3.0.4 RC1

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

3.0.4 RC1

Jaroslav Hajek-2
Hello,

the 3.0.4 RC1 tarballs are available at
http://artax.karlin.mff.cuni.cz/~hajej2am/ulozna/octave/

waiting for build reports...

PS. This is almost certainly the last of 3.0.x series. Stay tuned for 3.2.0...


--
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1 (mingw report)

tmacchant
Hello

I have built octave-3.0.4RC1 on mingw by gcc-3.4.5.
(My original build. It is not the same as bejamin:)

The build was successful.

However, errors on fileparts should be fixed for octave native to windows.

Regards

Tatsuro

***************************************
Integrated test scripts:

1  c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\src/DLD-FUNCTIONS\rand.cc  PASS   56/57   FAIL 1
2  c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\src\ov-fcn-handle.cc  PASS    2/3    FAIL 1
3  c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\scripts/miscellaneous\fileparts.m  PASS    3/9  
FAIL 6

Fixed test scripts:

4  test_string.m .......................................... PASS  130/131  FAIL 1

Summary:

  PASS   4051
  FAIL      9

There were 2 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 fntest.log for details).
Skipped tests are features that are disabled in this version
of Octave as the needed libraries were not present when Octave
was built

1493 (of 1669) files have no tests.  Please help improve Octave by
contributing tests for these files (see the list in the file fntests.log).
***************

Note:
In my build, test_io.m has seemed to fall into an inifinte loop (this time also) so that it has been
skipped.

****
 1
****
c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\src/DLD-FUNCTIONS\rand.cc  PASS   56/57   FAIL 1

 randp("seed",1);
 assert(randp(1e9,1,6),[1000006208 1000012224 999981120 999963520 999963072 999981440],1e-6)
!!!!! test failed
error: assert (randp (1e9, 1, 6),[1000006208, 1000012224, 999981120, 999963520, 999963072,
999981440],1e-6) expected
  1.0000e+009  1.0000e+009  9.9998e+008  9.9996e+008  9.9996e+008  9.9998e+008
but got
  1.0000e+009  1.0000e+009  9.9998e+008  9.9996e+008  9.9996e+008  9.9998e+008
maximum absolute error 27 exceeds tolerance 1e-006

!!!!!!!!!!!
To use relative tolerance , the third argument is give as minus value : -1e-6:

So please apply

--- a/src/DLD-FUNCTIONS/rand.cc Fri Jun 16 15:46:22 2009 -0400
+++ b/src/DLD-FUNCTIONS/rand.cc Sat Jun 17 12:47:07 2009 +0900
@@ -950,7 +950,7 @@
 %!test
 %! % Test fixed state
 %! randp("seed",1);
-%! assert(randp(1e9,1,6),[1000006208 1000012224 999981120 999963520 999963072 999981440],1e-6)
+%! assert(randp(1e9,1,6),[1000006208 1000012224 999981120 999963520 999963072 999981440],-1e-6)
 %!test
 %! if (__random_statistical_tests__)
 %!   % statistical tests may fail occasionally.


****
 2
****
2  c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\src\ov-fcn-handle.cc  PASS    2/3    FAIL 1
This has been reported by Benjamin but no good solution seems to be found yet.

 a = 2;
 f = @(x) a + x;
 g = @(x) 2 * x;
 h = @log2;
 f2 = f;
 g2 = g;
 h2 = h;
 nm = tmpnam();
 unwind_protect
   save ("-text", nm, "f2", "g2", "h2");
   clear f2 g2 h2
   load (nm);
   assert (f(2),f2(2));
   assert (g(2),g2(2));
   assert (g(3),g2(3));
   unlink (nm);
   save ("-text", nm, "f2", "g2", "h2");
 unwind_protect_cleanup
   unlink (nm);
 end_unwind_protect
!!!!! test failed
error: function handle points to non-existent function

