Re: [FORGE odepkg] Update, deprecation and FORTRAN warnings

[c.]

On 4 Mar 2014, at 11:29, Juan Pablo Carbajal <address@hidden> wrote:

> On Tue, Mar 4, 2014 at 9:45 AM, Juan Pablo Carbajal
> <address@hidden> wrote:
>> On Mon, Mar 3, 2014 at 6:13 PM, Mike Miller <address@hidden> wrote:
>>> On Mon, Mar 03, 2014 at 01:00:48 +0100, Juan Pablo Carbajal wrote:
>>>> On Mon, Mar 3, 2014 at 12:48 AM, Mike Miller <address@hidden> wrote:
>>>>> On Mon, Mar 03, 2014 at 00:32:20 +0100, Juan Pablo Carbajal wrote:
>>>>>> Mike,
>>>>>> 
>>>>>> I tried to import your patch and I get
>>>>>> 
>>>>>> 28 out of 28 hunks FAILED -- saving rejects to file src/cash.diff.rej
>>>>>> patching file src/daskr.diff
>>>>>> Hunk #1 FAILED at 0
>>>>>> 1 out of 1 hunks FAILED -- saving rejects to file src/daskr.diff.rej
>>>>>> abort: patch failed to apply
>>>>>> 
>>>>>> I am working with rev 31cb9aacfac5 tip
>>>>> 
>>>>> Hmm, I am on the same revision, and I was just able to import the patch
>>>>> again on that revision cleanly, I'm not sure why that doesn't work for
>>>>> you.
>>>>> 
>>>>> --
>>>>> mike
>>>> 
>>>> Mercurial Distributed SCM (version 2.8.2+3-082b2930fe2c)
>>> 
>>> Still not sure, but here are the cash.diff and daskr.diff files that you
>>> should be able to drop directly into the src directory. Does this work
>>> for you?
>>> 
>>> --
>>> mike
>> 
>> I created patches from your files and applied them... the files where
>> quite different, but everything seems to work...so no questions asked.
>> 
>> I pushed the changes, here odepkg compiles without warnings. I will
>> proceed to compare with the SoCiS repository.
>> 
>> @Mike: is there a diagnose log one could check to see why odepkg is
>> crashing in PowerPC (?)
> 
> @Carlo,
> 
> I updated the SoCiS repository with the changes. I did not marge cause
> I noticed that there are fewer functions tests than in the current OF
> package. Do you know why is that?

I also noticed that but I don't know the exact reason for that.
I'm cc-ing Roberto so he can comment on this.

Roberto, is there a reason why you reduced the number of tests?
Even for those functions you have almost rewritten from scratch
you could have kept most of the previous tests, what are the ones you
removed? Did you find them obsolete or irrelevant? were they testing
removed functionalities? were they using incompatible syntax?

> I also see loads of changes in ode45.m which I do not understand. Did
> they also edited this (and other older) functions?

Yes, one of the objectives of the project was to unify the structure
of the odexy solvers + odeset/odeget to make them easier to mantain, 
to change their interface to make it more matlab compatible and to 
eventually adapt them so they could be moved into core.

Of this three tasks I think the firt is essentially complete, the second is
partially done but the third still requires quite a bit of work.

I asked Roberto to update the project description in the wiki to reflect 
this status, but I haven't checked if he did that yet, he seems to be
very buisy with his PhD project at the moment.

> It would be good if the students try to do the merge. I am
> particularly concern for the missing tests.

Yes, I am waiting for Roberto to be a bit less buisy with his PhD project
so he can help with the merge.

@JPi, anyway, if you can help with reviewing and merging the code 
your contribution is very welcome!

c.