Discussion about the development and maintenence of Octave.

Robert T. Short | 1 Jun 2009 06:17

Patch for OOP example.

Attached is a patch to the polynomial object example.  This fixes 
several things.

-  Tabs in the source files made the documentation goofy.
-  There were several methods that didn't work.  They seem to work now.
-  Some simplifications.

How is this going to work without savannah?  My tip is out of date and 
will probably get worse.

Should we just hold these patches?

Bob
--
Robert T. Short, Ph.D.
PhaseLocked Systems

# HG changeset patch
# User Robert T. Short <octave <at> phaselockedsystems.com>
# Date 1243829491 25200
# Node ID 97e9fcb1a36e70652a865c7cee966614339ecbfc
# Parent  e5fb8a804aa1e107e79e5203978a3a4bbdbda4f7
  * examples/ <at> polynomial: removed tabs from all functions so
    documentation would look right.  Fixed a bunch of methods
    that didn't work.  Added a method referred to in the docs
    that didn't exist.

diff -r e5fb8a804aa1 -r 97e9fcb1a36e examples/ <at> polynomial/display.m
--- a/examples/ <at> polynomial/display.m	Sun May 24 10:23:07 2009 -0700

(Continue reading)

headers

Jaroslav Hajek | 1 Jun 2009 17:42

Re: Patch for OOP example.

On Mon, Jun 1, 2009 at 6:17 AM, Robert T. Short
<octave <at> phaselockedsystems.com> wrote:
> Attached is a patch to the polynomial object example.  This fixes several
> things.
>
> -  Tabs in the source files made the documentation goofy.
> -  There were several methods that didn't work.  They seem to work now.
> -  Some simplifications.
>
> How is this going to work without savannah?  My tip is out of date and will
> probably get worse.
>
> Should we just hold these patches?
>
> Bob
> --
> Robert T. Short, Ph.D.
> PhaseLocked Systems
>

John, would you please apply this patch to the velveeta repo, so that
it gets a definite ID? I want to get it into RC5 and I want to keep
release-3-2-x a mere clone of mainstream as long as possible. I assume
that after savannah is back working, you'll push from your backup repo
to it.

thanks

--

-- 
RNDr. Jaroslav Hajek

(Continue reading)

headers

John W. Eaton | 2 Jun 2009 07:36

Re: Patch for OOP example.

On  1-Jun-2009, Jaroslav Hajek wrote:

| John, would you please apply this patch to the velveeta repo, so that
| it gets a definite ID? I want to get it into RC5 and I want to keep
| release-3-2-x a mere clone of mainstream as long as possible. I assume
| that after savannah is back working, you'll push from your backup repo
| to it.

I'm traveling now and have no time to do it, plus I'm having trouble
even reaching my home system where my main archive is.  I don't have
anything that is not committed to the velveeta archive, so if you want
to clone that and add patches to your own copy, I will sync with you
when I return on Wednesday.

jwe

headers

Robert T. Short | 1 Jun 2009 19:10

Re: Patch for OOP example.

Jaroslav & John,

I can easily resubmit this patch later if it is a pain to deal with. The changes I made won't impact anything outside of the one example suite. Examples with errors are kind of a bummer, but no major problem.

Bob

Jaroslav Hajek wrote:

On Mon, Jun 1, 2009 at 6:17 AM, Robert T. Short <octave <at> phaselockedsystems.com> wrote:
Attached is a patch to the polynomial object example. This fixes several things. - Tabs in the source files made the documentation goofy. - There were several methods that didn't work. They seem to work now. - Some simplifications. How is this going to work without savannah? My tip is out of date and will probably get worse. Should we just hold these patches? Bob -- Robert T. Short, Ph.D. PhaseLocked Systems
John, would you please apply this patch to the velveeta repo, so that it gets a definite ID? I want to get it into RC5 and I want to keep release-3-2-x a mere clone of mainstream as long as possible. I assume that after savannah is back working, you'll push from your backup repo to it. thanks

headers

Jaroslav Hajek | 2 Jun 2009 10:21

Re: Patch for OOP example.

On Mon, Jun 1, 2009 at 7:10 PM, Robert T. Short
<octave <at> phaselockedsystems.com> wrote:
> Jaroslav & John,
>
> I can easily resubmit this patch later if it is a pain to deal with.  The
> changes I made won't impact anything outside of the one example suite.
> Examples with errors are kind of a bummer, but no major problem.
>
> Bob
>

I applied the patch to the release-3-2-x repository.
I noticed that a few methods (subsasgn, subsref) still didn't work or
worked differently than mentioned in the docs.
In particular, I updated subsasgn and subsref to handle the full index
chain, i.e. stuff like
a.poly(1:2) = [3,4]
will work correctly.
I added a notice about this to the docs. I also changed the {}
semantics so that a{i} means the coefficient of x^i, as the
documentation says. a(i) = x gripes (makes no sense).
There's still room for more error checks and improvements (such as
stripping leading zeros after assignments), but I hope the examples
need not be perfect.

cheers

--

-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz

headers

Robert T. Short | 2 Jun 2009 15:15

Re: Patch for OOP example.

