Tim:

Unfortunately, I haven't made much real progress on this since my initial
post of the gist.

The last pass I took at it I was making attempts to clean up the julia
initialization (which led to at least one stillborn PR).

My motivation had flagged a bit, as:
 1. I have limited extra cycles that are already spread thin across a
variety of personal and professional projects;
 2. I have been making an effort to improve my pure-julia workflow (still
very much a work in progress);
 3. for me, matlab is an at work, on windows thing, and I've recently
stopped using julia on windows; [It's a shame, but, unfortunately, the
windows version is simply not good enough for everyday use: significantly
slower startup times, partial or missing backtraces, unacceptably slow git
underneath the package manager, etc. Now my professional uses of julia are
done in an Ubuntu VM hosted on my windows machine. I'd volunteer to make it
better, but see #1.]
 4. I haven't been able to justify to myself that I needed it enough to
invest the time; and,
 5. no one else has needed it.

Of course, #5 can come off the list now. Besides, I'm more than happy to do
what I can to help out one of the most prolific and generous members of the
Julia community.

I've just created Mex.jl <https://github.com/twadleigh/Mex.jl>, a stub of a
repo on github and added you as a contributor. It will have to wait until I
get to work on Monday (alas, no remote access---against company policy),
but I will populate it with whatever I have and can probably donate at
least a couple of hours everyday until it's at the point where it meets
your needs.

I was initially thinking that this stuff could be included in the MATLAB.jl
package, but, looking into it more, there were a lot of choices made
therein that are specific to the matlab-from-julia model. Still very usable
for this effort, but probably better as a dependency.

We can continue this thread on github.

I'm looking forward to the collaboration!


On Sat, Feb 28, 2015 at 12:08 PM, Tim Holy <[email protected]> wrote:

> Tracy, any progress on this? I have a sudden need for this myself, and I'd
> be happy to help if there's stuff that needs finishing.
>
> --Tim
>
>
> On Friday, December 5, 2014 at 10:53:27 AM UTC-6, Tim Holy wrote:
>>
>> Sounds good. It would be for interop in both ways, then.
>>
>> --Tim
>>
>> On Friday, December 05, 2014 08:40:39 AM Tracy Wadleigh wrote:
>> > Rather than doing my own package, I think it makes more sense to try to
>> add
>> > this functionality to the existing MATLAB.jl package, where, for
>> instance,
>> > they already have mxArray marshaling.
>>
>>

Reply via email to