Mark Hindess wrote:
> Nathan, yeah that's probably it. It's caught me out a couple of times.
>
> Is that swing test fix ready? If not perhaps we should just exclude
> it, and then I can dump the with.awt.swing property for good?
I've reviewed and applied the TransferHandler fix, but I see anot
nope, did that too. That was the first thing I thought of. Tried JUnit
4.x and Junit 3.8.1
Elena Semukhina wrote:
> On 10/8/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
>>
>> still on that new machine.
>>
>> classlib builds fine, but "ant test" results in it appearing to be very,
>> very c
On 10/8/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
still on that new machine.
classlib builds fine, but "ant test" results in it appearing to be very,
very confused when trying to compile, with the first module,
accessibility.
It can't find things like "BasicSwingTestCase", although I c
nope, that wasn't it. I did a ant clean, ant , and test
still broke...
Mark Hindess wrote:
Nathan, yeah that's probably it. It's caught me out a couple of times.
Is that swing test fix ready? If not perhaps we should just exclude
it, and then I can dump the with.awt.swing property for go
Nathan, yeah that's probably it. It's caught me out a couple of times.
Is that swing test fix ready? If not perhaps we should just exclude
it, and then I can dump the with.awt.swing property for good?
Regards,
Mark.
On 7 October 2006 at 14:29, "Nathan Beyer" <[EMAIL PROTECTED]> wrote:
> Are
That could be it. If that case, we have a problem...
Nathan Beyer wrote:
Are you running build with "-Dwith.awt.swing=true"? I had some weird
problems yesterday and it just seemed that I wasn't consistently using
that property for all ant runs.
-Nathan
On 10/7/06, Geir Magnusson Jr. <[EMAIL PR
Are you running build with "-Dwith.awt.swing=true"? I had some weird
problems yesterday and it just seemed that I wasn't consistently using
that property for all ant runs.
-Nathan
On 10/7/06, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
still on that new machine.
classlib builds fine, but "an
still on that new machine.
classlib builds fine, but "ant test" results in it appearing to be very,
very confused when trying to compile, with the first module,
accessibility.
It can't find things like "BasicSwingTestCase", although I can confirm
that it was build in test_support...
I figu
IMHO, the compilation problem is already fixed. At least I was able to
build recent classlib with eclipse compiler (using blackdown-jre-1.4.2
and drlvm).
Ah! Looks like the default rmi implementation was changed and it
doesn't contain the problem. The implementation relying on j.u.Scanner
is exclu
Tim Ellison wrote:
Vladimir Gorr wrote:
maybe does it make sense adding the RI rt.jar to CLASSPATH for the
ECJ-based
compiler?
In this case all we didn't implemented will be taken from RI. Is this a
stupid idea?
You can add the RI's class libraries using a command-line option if you
so
Vladimir Gorr wrote:
On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
That's right. When we compile with Sun's compiler it compiles against
the Sun class libraries, and when we compile with the Harmony
(ECJ-based) javac it compiles against the Harmony class libraries.
Tim,
maybe does i
Ivan Volosyuk wrote:
I have good idea. Let's move to eclipse compiler in classlib's build
process. :)
I believe we did this a long time ago...
geir
This will allow us to build classlib using number of available JREs1.4.
--
Ivan
2006/5/26, Vladimir Gorr <[EMAIL PROTECTED]>:
On 5/26/06, Ti
Ivan Volosyuk wrote:
> I have good idea. Let's move to eclipse compiler in classlib's build
> process. :)
There are instructions in the build file for how to switch the default
build compiler to use ECJ.
> This will allow us to build classlib using number of available JREs1.4.
The build needs a
I have good idea. Let's move to eclipse compiler in classlib's build process. :)
This will allow us to build classlib using number of available JREs1.4.
--
Ivan
2006/5/26, Vladimir Gorr <[EMAIL PROTECTED]>:
On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
>
> Vladimir Gorr wrote:
> > maybe doe
On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
Vladimir Gorr wrote:
> On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
>>
>> Vladimir Gorr wrote:
>> > maybe does it make sense adding the RI rt.jar to CLASSPATH for the
>> > ECJ-based
>> > compiler?
>> > In this case all we didn't implemente
Vladimir Gorr wrote:
> On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
>>
>> Vladimir Gorr wrote:
>> > maybe does it make sense adding the RI rt.jar to CLASSPATH for the
>> > ECJ-based
>> > compiler?
>> > In this case all we didn't implemented will be taken from RI. Is this a
>> > stupid idea?
>
On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
Vladimir Gorr wrote:
> maybe does it make sense adding the RI rt.jar to CLASSPATH for the
> ECJ-based
> compiler?
> In this case all we didn't implemented will be taken from RI. Is this a
> stupid idea?
You can add the RI's class libraries usin
Vladimir Gorr wrote:
> maybe does it make sense adding the RI rt.jar to CLASSPATH for the
> ECJ-based
> compiler?
> In this case all we didn't implemented will be taken from RI. Is this a
> stupid idea?
You can add the RI's class libraries using a command-line option if you
so choose, but I don;t
On 5/26/06, Tim Ellison <[EMAIL PROTECTED]> wrote:
That's right. When we compile with Sun's compiler it compiles against
the Sun class libraries, and when we compile with the Harmony
(ECJ-based) javac it compiles against the Harmony class libraries.
Tim,
maybe does it make sense adding the
That's right. When we compile with Sun's compiler it compiles against
the Sun class libraries, and when we compile with the Harmony
(ECJ-based) javac it compiles against the Harmony class libraries.
You could pass in the -classpath to our javac to pick up the reference
implementation of the class
2006/5/26, Mark Hindess <[EMAIL PROTECTED]>:
We don't have java.util.Scanner and
so the only reason it is compiling is because the RI compiler assumes
its own bootstrap classes so finds its java.util.Scanner.
Yes, this can be a reason.
--
Alexey A. Petrenko
Intel Middleware Products Division
On 26 May 2006 at 12:18, "Alexey Petrenko" <[EMAIL PROTECTED]> wrote:
>
> 2006/5/26, Ivan Volosyuk <[EMAIL PROTECTED]>:
> >
> > I am experimenting with eclipse compiler ...
>
> I've just built fresh Harmony sources without any problems.
But not with the Eclipse compiler... which I think means it
I've just built fresh Harmony sources without any problems.
2006/5/26, Ivan Volosyuk <[EMAIL PROTECTED]>:
I am experimenting with eclipse compiler from
org/apache/harmony/tools/javac/Main. The point is that I don't have
java5 at home computer, but I figured out that eclipse compiler in
ecj_3.2RC
I am experimenting with eclipse compiler from
org/apache/harmony/tools/javac/Main. The point is that I don't have
java5 at home computer, but I figured out that eclipse compiler in
ecj_3.2RC5.jar can be good substitution. I have tweaked the tools
javac wrapper and managed to start ant with that co
24 matches
Mail list logo