[ pdl-Bugs-1221742 ] Failure of some PDL tests with perl-5.8.7

Bugs item #1221742, was opened at 2005-06-16 08:18
Message generated for change (Comment added) made by zowie
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100612&aid=1221742&group_id=612

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: other
Group: None
>Status: Closed
Resolution: None
Priority: 5
Submitted By: Sisyphus (sisyphus_)
Assigned to: Nobody/Anonymous (nobody)
Summary: Failure of some PDL tests with perl-5.8.7

Initial Comment:
On both Linux (Mandrake-9.1) and Win32 (Windows 
2000), trying to build PDL
results in (the same) failures with limits_keyspecs.t and
limits_normalize_dsets.t.
----------------------------------------
D:\comp58_M\PDL>perl -Mblib t\limits_keyspecs.t
1..12
ok 1 - x<n>p&f
ok 2 - <n>p&f
ok 3 - x,<n,>p,&f
ok 4 - x <n >p &f
ok 5 - <n>p&f
ok 6 - &f
ok 7 - x
not ok 8 - &
#     Failed test (t\limits_keyspecs.t at line 68)
ok 9 - undef
not ok 10 - <>&
#     Failed test (t\limits_keyspecs.t at line 68)
ok 11 - =s
ok 12 - <<
# Looks like you failed 2 tests of 12.
--------------------------------------
--------------------------------------
D:\comp58_M\PDL>perl -Mblib 
t\limits_normalize_dsets.t
1..21
Error - tried to use an unknown data structure as a PDL 
at
D:/perl58_M/5.8.7/lib/Test/More.pm line 1164, <DATA> 
line 206.
# Looks like your test died before it could output 
anything.
-------------------------------------

The problem is with the perl that's running the show, not 
the PDL that was
built using perl-5.8.7. There's no problem running:

D:\comp58_M\PDL>\perl\5.8.6\bin\perl -Mblib 
t\limits_keyspecs.t

and:

D:\comp58_M\PDL>\perl\5.8.6\bin\perl -Mblib 
t\limits_normalize_dsets.t

The failures look to be the result of 
Test::Builder::_unoverload() which is new (and 
undocumented, and untested) in 5.8.7.

One further suspects that operating systems other than 
the 2 mentioned above will be similarly affected.

Cheers,
Rob

----------------------------------------------------------------------

Comment By: Craig DeForest (zowie)
Date: 2006-07-18 16:59

Message:
Logged In: YES 
user_id=20200

Bug fixed in CVS.
Thanks for reporting the problem!


----------------------------------------------------------------------

Comment By: Sisyphus (sisyphus_)
Date: 2005-09-08 06:11

Message:
Logged In: YES 
user_id=1195342

Committed Chris Marshall's patched version of 
limits_keyspecs.t to the CVS rep today (8-Sept-05).

Rob (Sisyphus)

----------------------------------------------------------------------

Comment By: Chris Marshall (marshallch)
Date: 2005-09-02 17:20

Message:
Logged In: YES 
user_id=44920

The problem with limit_keyspecs.t was that hash values
containing the empty string were being set up as answers to
check against but the actual function being tested returns
undef and not the empty string.  My original report is on
the Patched page #1262686:

http://sourceforge.net/tracker/download.php?group_id=612&atid=300612&file_id=147884&aid=1262686

--Chris

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100612&aid=1221742&group_id=612