[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Proposal for built in tests.
From: |
Robert T. Short |
Subject: |
Re: Proposal for built in tests. |
Date: |
Tue, 23 Jun 2009 13:34:26 -0700 |
User-agent: |
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.21) Gecko/20090402 SeaMonkey/1.1.16 |
In order of importance (to me).
1. I just discovered this while I was working this morning and have
not characterized it carefully. I will do something in "test" to
duplicate it.
function x = somefunction( someobj )
set( someobj, 'someproperty', somevalue );
x = a computation;
endfunction
obj = SomeObject();
x = somefunction( obj );
- the property change performed in somefunction propagates outside
the function.
This shouldn't happen (and doesn't in MATLAB).
2. The loadobj function in the test classes causes a segmentation fault
(there is a comment
to this effect in the test script). In general, I need to redo the
way the loadobj method
is done to be compatible with MATLAB. MATLAB expects loadobj to
receive a struct
if the exemplar entry is not available and that is not the current
approach.
3. The following fails. Even though MATLAB loads the save file
properly, I am 99% sure
this is not an OOP bug, but rather a problem with the save
function. I went through
the .mat file and the count for the 'Cork' element is not correct.
./runoctave
> grk = Gork(); % Gork is a test class
> save -v6 'gorksave.mat' grk;
> clear;
> load gorksave.mat
4. addpath 'xxx/xxx/@xxx'
should error.
John W. Eaton wrote:
On 23-Jun-2009, Robert T. Short wrote:
| This won't happen quickly but is something I can do when I have just a
| few minutes to work on octave. I have a couple of nasty little OOP bugs
| to track down so when I have real time I will work on those.
What are the bugs?
jwe