Re: Moving ode changes to default branch

[c.]

On 14 Oct 2016, at 19:51, Rik <address@hidden> wrote:

>> 
>> According to 'hg blame' the previous version had been in the repository at 
>> least since 
>> revision afe9c529760d of december 2015 and had gone unnoticed ...
> 
> "If a tree falls in the forest, does it make a sound?"  If the code is
> never used, does it matter that there are bugs in it?  Apparently no one
> used this feature of ode23 so maybe we shouldn't be that concerned.
>  In an ideal world, all Octave functions would have tests and there would be 
> 100%
> line coverage, but the team is pretty small.

Indeed, that is actually what I meant when I said I consider the current
code more "stable" than the previous version.

It appears that, although it has been in the core for almost one year now,
the code for the ode solvers has been mainly been tested by me, Jacopo, 
Marco, Sebastian.

I do know the development team is small so I'm not surprised very few
others had time to dive into that code, especially as the original 
implementation that was moved into core in December 2015 from odepkg 
was extremely hard to follow and had plenty of duplicated code.

Since the summer of code projects started development was mainly done
in Francesco's BitBucket repository
Testing went on there essentially by the same group of people who had been
looking at the code in core before plus the two SOC students Francesco, Chiara.

That is a nonnegligible team size compared to the total Octave workforce,
that's why I consider the current code more reliable than what was there before.

My recent changesets are mainly meant to review, adapt and apply changes made 
there into the savannah repository.

They contained fixes to issues that were discussed as far back as OctConf 
2103 and threads such as this one:

http://octave.1599824.n4.nabble.com/List-of-known-good-options-tp4674079p4674096.html

which is about the same age. 

4.2 will be the first Octave version to include the Matlab style
ode solvers in the core, this in some sense guarantees there can be
no "regressions" caused by any change to them ;)

As this is the first release of odexx.m with Octave I rushed a little bit
to get the changes in the release so that the code can recieve more testing
by the general public and we can get useful feedback on it.

If the changes do not get accepted we will probably just get notified of
issues we already know about since as long as three years and that we 
have already dealt with.

That's why having limited time and knowing the release was "imminent"
I temporarily disregarded matters such as indentation and coding style 
in private utility functions such as ode_defaults.m.

It seems the release was not so imminent after all, so I'd be happy to go 
through
the code one more time and clean it up before release if you'd like me to.

c.

P.S. I'm on IRC right now if you can join it can help speed up the review 
process,
we can then post the logs to the list or tracker for the record.