****
 3
****
  c:\usr\Tatsu\mingwhome\octaves\octave-3.0.4RC1\scripts/miscellaneous\fileparts.m  PASS    3/9  
FAIL 6
 [d, n, e] = fileparts ("/file.ext");
 assert (strcmp (d, "/") && strcmp (n, "file") && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, "/") && strcmp (n, "file") && strcmp (e, ".ext")) failed
  ***** test
 [d, n, e] = fileparts ("dir/file.ext");
 assert (strcmp (d, "dir") && strcmp (n, "file") && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, "dir") && strcmp (n, "file") && strcmp (e, ".ext")) failed
  ***** test
 [d, n, e] = fileparts ("./file.ext");
 assert (strcmp (d, ".") && strcmp (n, "file") && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, ".") && strcmp (n, "file") && strcmp (e, ".ext")) failed
  ***** test
 [d, n, e] = fileparts ("d1/d2/file.ext");
 assert (strcmp (d, "d1/d2") && strcmp (n, "file") && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, "d1/d2") && strcmp (n, "file") && strcmp (e, ".ext")) failed
  ***** test
 [d, n, e] = fileparts ("/d1/d2/file.ext");
 assert (strcmp (d, "/d1/d2") && strcmp (n, "file") && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, "/d1/d2") && strcmp (n, "file") && strcmp (e, ".ext")) failed
  ***** test
 [d, n, e] = fileparts ("/.ext");
 assert (strcmp (d, "/") && strcmp (n, char (zeros (1, 0))) && strcmp (e, ".ext"));
!!!!! test failed
error: assert (strcmp (d, "/") && strcmp (n, char (zeros (1, 0))) && strcmp (e, ".ext")) failed

*************************
The above seems to be filesep() is '\' but not '/' in windows.

octave.exe:11> [d, n, e] = fileparts ("\\file.ext")
d = \
n = file
e = .ext
octave.exe:12> filesep
ans = \

****
 4
****
For No.4, I think that I have already sent


--- a/ChangeLog Tue Aug 19 16:15:52 2008 -0400
+++ b/ChangeLog Fri Aug 22 13:37:16 2008 +0900
@@ -1,3 +1,7 @@
+2008-08-22  Tatsuro MATSUOKA  <tmacchant@...>
+
+       test/octave.test/string/isprint-1.m in test_string.m for fixed.
+
 2008-08-19  Jaroslav Hajek <highegg@...>
 
        Version 3.0.2 released.
diff -r 4908e39553b0 test/test_string.m
--- a/test/test_string.m        Tue Aug 19 16:15:52 2008 -0400
+++ b/test/test_string.m        Fri Aug 22 13:37:16 2008 +0900
@@ -532,7 +532,7 @@
 %! result = zeros (1, 128);
 %!
 %! result (33:127) = 1;
-%! if (ispc () && ! isunix ())
+%! if (size(findstr (octave_config_info('canonical_host_type'), 'msvc'),2))
 %!   result(10) = 1;
 %! endif
 %!



--------------------------------------
Power up the Internet with Yahoo! Toolbar.
http://pr.mail.yahoo.co.jp/toolbar/
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1 (cygwin)

tmacchant
In reply to this post by Jaroslav Hajek-2
Hello

I have built octave-3.0.4RC1 on mingw by gcc-3.4.4-sjlj.
The depndencies used is prepared in cygwin, most parts of which have been contructed by Marco
(Thanks!).

gcc-4 build was failed at lioctave concering suitesparse. So that I used this time gcc-3.4.4 -sjlj.
(Environment using gcc-3.4.4-dw2 is now only saved in a PC in the university.)

The build was successful.

Regards

Tatsuro

************
Integrated test scripts:

1.  src/DLD-FUNCTIONS/rand.cc .............................. PASS   56/57   FAIL 1  
2. scripts/specfun/nchoosek.m ............................. PASS    1/2    FAIL 1  

Fixed test scripts:

3. test_system.m .......................................... PASS  125/126  FAIL 1

Summary:

  PASS   4134
  FAIL      3

There were 2 expected failures (see fntests.log for details).

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

1493 (of 1670) files have no tests.  Please help improve Octave by
contributing tests for these files (see the list in the file fntests.log).
****************************************


No.1 is the same as that for the mingw case.
http://www.nabble.com/3.0.4-RC1-to21495212.html#a21512989

No.3 comes from my irregular setting of home directries.

No.2 is new error
>>>>> processing /cygdrive/c/usr/tatsu/mingwhome/octaves/octave-3.0.4RC1/scripts/specfun/nchoosek.m
  ***** assert (nchoosek(100,45), bincoeff(100,45))
!!!!! test failed
error: assert (nchoosek (100, 45),bincoeff (100, 45)) expected
 6.1448e+28
but got
 6.1448e+28
values do not match

>From octave prompt,
octave:10> nchoosek(100,45)
ans =  6.14484712141365e+28
octave:11> bincoeff(100,45)
warning: A(I) = X: X does not have same shape as I
warning: A(I) = X: X does not have same shape as I
ans =  6.14484712141348e+28
octave:12> (nchoosek(100,45)-bincoeff(100,45))/ nchoosek(100,45)
warning: A(I) = X: X does not have same shape as I
warning: A(I) = X: X does not have same shape as I
ans =  2.83428762991993e-14              

Is relative reference of 3e-14 allowable?
If it is allowable,
The below change is desireble.
(nchoosek(100,45), bincoeff(100,45))
 |
 V
(nchoosek(100,45), bincoeff(100,45),-1e-13)

Have to I prepare in the form of changeset the above?



