emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: feature/icomplete-vertical


From: Ergus
Subject: Re: feature/icomplete-vertical
Date: Sat, 19 Sep 2020 14:57:27 +0200


On Sat, Sep 19, 2020 at 11:56:19AM +0000, Gregory Heytings via Emacs 
development discussions. wrote:

- when the current line after the prompt is too long the prompt disappears

Yes, set visual-line-mode -1 for this. Is the only solution so far and what I will do in the setup probably.


No, it is not the only solution, my code handles that case without using visual-line-mode.

If there is visual-lines -1 + truncate-lines that takes care of this in
a consistent way using the internal infrastructure. Why should I use
something else?

By the way, I just checked it (with your code) and apparently enabling visual-line-mode does not solve that problem.

Did you added truncate-lines locally?

As I said before, the patch was preliminar and needs some fixes I do
slowly as I am not a lisper. I just diff my local changes to send them
to you


Did you applied the patch cleanly?


As far as I can see yes. I already tried your code twice, did you try mine? I attach the improved version which handles the minibuffer-only case.


The changes indeed do only one thing at the time and it separated the
vertical from the horizontal completely. Which is IMO simpler that just
add if and else conditions here and there or hard code the mapcar
ignoring the user separator.

Anyway, IMO you not are using the best approach on this project, you are trying to do too many things at once.

You can do it if you want. No problem with that. Actually I tried to
convince the creator of the external package to do it and asked in the
mailing list some months ago if anyone wanted to do it and nobody
replied.

Completion mechanisms are a very complex thing, and I believe the best way is to do small incremental improvements with minimal changes, and to test them thoroughly before implementing the next improvement.

This is not a big change at all. This is just some cosmetics to an
output. I don't interact with he completion system and the code is not
complex at all.

Actually it is just 2 functions. The rest is the same just in a
different place. But actually your code and mine do exactly the same
just using different conditions.

If you see my first commits it was almost as simple as yours but then
reports started coming and I needed to fix them unless some of them
(like the disappearing prompt) were old bug already reported.

IMO this is an incremental change, but I don't try to compete for a
smaller patch but simpler and free of known issues. Like the ones you
are reporting.

Best
Ergus


reply via email to

[Prev in Thread] Current Thread [Next in Thread]