Re: [Matlab-emacs-discuss] git repo at sourceforge.

[Odd A. <odd...@gm...>]

Hi, Uwe,

Honestly, I wrote the patch as a workaround since the new versions of
Matlab no longer emit hotlinks in terminal mode.
It works by remapping the emacs debugger keyboard shortcuts to new
functions that insert these hotlinks themselves where Matlab used to do
so.  They are then picked up by emacs-matlab as if they had been emitted by
Matlab.

I wrote this primarily to be able to keep debugging my Matlab scripts in
emacs, since this no longer worked when I installed Matlab2015b.  My
knowledge of matlab-emacs is however quite superficial, so I cannot confirm
whether there are no unintended side effects.  I posted it on the mailing
list in the hope it could be useful, as several people have reported the
same problem.



On Sat, Mar 12, 2016 at 6:54 PM, Uwe Brauer <ou...@ma...> wrote:

> >>> "Eric" == Eric Ludlam <Eri...@ma...> writes:
>
>
>    > ---
>
>    > Hi Uwe,
>
>    > In the distant past, Emacs Matlab debugging worked by parsing the
>    > raw output text. It was a bit finicky but worked ok. That broke
>    > when the text output suddenly started emitting html style hotlinks
>    > as part of the java desktop. I re-wrote the parsing engine to focus
>    > on the hotlinks because they were very regular and reliable. It
>    > turns out, that the emitting of hotlinks was a bug that they ever
>    > came out on a generic terminal, so that was fixed, so now there is
>    > nothing for the parsing engine to pay attention to.
>
>    > I like the idea of having new db commands unique to Emacs so that
>    > we control the output. If we turn the hotlink feature off
>    > (currently being turned on) and switch to these new commands,
>    > things should work in older MATLABs as well.
>
> So to be sure: that patch we are talking about does these 2 things?
> Or must the hotlink feature turned off manually? If so then this should
> be mentioned in the release notes and/or the doc string.
>
>    > It will take a bit of testing to make sure it works reliably, and I
>    > think the # folks on this mailing list has gotten pretty small. I
>    > think it's worth going for it and expanding to all the relevant
>    > debug commands, and posting a request for testing when the patch is
>    > in.
>
>
> Ok I could apply the patch and change the matlab version number to make
> clear that there is something *new*, but I want to be sure not to break
> backward compatibility, also out of egoistic motives. I am still stuck
> on my laptop to Matlab 2010 and don't want to break things.
>
>
>    > It doesn't help with error output, and jump-to-last-error though.
>    > That would still need to be solved.
> Ok
>
> Uwe
>