gmane.comp.gnu.octave.maintainers

Re: bzr for Octave

Thomas Weber — 2013-06-19T21:04:08

Oh, I know what they are doing. I am usually fine with either English or
German. The problem comes when I need to send someone instructions and a
website expects me to dive into my browser configuration to get the
English version. 

Thomas

Re: octave and postgres examples

Thomas Grzybowski — 2013-06-19T20:39:34

Hi,

I was thinking that this was a documentation concern, but I have taken 
your advice and posted to the list <help< at >octave.org>.

Thanks,
Thomas Grzybowski


On 06/19/2013 04:33 PM, Jordi Gutiérrez Hermoso wrote:

Re: octave and postgres examples

Jordi Gutiérrez Hermoso — 2013-06-19T20:33:44

Isn't this question better for the help list <help< at >octave.org> than
for the developer list?

- Jordi G. H.

Re: [GM-core] Does GraphicsMagick encapsulate littlecms functionality?

Patrick Noffke — 2013-06-19T20:31:58

On Wed, Jun 19, 2013 at 2:44 PM, Bob Friesenhahn
<bfriesen< at >simple.dallas.tx.us> wrote:

Thanks!

I'm copying the Octave maintainer's list.

Pat

octave and postgres examples

Thomas Grzybowski — 2013-06-19T20:23:09

Hi,

I have been trying to connect octave to my postgres database using the 
'database' Package - obtained from Octave-Forge - SourceForge, and 
written and maintained by Olaf Till.

I am able to create a connection object ("conn"), as so:

conn = pq_connect(setdbopts("host", "localhost","dbname", "tg", "user", 
"tg", "password", ""));

I know some SQL, but I am unfamiliar with the octave side of things; I 
am trying to use the command "pq_exec_params", but the documentation is 
a little sparse on the syntax.

I could really use some examples of reading and writing tables to and 
from octave arrays.  If anyone could post some example code that would 
be very much appreciated.

Thanks,
Thomas Grzybowski

RE: blog update #2

Marco Vassallo — 2013-06-19T20:14:16

From: marcovass89< at >hotmail.it
To: carlo.defalco< at >gmail.com
Subject: RE: blog update #2
Date: Tue, 18 Jun 2013 14:19:28 +0200
CC: octave-maintainers< at >octave.org

Hi,

thanks for the modifications you have made to the code.

I've some problem while compiling mshm_dolfin_write.cc:

mshm_dolfin_write.cc:47:38: error: ‘octave_map::const_iterator’ has no member named ‘matrix_value’

and I don't see why it is not working.

I'm testing the "mshm_dolfn_read" and "mshm_dolfin_write" functions after having added the treatment of the boundary cells/sides and I'll add them to the repository very soon.

About the mshm_dolfin_write I have some enormous "if" statement but at the moment I don't see any other possibilities.

For the mshm_dolfin_refine I've some problem while recovering the information from the refined mesh and so I'm now investigating better dolfin, which will be also useful for writing the blog post about boundaries.

Marco

Re: [Pkg-octave-devel] Bug#706376: Bug#706376: Bug#706376:Bug#706376: Bug#706376: octave: sparse matrix n*2^16

Jordi Gutiérrez Hermoso — 2013-06-19T19:53:42

We are assuming it in our code. In numel for one. And in places like
whatever processes A(idx) for a logical index idx. We're not making
special cases in these places, "if (sparse_type)
dont_linearly_index();"

Each of these places that assume that linear indexing works needs to
be patched to check for sparse types.


We are fixing one symptom of a larger bug, a bug that is present in
many different locations.


That would delay the problem for a nontrivial amount of time. It would
be nice, but it wouldn't fix the problem.

There are other things we could do.

(1) Avoid linear indexing for sparse matrices as much as possible,
i.e. check everywhere we can think of for sparse matrix types. You've
mentioned a few more places where this should be checked.

(2) Warn when creating sparse matrices with large indices that some
operations may not work, or clearly error out when those operations
are attempted.

(3) Abstract octave_idx_type so that it doesn't actually use 32-bit
ints for sparse matrices.


Well, numel

Re: SOCIS 2013

Nir Krakauer — 2013-06-19T19:52:05

Yes. See http://wiki.octave.org/Summer_of_Code_Project_Ideas for a
list. For SOCIS, there is special interest in proposals along the
lines of http://wiki.octave.org/Summer_of_Code_Project_Ideas#Geometric_integrators_for_Hamiltonian_Systems

Re: SOCIS 2013

c. — 2013-06-19T19:52:09

On 16 Jun 2013, at 18:06, Marco Cecchetti <mrcekets< at >gmail.com> wrote:

Yes.

Here: http://wiki.octave.org/Summer_of_Code_Project_Ideas

c.

Re: libgomp: Thread creation failed: Resource temporarily unavailable

Jordi Gutiérrez Hermoso — 2013-06-19T19:33:58

Please move this to the help list <help< at >octave.org>, not the maintainers' list.

