OpenSolaris - update - audio check

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

OpenSolaris - update - audio check

Jim Langston-2
Hi all,

There is an issue during testing with .wav
files not reading which I posted to the alias
awhile ago titled - audio check -, I have isolated
the problem, created a test case and I will be
submitting a bug against the Sun C++ compiler.

In the meantime, all the other tests pass, and I
just need to start finding the plotting libraries to run
on OpenSolaris and rebuild Octave with those libraries
to do plotting.

For those on OpenSolaris, my development env:

OpenSolaris 2008.11

w/

Sun Studio 2008/10/22

w/

Octave 3.0.3

Jim

--
/////////////////////////////////////////////

Jim Langston
Sun Microsystems, Inc.

(877) 854-5583 (AccessLine)
(513) 702-4741 (Cell)
AIM: jl9594
[hidden email]

_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

Jim Langston-2
Hi all,

Could someone take a look at the dialog I've been having
with the Sun Developer Forum on underflow, this has to do
with what is happening with reading audio files and what
feedback I'm getting back on the forum.

http://forums.sun.com/thread.jspa?messageID=10533624

I would appreciate any thoughts on a possible solution

Thanks,

Jim

/////////////////////////



Jim Langston wrote:

> Hi all,
>
> There is an issue during testing with .wav
> files not reading which I posted to the alias
> awhile ago titled - audio check -, I have isolated
> the problem, created a test case and I will be
> submitting a bug against the Sun C++ compiler.
>
> In the meantime, all the other tests pass, and I
> just need to start finding the plotting libraries to run
> on OpenSolaris and rebuild Octave with those libraries
> to do plotting.
>
> For those on OpenSolaris, my development env:
>
> OpenSolaris 2008.11
>
> w/
>
> Sun Studio 2008/10/22
>
> w/
>
> Octave 3.0.3
>
> Jim
>
>  


--
/////////////////////////////////////////////

Jim Langston
Sun Microsystems, Inc.

(877) 854-5583 (AccessLine)
(513) 702-4741 (Cell)
AIM: jl9594
[hidden email]

_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

John W. Eaton
Administrator
On  5-Dec-2008, Jim Langston wrote:

| Could someone take a look at the dialog I've been having
| with the Sun Developer Forum on underflow, this has to do
| with what is happening with reading audio files and what
| feedback I'm getting back on the forum.
|
| http://forums.sun.com/thread.jspa?messageID=10533624
|
| I would appreciate any thoughts on a possible solution

If I understand correctly from that thread you are saying that there
may be a bug in the way Octave's c_file_ptr_buf class is implemented.
It's certainly possible that there is a bug, but I don't know enough
about how these functions are supposed to work.  I admit to simply
copying and modifying some libstdc++ code to create the c_file_ptr_buf
class, so I didn't really understand it fully when I wrote it, but it
seemed to work properly (back in the day, there wasn't exactly much
documentation for the stream library internals...).  I don't object to
changing the c_file_ptr_buf class, but before making changes I'd like
to have some definitive statement about what the proper implementation
is.

jwe
_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

Jim Langston-2
Hi John,

Thanks, I have been looking at this from a couple of
directions, the first, that there is a bug in the Studio
compiler, which is forcing the underflow to loop, the
second, from the Apache C++ stdlib perspective, which
is where most of the dialog is coming from. From what
I understand, if underflow is handled via an underflow method,
then a buffer space must be set up to deal with the stream,
otherwise, the state is unknown. I understand what underflow
is trying to do, which I thought was fine, but the Apache C++
engineer seemed adamant that the buffer must be handled
in the way he was prescribing. I have not looked into re-writing the
underflow method, that is, not until I get feedback from the Sun compiler
engineers on a possible bug.

I was also thinking of trying to re-compile everything with Apache C++
stdlib
to see if what the engineer thought would happen, does. But, I have not
done so yet.