Cool. Thanks.

Examples should be simple and easy to follow, not perfectly designed code! Extensive error checking (etc.) often takes away from the message you are trying to convey.

Bob

Jaroslav Hajek wrote:

On Mon, Jun 1, 2009 at 7:10 PM, Robert T. Short <octave <at> phaselockedsystems.com> wrote:
Jaroslav & John, I can easily resubmit this patch later if it is a pain to deal with. The changes I made won't impact anything outside of the one example suite. Examples with errors are kind of a bummer, but no major problem. Bob
I applied the patch to the release-3-2-x repository. I noticed that a few methods (subsasgn, subsref) still didn't work or worked differently than mentioned in the docs. In particular, I updated subsasgn and subsref to handle the full index chain, i.e. stuff like a.poly(1:2) = [3,4] will work correctly. I added a notice about this to the docs. I also changed the {} semantics so that a{i} means the coefficient of x^i, as the documentation says. a(i) = x gripes (makes no sense). There's still room for more error checks and improvements (such as stripping leading zeros after assignments), but I hope the examples need not be perfect. cheers

headers

John W. Eaton | 2 Jun 2009 15:02

Re: Patch for OOP example.

On  2-Jun-2009, Jaroslav Hajek wrote:

| I applied the patch to the release-3-2-x repository.
| I noticed that a few methods (subsasgn, subsref) still didn't work or
| worked differently than mentioned in the docs.
| In particular, I updated subsasgn and subsref to handle the full index
| chain, i.e. stuff like
| a.poly(1:2) = [3,4]
| will work correctly.

Savannah seems to be back now, so please go ahead and push any 3.2.x
changes there too, and then I'll update from there.

Thanks,

jwe

headers

Jaroslav Hajek | 2 Jun 2009 15:15

Re: Patch for OOP example.

On Tue, Jun 2, 2009 at 3:02 PM, John W. Eaton <jwe <at> octave.org> wrote:
> On  2-Jun-2009, Jaroslav Hajek wrote:
>
> | I applied the patch to the release-3-2-x repository.
> | I noticed that a few methods (subsasgn, subsref) still didn't work or
> | worked differently than mentioned in the docs.
> | In particular, I updated subsasgn and subsref to handle the full index
> | chain, i.e. stuff like
> | a.poly(1:2) = [3,4]
> | will work correctly.
>
> Savannah seems to be back now, so please go ahead and push any 3.2.x
> changes there too, and then I'll update from there.
>
> Thanks,
>
> jwe
>

Okay, done. 129 changesets were missing.

regards

--

-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz

headers

John W. Eaton | 2 Jun 2009 15:26

Re: Patch for OOP example.

On  2-Jun-2009, Jaroslav Hajek wrote:

| Okay, done. 129 changesets were missing.

Ugh.  It's good to have some redundancy and/or backups...

jwe

headers

Jaroslav Hajek | 2 Jun 2009 15:33

Re: Patch for OOP example.

On Tue, Jun 2, 2009 at 3:26 PM, John W. Eaton <jwe <at> octave.org> wrote:
> On  2-Jun-2009, Jaroslav Hajek wrote:
>
> | Okay, done. 129 changesets were missing.
>
> Ugh.  It's good to have some redundancy and/or backups...
>
> jwe
>

Yeah. Fortunately, with Mercurial, we actually have dozens of backups
spread over the world :)

--

-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz

headers

Rob Mahurin | 1 Jun 2009 16:54

Re: Patch for OOP example.

On Jun 1, 2009, at 12:17 AM, Robert T. Short wrote:
> How is this going to work without savannah?  My tip is out of date  
> and will probably get worse.
>
> Should we just hold these patches?

Pull from jwe's backup repository to stay current.

$ hg pull
pulling from http://www.octave.org/hg/octave
abort: error: Connection refused
$ hg pull http://velveeta.che.wisc.edu/cgi-bin/hgwebdir.cgi/octave
pulling from http://velveeta.che.wisc.edu/cgi-bin/hgwebdir.cgi/octave
searching for changes
[...]

Rob

--

-- 
Rob Mahurin
Department of Physics and Astronomy
University of Tennessee 		865 207 2594
Knoxville, TN 37996 			rob <at> utk.edu

headers

Robert T. Short | 1 Jun 2009 17:16

Re: Patch for OOP example.

Thanks.

Rob Mahurin wrote:
> On Jun 1, 2009, at 12:17 AM, Robert T. Short wrote:
>> How is this going to work without savannah?  My tip is out of date 
>> and will probably get worse.
>>
>> Should we just hold these patches?
>
> Pull from jwe's backup repository to stay current.
>
> $ hg pull
> pulling from http://www.octave.org/hg/octave
> abort: error: Connection refused
> $ hg pull http://velveeta.che.wisc.edu/cgi-bin/hgwebdir.cgi/octave
> pulling from http://velveeta.che.wisc.edu/cgi-bin/hgwebdir.cgi/octave
> searching for changes
> [...]
>
> Rob
>