- Jordi G. H.

libgomp: Thread creation failed: Resource temporarily unavailable

yuer — 2013-06-18T04:38:08

Hi,
I am running Octave 3.7.2 64bit on CentOS6. every time when I run a batch of
jobs, the first 12 jobs are fine but after 12 octave process are running,
new octave process will give error "libgomp: Thread creation failed:
Resource temporarily unavailable".

what's libgomp? is it called by octave or not?

any help?

thanks

yuer



--
View this message in context: http://octave.1599824.n4.nabble.com/libgomp-Thread-creation-failed-Resource-temporarily-unavailable-tp4654402.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.

Re: [Pkg-octave-devel] Bug#706376: Bug#706376: Bug#706376: Bug#706376:octave: sparse matrix n*2^16

David Bateman — 2013-06-19T18:06:14

Who is assuming linear indexing works for all matrix types ? Where
exactly is that stated ? If its in then manual that its a bug in the
manual as linear indexing can never work correctly in this case because
the underlying 32 integer type won't allowed it and in fact octave as it
stand throws an error !!!


The bug report was in "trace" that called "isempty" on a sparse matrix.
Neither function needs "numel" or "linear indexing". We aren't patching
around anything, we are fixing a bug


So you advocate everyone moving to 64-bit indexing ? In that case what
happens when we get the same bug report for s(2^32, 2^32) ? Is that a
"disease" too.  Certainly there is a limitation on linear indexing and
numel for sparse matrices. We should document it and make as many things
as possible work with sparse matrices and be done with it.


There is no disease, and unless you want to artificially limit the size
of sparse matrices that can be treated such that numel is less than 2^31
for 32 bit indexing and 2^63 for 64 bit

Re: quaternion package to hg

Lukas Reichlin — 2013-06-18T17:38:04

On 18.06.2013, at 16:48, Carnë Draug <carandraug< at >octave.org> wrote:


Hi Carnë

The packages

extra/quaternion-legacy
extra/control-legacy

are outdated and should not be included with quaternion and control. I just hesitated to delete them. Maybe you can leave them in the remaining svn repo.

Lukas

SOCIS 2013

Marco Cecchetti — 2013-06-16T16:06:26

Hi all,
I would like to know if your organization will apply for SOCIS this year,
and in case, if there is a project idea list somewhere.

Thanks,

Re: [Pkg-octave-devel] Bug#706376: Bug#706376: Bug#706376: octave:sparse matrix n*2^16

David Bateman — 2013-06-15T15:00:14

That a bit of a specious argument. "Because we can't solve problem B we
shouldn't solve problem A". Taking this argument to the absurd this
shouldn't work either

n = 2^16;
s = sparse (1:n,1:n,1);
t = s * s;

as t will have more elements than can be indexed with octave_idx_type.
Clear it works. So essentially you're saying that sparse matrices with
32-bit indexing and numel larger than 2^31 are useless!!

A lot of attention was made in the sparse implementation to not rely
either on linear indexing or the value of numel to avoid this issue. I'd
think that any function or operator that relies on either for sparse
matrices, or at least when it doesn't have to, is buggy.

Ed is right the is_empty method of the octave_base_value class should be
specialized for sparse matrices as in the attached changeset as it has
no need to rely on numel. Then any function that relies on isempty
should now work for sparse matrices.

David

# HG changeset patch
# User David Bateman <dbateman< at >free.fr>
# Date 1371308306 -7200
# No

Re: Google Summer of Code - LaTeX processing

Andrej Lojdl — 2013-06-19T19:28:27

I tried without -j 4 , still the same error. Here is diff as attachment.

--
Andrej

Re: Some basic hg tips to improve usability

Jordi Gutiérrez Hermoso — 2013-06-19T17:26:49

Now that you know how to more effectively use Mercurial, why not use
it to generate the required doc patch? ;-)

- Jordi G. H.

Re: Google Summer of Code - LaTeX processing

Patrick Noffke — 2013-06-19T17:19:44

You shouldn't need to re-run configure.

Try without -j 4.

Can you also post your current output of hg diff.  Please email as an
attachment.

Pat

Re: Audio playback and classes

Jordi Gutiérrez Hermoso — 2013-06-19T16:55:58

We use two underscores both front and back for functions that are
supposed to be internal and not called by the user.

We don't universally uphold this convention, but we should.

- Jordi G. H.

Re: Audio playback and classes

Vytautas Jancauskas — 2013-06-19T16:40:05

What is a good naming convention for these low level functions that
will get wrapped by octave class methods? I see < at >ftp adds two
underscores in front.

Re: Some basic hg tips to improve usability

Torsten — 2013-06-19T16:31:10

I suggest to add these tips to "Appendix D Contributing Guidelines" of
the octave manual.

Torsten