Thanks,

Jim

/////////////////////

 John W. Eaton wrote:

> On  5-Dec-2008, Jim Langston wrote:
>
> | Could someone take a look at the dialog I've been having
> | with the Sun Developer Forum on underflow, this has to do
> | with what is happening with reading audio files and what
> | feedback I'm getting back on the forum.
> |
> | http://forums.sun.com/thread.jspa?messageID=10533624
> |
> | I would appreciate any thoughts on a possible solution
>
> If I understand correctly from that thread you are saying that there
> may be a bug in the way Octave's c_file_ptr_buf class is implemented.
> It's certainly possible that there is a bug, but I don't know enough
> about how these functions are supposed to work.  I admit to simply
> copying and modifying some libstdc++ code to create the c_file_ptr_buf
> class, so I didn't really understand it fully when I wrote it, but it
> seemed to work properly (back in the day, there wasn't exactly much
> documentation for the stream library internals...).  I don't object to
> changing the c_file_ptr_buf class, but before making changes I'd like
> to have some definitive statement about what the proper implementation
> is.
>
> jwe
> _______________________________________________
> Help-octave mailing list
> [hidden email]
> https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
>  


--
/////////////////////////////////////////////

Jim Langston
Sun Microsystems, Inc.

(877) 854-5583 (AccessLine)
(513) 702-4741 (Cell)
AIM: jl9594
[hidden email]

_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

Jim Langston-2
I have gotten a patch for the underflow loop from the compiler
engineers, the underflow no longer hangs, so the tests have been able
to proceed, here are the results, any thoughts ??

Thanks,

Jim

//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

[hidden email] gmake check
gmake -f octMakefile check
gmake[1]: Entering directory `/export/home/langston/OCTAVE/octave-3.0.3'
gmake -C test check
gmake[2]: Entering directory `/export/home/langston/OCTAVE/octave-3.0.3/test'
../run-octave --norc --silent --no-history ./fntests.m .

Integrated test scripts:

