Re: Octave 3.6.2-rc0 release candidate available for ftp

[Michael Goffioul]

On Fri, May 11, 2012 at 9:22 PM, John W. Eaton <address@hidden> wrote:
> The first release candidate of Octave 3.6.1 is now available from
> alpha.gnu.org in the directory /gnu/octave:
>
>  4a8b92ece1fdc92f2e7e17e1c8d23364  octave-3.6.2-rc0.tar.bz2
>  077b58a9663c0cf4a0450e1c7516e82a  octave-3.6.2-rc0.tar.gz
>
>  -rw-r--r-- 1 3003 3002 15437714 May 11 19:55 octave-3.6.2-rc0.tar.bz2
>  -rw-r--r-- 1 3003 3002 17994768 May 11 19:57 octave-3.6.2-rc0.tar.gz
>
> Even if you have been building Octave regularly from the Mercurial
> archive, it would help if you could build from this release candidate
> and report any problems.  We could especially use reports for Windows
> and OS X systems.

On Windows/MSVC (VS2008):
==============================================
Summary:

  PASS  10176
  FAIL      6

There were 4 expected failures (see fntests.log for details).

Expected failures are known bugs.  Please help improve Octave
by contributing fixes for them.

There were 11 skipped tests (see fntests.log for details).
Skipped tests are features that are disabled in this version of Octave
because the needed libraries were not present when Octave was built.
==============================================

The failures are:
- 2 in data.cc, due to frexp(Inf) => NaN
- 1 in mappers.cc due to isprint(char(9)) == true
- 1 in mappers.cc as follows:

  ***** test
 x = [-1, 0, 1, Inf];
 v = [Inf, Inf, 0, Inf];
 assert (lgamma(x), v);
 assert (lgamma(single (x)), single (v));
!!!!! test failed
assert (lgamma (single (x)),single (v)) expected
   Inf   Inf     0   Inf
but got
       Inf + 3.14159i       Inf + 0.00000i   0.00000 + 0.00000i
Inf + 0.00000i
complex != real

Looking at the code in liboctave/lo-specfunc.cc, in rc_lgamma float
implementation, I think it's due to the fact that the variable sgngam
is not initialized (the same problem may appear in the double
implementation), but not all code paths are assigning it, so the final
test for < 0 is undefined (Note: HAVE_LGAMMAF_R is not defined).

- 1 in speed.m as follows:

  ***** test
 [order, n, T_f1, T_f2] = speed ("airy (x)", "x = rand (n, 10)", [100, 1000]);
 assert (isstruct (order));
 assert (size (order), [1, 1]);
 assert (fieldnames (order), {"p"; "a"});
 assert (isnumeric (n));
 assert (length (n) > 10);
 assert (isnumeric (T_f1));
 assert (size (T_f1), size (n));
 assert (isnumeric (T_f2));
 assert (length (T_f2) > 10);
!!!!! test failed
assert (length (n) > 10) failed

Executing the test code in octave results in the following, I don't
know where the problem is:

octave.exe:14> [order, n, T_f1, T_f2] = speed ("airy (x)", "x = rand
(n, 10)", [100, 1000])
order =

  scalar structure containing the fields:

    p = 9.6404e-005
    a =  0.014990

n =

    848   1000

T_f1 =

   0.015000   0.015000

T_f2 =

   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0

- 1 in test_system.m (I think the same happens on MinGW), as follows:

  ***** test
 orig_umask = umask (0);
 nm = tmpnam ();
 id = fopen (nm, "wb");
 s1 = stat (nm);
 fclose (id);
 unlink (nm);

 umask (777);
 nm = tmpnam ();
 id = fopen (nm, "wb");
 s2 = stat (nm);
 fclose (id);
 unlink (nm);

 assert (deblank (s1.modestr), "-rw-rw-rw-");
 assert (deblank (s2.modestr), "----------");
 # Restore original umask value
 umask (orig_umask);
!!!!! test failed
assert (deblank (s2.modestr),"----------") expected
----------
but got
-r--r--r--

During the test, one can also see a bunch of "permission denied"
messages printed.

Michael.