Re: More format() changes

[Rik]

On 10/16/2019 09:11 AM, John W. Eaton wrote:
> On 10/16/19 11:16 AM, Rik wrote:
>
>> Alternatively, we could use Octave's own unwind_protect
>> class and the protect_var method, but I'm not sure if that does what we
>> want in this instance which is to let the change happen, unless an error is
>> thrown, in which case the variable needs to be restored.
>
> This approach is probably the simplest thing and most like other code in
> Octave that needs to reset some internal program state on an error.  To
> prevent the changes from reverting if there is no error, you can discard
> the actions of an unwind_protect object.  I think the attached patch
> should do it.
>
> It would be a little easier if all format settings were stored in a
> single object instead of as individual file-scope or global variables.
> I'm headed in that direction anyway, as I think all this info should be
> stored as part of the interpreter object instead of globally, but that's
>  a project for another day.

I completely agree about encapsulating all of these bits of information in
a single object for ease of use, and also that I don't have time for that
just now.  Let me take a look at your diffs and understand it, and then
I'll check it in.

--Rik