  src/DLD-FUNCTIONS/bsxfun.cc ............................ PASS   55/55 
  src/DLD-FUNCTIONS/cellfun.cc ........................... PASS   31/31 
  src/DLD-FUNCTIONS/conv2.cc ............................. PASS    1/1  
  src/DLD-FUNCTIONS/dispatch.cc .......................... PASS   13/13 
  src/DLD-FUNCTIONS/fft.cc ............................... PASS    9/9  
  src/DLD-FUNCTIONS/fsolve.cc ............................ PASS    4/4  
  src/DLD-FUNCTIONS/luinc.cc .............................
  src/DLD-FUNCTIONS/matrix_type.cc ....................... PASS   51/51 
  src/DLD-FUNCTIONS/max.cc ............................... PASS   12/12 
  src/DLD-FUNCTIONS/rand.cc .............................. PASS   57/57 
  src/DLD-FUNCTIONS/regexp.cc ............................ PASS   90/90 
  src/DLD-FUNCTIONS/spkron.cc ............................ PASS    2/2  
  src/DLD-FUNCTIONS/spqr.cc .............................. PASS    2/2  
  src/DLD-FUNCTIONS/tsearch.cc ........................... PASS    6/6  
  src/data.cc ............................................ PASS   10/20   FAIL 10
  src/ov-fcn-handle.cc ................................... PASS    3/3  
  src/ov-struct.cc ....................................... PASS   16/16 
  src/pt-idx.cc .......................................... PASS    5/5  
  src/strfns.cc .......................................... PASS   22/22 
  src/syscalls.cc ........................................ PASS    1/1  
  liboctave/CMatrix.cc ................................... PASS    6/6  
  liboctave/dMatrix.cc ................................... PASS    6/6  
  scripts/audio/wavwrite.m ............................... PASS    0/3    FAIL 3
  scripts/elfun/acosd.m .................................. PASS    3/3  
  scripts/elfun/acotd.m .................................. PASS    3/3  
  scripts/elfun/acscd.m .................................. PASS    3/3  
  scripts/elfun/asecd.m .................................. PASS    3/3  
  scripts/elfun/asind.m .................................. PASS    3/3  
  scripts/elfun/atand.m .................................. PASS    3/3  
  scripts/elfun/cosd.m ................................... PASS    5/5  
  scripts/elfun/cotd.m ................................... PASS    5/5  
  scripts/elfun/cscd.m ................................... PASS    5/5  
  scripts/elfun/secd.m ................................... PASS    5/5  
  scripts/elfun/sind.m ................................... PASS    5/5  
  scripts/elfun/tand.m ................................... PASS    5/5  
  scripts/general/__isequal__.m .......................... PASS   26/26 
  scripts/general/accumarray.m ........................... PASS   11/11 
  scripts/general/arrayfun.m ............................. PASS    1/1  
  scripts/general/bitcmp.m ............................... PASS   15/15 
  scripts/general/blkdiag.m .............................. PASS    8/8  
  scripts/general/cell2mat.m ............................. PASS    2/2  
  scripts/general/circshift.m ............................ PASS    4/4  
  scripts/general/cplxpair.m ............................. PASS   13/13 
  scripts/general/cumtrapz.m ............................. PASS    6/6  
  scripts/general/deal.m ................................. PASS    2/2  
  scripts/general/interp1.m .............................. PASS  144/144
  scripts/general/interp2.m .............................. PASS    6/6  
  scripts/general/interp3.m .............................. PASS    1/1  
  scripts/general/interpft.m ............................. PASS    2/5    FAIL 3
  scripts/general/issymmetric.m .......................... PASS   11/12   FAIL 1
  scripts/general/logical.m .............................. PASS    8/8  
  scripts/general/lookup.m ............................... PASS   13/13 
  scripts/general/mod.m .................................. PASS   25/25 
  scripts/general/nthroot.m .............................. PASS    3/3  
  scripts/general/polyarea.m ............................. PASS    4/4  
  scripts/general/rem.m .................................. PASS    4/4  
  scripts/general/repmat.m ............................... PASS   16/16 
  scripts/general/runlength.m ............................ PASS    1/1  
  scripts/general/sortrows.m ............................. PASS    2/2  
  scripts/general/structfun.m ............................ PASS    1/1  
  scripts/general/sub2ind.m .............................. PASS   24/24 
  scripts/general/trapz.m ................................ PASS    6/6  
  scripts/geometry/convhull.m ............................ PASS    0/1    FAIL 1
  scripts/geometry/delaunay.m ............................ PASS    0/1    FAIL 1
  scripts/geometry/delaunay3.m ........................... PASS    0/1    FAIL 1
  scripts/geometry/dsearch.m ............................. PASS    2/2  
  scripts/geometry/dsearchn.m ............................ PASS    6/6  
  scripts/geometry/griddata.m ............................ PASS    0/1    FAIL 1
  scripts/geometry/griddata3.m ........................... PASS    0/2    FAIL 2
  scripts/geometry/griddatan.m ........................... PASS    0/2    FAIL 2
  scripts/geometry/tsearchn.m ............................ PASS    5/5  
  scripts/image/imshow.m ................................. PASS    3/3  
  scripts/linear-algebra/cond.m .......................... PASS    5/5  
  scripts/linear-algebra/condest.m ....................... PASS    0/4    FAIL 4
  scripts/linear-algebra/onenormest.m .................... PASS    0/4    FAIL 4
  scripts/miscellaneous/bincoeff.m ....................... PASS    3/3  
  scripts/miscellaneous/compare_versions.m ............... PASS   30/30 
  scripts/miscellaneous/fullfile.m ....................... PASS   19/19 
  scripts/miscellaneous/getfield.m ....................... PASS    2/2  
  scripts/miscellaneous/inputname.m ...................... PASS    2/2  
  scripts/miscellaneous/setfield.m ....................... PASS    2/2  
  scripts/miscellaneous/substruct.m ...................... PASS    6/6  
  scripts/plot/hist.m .................................... PASS    7/7  
  scripts/plot/orient.m .................................. PASS    7/7  
  scripts/polynomial/pchip.m ............................. PASS    7/7  
  scripts/polynomial/residue.m ........................... PASS    4/4  
  scripts/polynomial/spline.m ............................ PASS   16/16 
  scripts/set/intersect.m ................................ PASS    1/1  
  scripts/set/ismember.m ................................. PASS   31/31 
  scripts/set/setdiff.m .................................. PASS    7/7  
  scripts/set/setxor.m ................................... PASS    1/1  
  scripts/set/unique.m ................................... PASS    9/9  
  scripts/signal/detrend.m ............................... PASS    3/3  
  scripts/signal/freqz.m ................................. PASS    3/3  
  scripts/signal/sinetone.m .............................. PASS    1/1  
  scripts/sparse/nonzeros.m .............................. PASS    2/2  
  scripts/sparse/pcg.m ................................... PASS    0/4    FAIL 4
  scripts/sparse/pcr.m ................................... PASS    0/3    FAIL 3
  scripts/sparse/speye.m ................................. PASS    5/5  
  scripts/sparse/spfun.m ................................. PASS    4/4  
  scripts/sparse/spones.m ................................ PASS    3/3  
  scripts/sparse/spstats.m ............................... PASS    1/1  
  scripts/specfun/betaln.m ............................... PASS    3/3  
  scripts/specfun/factorial.m ............................ PASS    5/5  
  scripts/specfun/legendre.m ............................. PASS    1/1  
  scripts/special-matrix/hadamard.m ...................... PASS    2/3    FAIL 1
  scripts/special-matrix/hankel.m ........................ PASS    5/5  
  scripts/special-matrix/magic.m ......................... PASS    3/4    FAIL 1
  scripts/statistics/base/mode.m ......................... PASS   12/12 
  scripts/statistics/distributions/binornd.m ............. PASS    2/2  
  scripts/statistics/tests/kruskal_wallis_test.m ......... PASS    1/1  
  scripts/strings/blanks.m ............................... PASS    3/3  
  scripts/strings/dec2base.m ............................. PASS    4/4  
  scripts/strings/index.m ................................ PASS   10/10 
  scripts/strings/mat2str.m .............................. PASS    3/3  
  scripts/strings/strcat.m ............................... PASS    3/3  
  scripts/strings/strtok.m ............................... PASS   26/26 
  scripts/strings/strvcat.m .............................. PASS    3/3  
  scripts/testfun/assert.m ............................... PASS   55/55 
  scripts/testfun/example.m .............................. PASS    4/4  
  scripts/testfun/test.m ................................. PASS   46/46 
  scripts/time/calendar.m ................................ PASS    2/2  
  scripts/time/datenum.m ................................. PASS    6/6  
  scripts/time/datestr.m ................................. PASS   32/32 
  scripts/time/datevec.m ................................. PASS   11/11 
  scripts/time/eomday.m .................................. PASS   11/11 
  scripts/time/etime.m ................................... PASS    5/5  
  scripts/time/weekday.m ................................. PASS   19/19 

