Re: Default merged to stable for upcoming 4.2 release

[Jacopo Corno]

Dear all,

While fixing some compatibility issues for the release of odepkg I realized there is a minor bug in the core private functions, namely the function arguments are not passed for the first step.

I attach a patch to fix the problem. Can this go in the release or is it too late?

I'll keep working on the package to make it ready ASAP.

Cheers,
Jack

On 05/10/2016 10:16, Chiara Segala wrote:

2016-10-04 16:47 GMT+02:00 c. <address@hidden>:

> I fully support this. Thanks! Let me know, if you need specific help with something.

If you and Jacopo could test odepkg with the current release candidates and prepare a release of the package that would be great!
Marco, how long would it take to merge chiara's new solvers into odepkg?

> One comment: we should also discuss what to do with dassl and daspk and when. I think we agreed to move them to odepkg eventually (and possibly change dassl_options and daspk_options to odeset for better compatibility).

I don't remember this discussion, so I'm not sure how hard it is to do this move ...

I suspect it might be hard to do because the DAE class they are based upon is very
closely entagled with liboctave and it might take quite some work to turn it into
a dynamic extension. But all this needs to be checked more carefully ...

Anyway I would definitely do this before at least Octave 4.6, an implicit solver
for descriptor systems is a very important feature and until ode15{s,i} are very
well tested I wouldn't remove the existing alternative from core ...

c.

Dear all,

I have to fix some things of my solvers,
then I will adapt them to the changes made by Francesco during the GSoC.
My work should be ready in about a week.

Chiara

-- 

Jacopo Corno M.Sc.
Graduate School CE
S4|10-232
Dolivostraße 15
D-64293 Darmstadt / Germany

Email: address@hidden

22608.patch
Description: Text Data