--------------------------------------
Power up the Internet with Yahoo! Toolbar.
http://pr.mail.yahoo.co.jp/toolbar/
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1 (cygwin)

Marco Atzeri

--- Tatsuro MATSUOKA  ha scritto:

> Hello
>
> I have built octave-3.0.4RC1 on mingw by
> gcc-3.4.4-sjlj.
> The depndencies used is prepared in cygwin, most
> parts of which have been contructed by Marco
> (Thanks!).
>
> gcc-4 build was failed at lioctave concering
> suitesparse. So that I used this time gcc-3.4.4
> -sjlj.

It is probably due to the change for Suitesparse-3.2.
On cygwin the last one is 3.1.
I will update both the package for the 3.0.4 release.

> (Environment using gcc-3.4.4-dw2 is now only saved
> in a PC in the university.)
>
> The build was successful.
>
> Regards
>
> Tatsuro




     
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1 (cygwin)

tmacchant
Hello Marco Atzeri

Thank you for your reply.

--- Marco Atzeri <[hidden email]> wrote:

> It is probably due to the change for Suitesparse-3.2.
> On cygwin the last one is 3.1.
> I will update both the package for the 3.0.4 release.
>

I will wait for your 3.0.4 release. As far as my building, there is no intrinsic problem for 3.0.4
build on the cygwin. Two errors in fntest are the problems of numerical error and seem to be within
allowance.

Regards

Tatsuro



--------------------------------------
Power up the Internet with Yahoo! Toolbar.
http://pr.mail.yahoo.co.jp/toolbar/
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1

Rafael Laboissiere
In reply to this post by Jaroslav Hajek-2
* Jaroslav Hajek <[hidden email]> [2009-01-16 10:13]:

> Hello,
>
> the 3.0.4 RC1 tarballs are available at
> http://artax.karlin.mff.cuni.cz/~hajej2am/ulozna/octave/

Thanks for this RC tarball.  It builds fine and all tests pass when built as
a Debian package for the unstable distribution.

However, I noticed something strange: several graphics files distributed in
the tarball are empty.  E.g.:

$ tar tfvz ../octave-3.0.4-rc1.tar.gz | grep delaunay.png
-rw-rw-rw- hajek/users       0 2009-01-16 07:58 octave-3.0.4/doc/interpreter/HTML/delaunay.png
-rw-rw-rw- hajek/users       0 2009-01-16 07:58 octave-3.0.4/doc/interpreter/delaunay.png

Is this normal?
 
--
Rafael
Reply | Threaded
Open this post in threaded view
|

Re: 3.0.4 RC1

Benjamin Lindner
In reply to this post by Jaroslav Hajek-2

Jaroslav Hajek wrote:
> Hello,
>
> the 3.0.4 RC1 tarballs are available at
> http://artax.karlin.mff.cuni.cz/~hajej2am/ulozna/octave/
>
> waiting for build reports...
>

Building on mingw32/win32 using TDM's gcc-4.3.0-2 succeeds fine.
make check yields the following 3 errors:

d:\files\admin\octaveforge_svn\trunk\octave-forge\admin\Windows\mingw32\octave\octave-3.0.x\src\ov-fcn-handle.cc
  PASS    2/3    FAIL 1
 
d:\files\admin\octaveforge_svn\trunk\octave-forge\admin\Windows\mingw32\octave\octave-3.0.x\scripts/miscellaneous\fileparts.m
  PASS    3/9    FAIL 6

  test_poly.m ............................................ PASS   48/49
   FAIL 1

The first one is known, and not yet solved. Let's hope for 3.2.0 :)

The second one I see is also reported by Tatsuro and according to the
thread starting at
http://www.nabble.com/fileparts.m-in-octave-%28mingw-and-msvc%29-to21524425.html
it should be fixed. I need to update my repos and try again.

The third one tells
 >>>>> processing test_poly
   ***** error polyfit ([1, 2; 3, 4], [1, 2; 3, 4], 4);
!!!!! expected <.> but got no error

I'm not sure what this test should do, and why it should fail.

benjamin