary
OutputStream allocations.
Yes, this looks better. At least by using reset() on the
ByteArrayOutputStream we are eliminating one level of allocation and
potential copying. Aaron managed to convince me that th
Keith Seitz wrote:
So, after much discussion last week, I'm not sure where I stand in
making any progress at getting this stuff committed.
I've attached the outstanding patches awaiting more review and/or
approval.
Hi Keith,
Please post ChangeLog entries whenever you post a patch. I've mad
Hi,
Ziga Mahkovec a écrit :
Following up my previous patch[1], this patch enables the use of
BufferedImages for component double buffering. This results in much
better performance and surprisingly even fixes the Pango text rendering
issues.
What about the Batik tests ?
p.b.
_
Aaron Luchko wrote:
Ok, after some discussions with Bryce and Keith I've arrived at an
improved version of the patch. The first couple changes involve using an
array instead of the hashtable and tossing the IOException up from
_processOnePacket and exiting in run() when we get it.
The final cha
Hi,
I believe the attached patch addresses concerns over temporary
OutputStream allocations.
The changes in this patch are reflected in
gnu.classpath.jdwp.event.Event submission.
Comments/questions/concerns?
Keith
ChangeLog
* gnu/classpath/jdwp/transport/JdwpConnection.java (sendPacket)
Hi,
So, after much discussion last week, I'm not sure where I stand in
making any progress at getting this stuff committed.
I've attached the outstanding patches awaiting more review and/or
approval.
Comments/questions/concerns?
Keith
Index: gnu/classpath/jdwp/event/filters/IEventFilter.java
==
This seems like it must be duplicated somewhere, but I couldn't find
it quickly.
I also looked but could not find code to reuse.
There is getPackagePortion in Class, maybe making that
package-private and then using it would be good.
We're in java.io...
I wouldn't mind adding a utility metho
On Sat, 2005-06-25 at 15:51 +0200, Mark Wielaard wrote:
> Hi,
>
> While working with the native gtk+ awt peers I often need to find out
> where precisely a warning or error message is comming from. The
> following patch installs a default log handler that dumps a stacktrace
> whenever we encounter
On Sat, 2005-06-25 at 16:53 +0200, Ziga Mahkovec wrote:
> This patch fixes some of the problems[1] Mark was having when running
> the Swing demo with Java2D/Cairo. No text is shown yet due to some
> Pango issues, but at least the drawing part looks OK now:
>
> http://bootchart.org/misc/classpat
On Sun, 2005-06-26 at 16:55 +0200, Ziga Mahkovec wrote:
> Following up my previous patch[1], this patch enables the use of
> BufferedImages for component double buffering. This results in much
> better performance and surprisingly even fixes the Pango text rendering
> issues.
>
> Swing on Cairo n
> "Guilhem" == Guilhem Lavaux <[EMAIL PROTECTED]> writes:
Guilhem> We do care. :) So maybe I should bring back the block
Guilhem> concerning arm. If there are other machines which behaves the
Guilhem> same way we would have to implement something in configure
Guilhem> though.
Guilhem> Do you
> "Roman" == Roman Kennke <[EMAIL PROTECTED]> writes:
Roman> + public static Component[] getVisibleChildren(Container c)
Roman> + {
Roman> +Component[] children = c.getComponents();
Roman> +Vector visible = new Vector();
Roman> +for (int i = 0; i < children.length; i++)
Roman> +
> "Guilhem" == Guilhem Lavaux <[EMAIL PROTECTED]> writes:
Guilhem> Here is a quick patch to replace the existing magic in ieeefp.h to
Guilhem> detect endianness using a configure defined macro.
Guilhem> Comments before commit ?
How does this handle ARM machines with the weird 'double' layout?
> "Mark" == Mark Wielaard <[EMAIL PROTECTED]> writes:
Mark> The stack trace generation is only conditional on it being something
Mark> that already generates output on stderr by default. Since I think that
Mark> none of these error/warning messages should ever occur in our code (it
Mark> means
> "Daniel" == Daniel Bonniot <[EMAIL PROTECTED]> writes:
Daniel> Purpose: readResolve and writeReplace are currently only looked for in
Daniel> the current class, while accessible parent methods must also be
Daniel> considered. This is what this patch does.
Looks good to me fwiw. A couple mi
Hi,
I finally fixed the horribly broken BoxLayout. I have rewritten the
layout algorithm from scratch. It is now more comprehensible and correct
(I have a couple of testcases here that I commit to Mauve soon).
The layout algorithm now uses an inner interface (Direction) which
allows the algo
I fixed two smaller issues in javax.swing.Box:
2005-06-27 Roman Kennke <[EMAIL PROTECTED]>
* javax/swing/Box.java
(createGlue): Return Short.MAX_VALUE instead of Integer.MAX_VALUE as
dimension in the Filler component.
(createHorizontalGlue): Return a new Filler obje
JComponent.getPreferredSize lets the ComponentUI override a manually set
preferredSize. This is not allowed. The priority is as follows:
(0. constrained by minimumSize)
1. user set preferredSize
2. ComponentUI preferredSize
3. Container preferredSize (LayoutManager)
4. Component preferredSize
T
On Mon, 2005-06-27 at 08:31 -0600, Tom Tromey wrote:
> > "Guilhem" == Guilhem Lavaux <[EMAIL PROTECTED]> writes:
>
> Guilhem> Here is a quick patch to replace the existing magic in ieeefp.h to
> Guilhem> detect endianness using a configure defined macro.
> Guilhem> Comments before commit ?
>
2005-06-27 Audrius Meskauskas <[EMAIL PROTECTED]>
org/omg/DynamicAny/DynAnyPackage/InvalidValue.java,
org/omg/DynamicAny/DynAnyPackage/InvalidValueHelper.java,
org/omg/DynamicAny/DynAnyPackage/TypeMismatch.java,
org/omg/DynamicAny/DynAnyPackage/TypeMismatchHelper.java,
org/omg/DynamicAny/Dy
2005-06-25 Audrius Meskauskas <[EMAIL PROTECTED]>
* org/omg/PortableServer/POAPackage/ServantAlreadyActive.java,
org/omg/PortableServer/POAPackage/ServantAlreadyActiveHelper.java: New
files.
/* ServantAlreadyActiveHelper.java --
Copyright (C) 2005 Free Software Foundation, Inc.
This file
Hi,
As I just got my FSF assignment, here my first patch (against HEAD).
Purpose: readResolve and writeReplace are currently only looked for in the
current class, while accessible parent methods must also be considered. This
is what this patch does.
This patch is fixing two mauve errors I s
I'm committing the attached patch which merges the new 1.5 features
of java.lang.Integer back into HEAD from the generics branch. These
don't use any new language features (although methods like valueOf()
have an indirect use in autoboxing) and simply seem to have landed
on the generics branch due
Hi,
Here is a quick patch to replace the existing magic in ieeefp.h to
detect endianness using a configure defined macro.
Comments before commit ?
Guilhem.
ChangeLog entry:
2005-06-27 Guilhem Lavaux <[EMAIL PROTECTED]>
* native/fdlibm/ieeefp.h: Removed all preprocessor magic to
27 matches
Mail list logo