Re: [bug-gtypist] further future of gtypist

[Felix Natter]

hello clutton,
hello Tim,

thanks for bringing this up.

I am currently contributing to Freeplane (Mind Mapping solution:
http://freeplane.org) and Debian (http://www.debian.org).
That is one reason why I haven't contributed much for a few years now.

The other reason is that we generate the command line parsing code
(master branch / 2.10) with an unreleased version of GNU gengetopt. This
does render it non-buildable in Debian (because even the command line code
must be generated there). I think we have to revert this change and use
normal getopt instead. But Tim should decide, as he wrote this part.

clutton <address@hidden> writes:
> I asked about adding a feature to gtypist a while ago. And now I
> understand why I haven't.
>
> I lurked in code several times in 2016 when I had some time. And now
> I've done it again having holidays. I've finished some other OSS
> contribution having free time but not gtypist.
>
> The code needs some kind of refactoring IMHO. The features were added
> and bugs were fixed and the logic now is rather complicated without
> reasons. I feel like I can't replace drill part without braking the
> whole thing.
>
> Two approaches can be taken, just codding and spending time with
> the code to get used to it. Or slowly rewriting/refactoring the whole
> thing.
>
> The code feels like there's too much logic injected here and there, oh,
> and there too, instead of some kind of design.

Yes. Until recently I haven't read clean code [1]. Now that I've read
it, the code looks like an insult ;-)

[1] http://www.goodreads.com/book/show/3735293-clean-code

> Don't think that I'm one of those guys who constantly want to rewrite
> everything, just telling you what I think about my very little time
> spending with sources.
>
> And I want you feedback, do you feel the same? Shall we just tinker it
> a little again to get what needed? Or shall there be some kind of
> refactoring? Both approaches have their good and bad sides.

How about converting the code to C++ classes while refactoring, I would
help with this!

When doing this, we can get away with the typeahead and word processor
features. I agree with clutton [2] that these are useless, since every
word processor / editor has its own heuristics (or none at all) here!

[2] https://lists.gnu.org/archive/html/bug-gtypist/2016-10/msg00003.html

Cheers and Best Regards,
-- 
Felix Natter