Fixed test scripts:

error: `xf' undefined near line 100 column 14


gmake[2]: Leaving directory `/export/home/langston/OCTAVE/octave-3.0.3/test'
gmake[1]: Leaving directory `/export/home/langston/OCTAVE/octave-3.0.3'
[hidden email]

////////////////////////////////////////////////////////////////////////////////

Jim Langston wrote:
Hi John,

Thanks, I have been looking at this from a couple of
directions, the first, that there is a bug in the Studio
compiler, which is forcing the underflow to loop, the
second, from the Apache C++ stdlib perspective, which
is where most of the dialog is coming from. From what
I understand, if underflow is handled via an underflow method,
then a buffer space must be set up to deal with the stream,
otherwise, the state is unknown. I understand what underflow
is trying to do, which I thought was fine, but the Apache C++
engineer seemed adamant that the buffer must be handled
in the way he was prescribing. I have not looked into re-writing the
underflow method, that is, not until I get feedback from the Sun compiler
engineers on a possible bug.

I was also thinking of trying to re-compile everything with Apache C++ 
stdlib
to see if what the engineer thought would happen, does. But, I have not
done so yet.

Thanks,

Jim

/////////////////////

 John W. Eaton wrote:
  
On  5-Dec-2008, Jim Langston wrote:

| Could someone take a look at the dialog I've been having
| with the Sun Developer Forum on underflow, this has to do
| with what is happening with reading audio files and what
| feedback I'm getting back on the forum.
| 
| http://forums.sun.com/thread.jspa?messageID=10533624
| 
| I would appreciate any thoughts on a possible solution

If I understand correctly from that thread you are saying that there
may be a bug in the way Octave's c_file_ptr_buf class is implemented.
It's certainly possible that there is a bug, but I don't know enough
about how these functions are supposed to work.  I admit to simply
copying and modifying some libstdc++ code to create the c_file_ptr_buf
class, so I didn't really understand it fully when I wrote it, but it
seemed to work properly (back in the day, there wasn't exactly much
documentation for the stream library internals...).  I don't object to
changing the c_file_ptr_buf class, but before making changes I'd like
to have some definitive statement about what the proper implementation
is.

jwe
_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
  
    


  


-- 
/////////////////////////////////////////////

Jim Langston
Sun Microsystems, Inc.

