Fixing completion

[Stefan Schimanski]

Attached is a patch to replace the manual interception of keys with proper LFUNs. In fact one half of it was already implemented. Now you can also set the key to hide the popup/inline completion.

Opinions?

Stefan

With this the remaining question is how to bind the LFUNs by default. Locally I have tried Tab and Escape and it works as before.

lfun-completion.patch
Description: Binary data

Am 07.09.2008 um 13:19 schrieb Uwe Stöhr:

>> Sorry but this is the rule: When you add a new feature build it so that it >> doesn't introduce a regression except this was a discussed decision.

>




> of course we shouldn't remove it. but the point is to do it in a way




So lets wait what Stefan and Vincent say.


> TAB very fast, so this is not a big problem.
>

> time before two hours :D




> to introduce 'fixes' which break another things.



regards Uwe

> it is not fair to be taught about the rules in the thread which start with > defective commit that was not discussed at all, in fact even announced. Why is this not fair? I explained why I committed the patch and why this is conform to the rules. > that both tab completion and tab navigation works. i think its possible, > but not through completer_->popupVisible() && ev->key() == Qt::Key_Tab. I fully agree, but until this is fixed, the TAB behaviour of LyX 1.5 needs to be restored and the completion feature be fixed. > I don't agree with you here: Now only completion doesn't work when pressing > yes and those two very fast tabs were exactly the way how completion saved my And now I have to write a report containing lots of tables, and navigating with the TAB saves me a lot of time. What is better? ;-) > as i said, its not either or; we will make both working. its just stupid No, the completion feature introduces 2 regressions, not my commit introduced a bug, it just uncovered the completion bug.