On Tue, 6 Sep 2005, Andy Dougherty wrote:

> On Tue, 6 Sep 2005, Nicholas Clark wrote:
> 
> > On Tue, Sep 06, 2005 at 10:15:26AM -0700, Bernhard Schmalhofer via RT wrote:
> > 
> > > Too bad. As I reached the end of intuitive bug fixing, I am now skipping
> > > 'examples/japh/japh13.pasm' in 't/examples/japh.t'.
> > > 
> > > Perhaps some MacOS X user can provide more insight.
> > 
> > Is it just OS X?
> > Or do all big endian platforms give the same result?
> 
> No, it's just a bad test.  On my big endian Solaris/SPARC system, it works 
> if the system is unloaded, but fails if the system is busy.  From the test 
> itself:
> 
>       2 threads, synced by sleeping time only
> 
> I propose we just remove the test.
> 
> Similarly, japh8.pasm starts as follows:
> 
>       # Little Endian only using struct

At a minimum, I'd recommend the following.  It still leaves one failing 
japh test.

diff -r -u parrot-current/examples/japh/japh13.pasm 
parrot-andy/examples/japh/japh13.pasm
--- parrot-current/examples/japh/japh13.pasm    Tue Sep  6 10:15:24 2005
+++ parrot-andy/examples/japh/japh13.pasm       Wed Sep  7 10:10:42 2005
@@ -6,7 +6,8 @@
 
 =head1 DESCRIPTION
 
-2 threads, synced by sleeping time only
+2 threads, synced by sleeping time only.  This is unreliable,
+especially on multiprocessor or heavily-loaded systems.
 
 =cut
 
diff -r -u parrot-current/t/examples/japh.t parrot-andy/t/examples/japh.t
--- parrot-current/t/examples/japh.t    Wed Sep  7 09:45:19 2005
+++ parrot-andy/t/examples/japh.t       Wed Sep  7 10:10:02 2005
@@ -17,6 +17,9 @@
 Test the JAPHs in 'examples/japh'.
 For now there are only JAPHs in PASM.
 
+The numbering is very confusing:  Test 1, for example, corresponds to
+japh3.pasm.  XXX
+
 =cut
 
 use strict;
@@ -25,14 +28,14 @@
 
 {
   # these JAPHs should work
-  foreach my $japh_num ( 3, 8, 10, 12 ) {
+  foreach my $japh_num ( 3, 10, 12 ) {
     test_japh($japh_num);
   } 
 } 
 
 # See [perl #37082] in the Parrot RT
 SKIP: {
-  skip 'japh13.pasm has problems under MacOS', 1;
+  skip 'japh13.pasm is unreliable.', 1;
 
   foreach my $japh_num ( 13 ) {
     test_japh($japh_num);
@@ -43,7 +46,7 @@
 {
   # these JAPHs are known to fail
   local $TODO = 'various reasons';
-  foreach my $japh_num ( 1, 2, 4, 5, 6, 7, 9, 11, 14, 15 ) {
+  foreach my $japh_num ( 1, 2, 4, 5, 6, 7, 8, 9, 11, 14, 15 ) {
     test_japh($japh_num);
   } 
 };
 

-- 
    Andy Dougherty              [EMAIL PROTECTED]

Reply via email to