Mark Wielaard | 5 Sep 21:26 2004

Re: Mauve patches for SpinnerListModel

Hi,

On Thu, 2004-08-19 at 02:37, Andrew John Hughes wrote:
> Attached is a series of Mauve tests for SpinnerListModel
> to go in the directory gnu/testlet/javax/swing/SpinnerListModel.
> What do I do about adding these?

You post them to mauve-patches@... :)
And/Or you ask Tom for permission to check them in yourself.

> Also, does anyone have any idea as to why Mauve only seems to test java.io and
> java.net when I run make check?  Using ant, I can get it to compile all the tests
> but it segfaults in java.lang.reflect.Array.newInstance.

For these kind of questions there is the
mauve-discuss@... mailinglist.
Personally I don't use ant but the batch_run and runner script (see the
README) which should automatically run all the tests. The README also
has some more examples on how to run mauve "the traditional way" (make
and autotools).

The crash in Array.newInstance() is a bug in jamvm (which I assume you
are using). CCed Robert. I thought I had reported it earlier, but maybe
I forgot. This is in the last multi-dimensional test:

    val = 0;
    try
      {
        x = Array.newInstance(String.class, new int[]
          {Integer.MAX_VALUE, Integer.MAX_VALUE});
(Continue reading)

Robert Lougher | 6 Sep 04:36 2004
Picon

Re: Mauve patches for SpinnerListModel

Hi,

I was all ready to release JamVM 1.2.0 but I saw this :)  I've fixed
it, but it's too late to get a release out tonight now (it's 3:30am
and it's work tomorrow).  I'll hopefully get my brother to put it up
on Sourceforge for me tomorrow.

Rob.

On Sun, 05 Sep 2004 21:26:08 +0200, Mark Wielaard <mark@...> wrote:
> Hi,
> 
> On Thu, 2004-08-19 at 02:37, Andrew John Hughes wrote:
> > Attached is a series of Mauve tests for SpinnerListModel
> > to go in the directory gnu/testlet/javax/swing/SpinnerListModel.
> > What do I do about adding these?
> 
> You post them to mauve-patches@... :)
> And/Or you ask Tom for permission to check them in yourself.
> 
> > Also, does anyone have any idea as to why Mauve only seems to test java.io and
> > java.net when I run make check?  Using ant, I can get it to compile all the tests
> > but it segfaults in java.lang.reflect.Array.newInstance.
> 
> For these kind of questions there is the
> mauve-discuss@... mailinglist.
> Personally I don't use ant but the batch_run and runner script (see the
> README) which should automatically run all the tests. The README also
> has some more examples on how to run mauve "the traditional way" (make
> and autotools).
(Continue reading)

Thomas Zander | 6 Sep 18:22 2004

Re: Mauve patches for SpinnerListModel


On Sunday 05 September 2004 21:26, Mark Wielaard wrote:
> > Also, does anyone have any idea as to why Mauve only seems to test
> > java.io and java.net when I run make check?  Using ant, I can get it
> > to compile all the tests but it segfaults in
> > java.lang.reflect.Array.newInstance.
>
> For these kind of questions there is the
> mauve-discuss@... mailinglist.

Immediately after reporting the bug :) you can alter the build.xml to only 
test the classes you wrote quite easilly.
In the "test" target there are some <fileset /> lines and all you have to 
do is to make sure the includes part points to the dir you want to test 
(it will take all subclasses as well if you use 2 *'s at the end)

Naturally you should remove the fileset tags you don't want (there are 2 by 
default)

--

-- 
Regards
Thomas
Noa Resare | 17 Sep 23:20 2004

[PATCH] xml output from mauve

Hello friends

I started hacking on kaffe a few days ago, and this lead me to looking
at mauve. While thinking about writing some scripts that do fun stuff
with the output from mauve test runs and had a look at the perl script
that lives in kaffe cvs I came to the conclusion that the mauve output
wasn't as easy to parse as I wanted it to be.

So, I did some afternoon hacking and came up with this patch to mauve
that outputs the results from a mauve run into an xml file for easy
parsing. I haven't written anything that does anything with the output
yet but I thought I could share this with you anyway.

Note that since this patch touches Makefile.am autoreconf should be re-
run also and the generated changes should go into cvs.

Where should I add documentation about this?

cheers/noa

--

-- 
And the lions ate the christians and the christians burned the witches,
and even I am out of explanations -- Ola Salo
gpg fingerprint: F3C4 AC90 B885 FE15 344B  4D05 220B 7662 A190 6F09
Attachment (mauve-xmlout.patch): text/x-patch, 13 KiB
Noa Resare | 18 Sep 22:47 2004

Re: [PATCH] xml output from mauve (updated)

On fre, 2004-09-17 at 23:20 +0200, Noa Resare wrote:
> So, I did some afternoon hacking and came up with this patch to mauve
> that outputs the results from a mauve run into an xml file for easy
> parsing. I haven't written anything that does anything with the output
> yet but I thought I could share this with you anyway.
> 

This updated patch handles mauve output containing & < and > properly.
It also adds Makefile.am to the changelog entry.

/noa

ps. a test result diff program (that looks for regressions) in very
rough form is available from http://resare.com/noa/testdiff

--

-- 
And the lions ate the christians and the christians burned the witches,
and even I am out of explanations -- Ola Salo
gpg fingerprint: F3C4 AC90 B885 FE15 344B  4D05 220B 7662 A190 6F09
Attachment (mauve-xmlout2.patch): text/x-patch, 14 KiB
Mark Wielaard | 19 Sep 23:51 2004

Re: [PATCH] xml output from mauve (updated)

Hi,

On Sat, 2004-09-18 at 22:47, Noa Resare wrote:
> On fre, 2004-09-17 at 23:20 +0200, Noa Resare wrote:
> > So, I did some afternoon hacking and came up with this patch to mauve
> > that outputs the results from a mauve run into an xml file for easy
> > parsing. I haven't written anything that does anything with the output
> > yet but I thought I could share this with you anyway.
> 
> This updated patch handles mauve output containing & < and > properly.
> It also adds Makefile.am to the changelog entry.
>
> ps. a test result diff program (that looks for regressions) in very
> rough form is available from http://resare.com/noa/testdiff

Very nice!

Maybe this should be written as a new XMLTestHarness class?
Although I don't really see a problem with just tagging it on the
SimpleTestHarness. Opinions anybody?

Are you sure the XML escaping is always correct?
Maybe it doesn't really matter in this case, but you might want to look
whether GNU Kawa (http://www.gnu.org/software/kawa/) has something
appropriate in the gnu.xml package.

Thanks for doing this. Having easy to compare mauve results is very
valuable.

Cheers,
(Continue reading)

Noa Resare | 20 Sep 08:48 2004

Re: [PATCH] xml output from mauve (updated)

On sön, 2004-09-19 at 23:51 +0200, Mark Wielaard wrote:
> Hi,
> 
> On Sat, 2004-09-18 at 22:47, Noa Resare wrote:
> > On fre, 2004-09-17 at 23:20 +0200, Noa Resare wrote:
> > > So, I did some afternoon hacking and came up with this patch to mauve
> > > that outputs the results from a mauve run into an xml file for easy
> > > parsing. I haven't written anything that does anything with the output
> > > yet but I thought I could share this with you anyway.
> > 
> > This updated patch handles mauve output containing & < and > properly.
> > It also adds Makefile.am to the changelog entry.
> >
> > ps. a test result diff program (that looks for regressions) in very
> > rough form is available from http://resare.com/noa/testdiff
> 
> Very nice!
> 
> Maybe this should be written as a new XMLTestHarness class?
> Although I don't really see a problem with just tagging it on the
> SimpleTestHarness. Opinions anybody?

Maybe that would be better. However, I don't think the current solution
is horribly ugly and I'd rather use that time trying to contribute to
fixing jvm bugs. I'm open to suggestions though :)

> Are you sure the XML escaping is always correct?
> Maybe it doesn't really matter in this case, but you might want to look
> whether GNU Kawa (http://www.gnu.org/software/kawa/) has something
> appropriate in the gnu.xml package.
(Continue reading)

Noa Resare | 20 Sep 18:03 2004

tests impossible to pass with class library written using only spec?

I looking around in some java.io mauve failures i found some failing
test code in gnu/testlet/java/io/OutputStreamWriter/jdk11.java reading 

	// Note that for java.io the canonical name as returned by
        // getEncoding() must be the "historical" name. ISO8859_1.
        harness.check(osw2.getEncoding(), "ISO8859_1", 
		"OutputStreamWriter(writer, encoding)");
        osw2.close ();
        osw2 = new OutputStreamWriter(baos2, "latin1");
        harness.check(osw2.getEncoding(), "ISO8859_1", 
		"OutputStreamWriter(writer, encoding) // alias");
        osw2.close ();

Ah, simple I thought. It must be a list in the java class library
specification that tells what the historical names are. Now i know there
is no such list. That leads me to a more general question: is it ok for
mauve to test functionality that there is no way to implement using only
the class library spec (or looking at the code of the of the tests
preformed and just returning what is asked for)?

cheers/noa

--

-- 
And the lions ate the christians and the christians burned the witches,
and even I am out of explanations -- Ola Salo
gpg fingerprint: F3C4 AC90 B885 FE15 344B  4D05 220B 7662 A190 6F09

Mark Wielaard | 20 Sep 21:22 2004

Re: tests impossible to pass with class library written using only spec?

Hi,

On Mon, 2004-09-20 at 18:03, Noa Resare wrote:
> Ah, simple I thought. It must be a list in the java class library
> specification that tells what the historical names are. Now i know there
> is no such list.

The Java Class libries Second Edition Volume 1 lists them under the
String constructor description. An online copy can be found at:
http://java.sun.com/j2se/1.4.2/docs/guide/intl/encoding.doc.html
in the column "Canonical Name for java.io and java.lang API"

Cheers,

Mark
ferdinand laurey | 22 Sep 08:55 2004

Fw: If you like shopping o.n,line, wh~'at are you wa,iti-ng for?

her lover, and destroy him by her confession.

-----Original Message-----
From: Jeremy Mcdonald [mailto:s <at> huts.com]
To: perry ritzearnest beraunted melchinquincy bogdenarnold kuchan
Sent: Wednesday, October, 2004 6:52 AM
Subject: If you like shopping o'n_line, wh'`at are you wa`iti`ng for?

Experience the reliable service and check out the l~o,w price on Xalatan,
Aten'-olol, Zyban,  and V~al'~ium. Find out m~or_e here o`n.line with us.
http://rmpi.h.discovered3291nx.com/56/

It was actually the fr^.ie-nd in one fitness club referred your site. At
first I were still n^.ot f,u_lly comfortable with o-rde~-r m--ed`s o_n^line
without seeing the m_.ed~s first. But as the o'rde,~r process went on, I
were m^or`e relaxed cause there was n`^o other hassel. Finally I received
the o_rde,^r del_i_vered to my door and it was n-`ot one end. Next time, I
will try to o'rde'_r m~'ed_s with your site again.   --Kanny G. MI
milieubeskyttelseshensyns19  kassedamers05 hennings maureres  lidkbene


Gmane