(877) 854-5583 (AccessLine)
(513) 702-4741 (Cell)
AIM: jl9594
[hidden email]

_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

John W. Eaton
Administrator
On 22-Jan-2009, Jim Langston wrote:

| I have gotten a patch for the underflow loop from the compiler
| engineers, the underflow no longer hangs, so the tests have been able
| to proceed, here are the results, any thoughts ??

You will have to look at the test/fntest.log file to find more detail
about the failures.

jwe
_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

Jim Langston-2
Thanks John, that cleared up alot of the failures

Jim

///////////////

John W. Eaton wrote:

> On 22-Jan-2009, Jim Langston wrote:
>
> | I have gotten a patch for the underflow loop from the compiler
> | engineers, the underflow no longer hangs, so the tests have been able
> | to proceed, here are the results, any thoughts ??
>
> You will have to look at the test/fntest.log file to find more detail
> about the failures.
>
> jwe
>  

--
/////////////////////////////////////////////

Jim Langston
Sun Microsystems, Inc.

(877) 854-5583 (AccessLine)
(513) 702-4741 (Cell)
AIM: jl9594
[hidden email]


_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

fntestLog.zip (5K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: OpenSolaris - update - audio check

John W. Eaton
Administrator
On 22-Jan-2009, Jim Langston wrote:

| Thanks John, that cleared up alot of the failures

  >>>>> processing /export/home/langston/OCTAVE/octave-3.0.3/src/data.cc
    ***** assert(norm(x,1), 20);
  !!!!! test failed
  error: norm: only implemented for double and complex values
  shared variables {
    x =

       1  -3   4   5  -7

  }

This failure and the others like it are not expected.  The constant

  x = [1, -3, 4, 5, -7];

should have class "double".

    ***** test
   A = [1:10; 1:10]/10;
   wavwrite("a.wav", A);
   [B, samples_per_sec, bits_per_sample] = wavread("a.wav");
   assert(A,B, 10^(-4));
   assert(samples_per_sec, 8000);
   assert(bits_per_sample, 16);
  !!!!! test failed
  error: mx_el_ne: nonconformant arguments (op1 is 1x0, op2 is 1x4)
    ***** test
   A=[1:10; 1:10] / 10;
   wavwrite("a.wav", A, 4000);
   [B, samples_per_sec, bits_per_sample] = wavread("a.wav");
   assert(A,B, 10^(-4));
   assert(samples_per_sec, 4000);
   assert(bits_per_sample, 16);
  !!!!! test failed
  error: mx_el_ne: nonconformant arguments (op1 is 1x0, op2 is 1x4)
    ***** test
   A=[1:10; 1:10] / 10;
   wavwrite("a.wav", A, 4000, 8);
   [B, samples_per_sec, bits_per_sample] = wavread("a.wav");
   assert(A,B, 10^(-2));
   assert(samples_per_sec, 4000);
   assert(bits_per_sample, 8);
  !!!!! test failed
  error: mx_el_ne: nonconformant arguments (op1 is 1x0, op2 is 1x4)

These are not expected failures.

    ***** test
   x = -3:0.5:3;
   y = abs (sin (x));
   assert (convhull (x, y, {"s","Qci","Tcv","Pp"}), [1;7;13;12;11;10;4;3;2;1])
  !!!!! test failed
  error: convhulln: not available in this version of Octave
  >>>>> processing /export/home/langston/OCTAVE/octave-3.0.3/scripts/geometry/delaunay.m

This test and the others that rely on qhull should be protected with
testif.  I think that is already done in the current development sources.

There are a number of other tests that fail to meet expected
tolerances by a small amount.  I don't think those are cause for
concern.  Maybe we should relax the tolerances slightly.

You should try to find out why norm is failing as that is likely to
cause some trouble.

jwe
_______________________________________________
Help-octave mailing list
[hidden email]
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave