making O/Ecompletion even better

Hi romain

Often when I edit code, the completion works well but I end up been forced to delete part of the text manually.
Here is the scenario


        classDefinedSelector foo...
        ^
        cursor

        Then I type meta
       
        the system proposes me metaclassDefinedSelector
        I type enter

        and I get

        metaclassDefinedSelector classDefinedSelector foo
                                            ^

        so I have to remove by hand classDefinedSelector.

So it would be good if we could eat the current text when doing the substitution.

Am I the only one getting such behavior?

Stef
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

CONTENTS DELETED

Hi Stef,

No you are not :).

I also mentioned once two more issues:
- Enter should not trigger completion because it interferes with code  
formatting. I would only keep tab as a means to trigger completion and  
to select something
- right now, when pressing the down key because you saw a good  
completion item in the second position there is a visible delay  
because the ECompletion is triggered. ECompletion should either be  
integrated in the OCompletion directly, or should be triggered only  
when we hit the bottom of the three items. In the current form, having  
the completion item in the second position is almost as bad as not  
having it at all.

Unfortunately, I do not have time to look into these issues, but I can  
offer testing :)

Cheers,
Doru


On 25 Apr 2010, at 15:45, Stéphane Ducasse wrote:

--
www.tudorgirba.com

"From an abstract enough point of view, any two things are similar."




_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On Apr 25, 2010, at 3:55 PM, Tudor Girba wrote:

> Hi Stef,
>
> No you are not :).
>
> I also mentioned once two more issues:
> - Enter should not trigger completion because it interferes with code formatting. I would only keep tab as a means to trigger completion and to select something

        yes you are probably right. I was in favor of this inverse... but I learned :)

> - right now, when pressing the down key because you saw a good completion item in the second position there is a visible delay because the ECompletion is triggered. ECompletion should either be integrated in the OCompletion directly, or should be triggered only when we hit the bottom of the three items. In the current form, having the completion item in the second position is almost as bad as not having it at all.
>
> Unfortunately, I do not have time to look into these issues, but I can offer testing :)

romain :)
please....... with the eye of the cat in shrek


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Stéphane Ducasse wrote

So it would be good if we could eat the current text when doing the substitution.

I would add one twist to that: only if the remaining part of the word matches the tail end of the selected completion.

Equally as disruptive as what you correctly discribed is:
        bar foo...
        ^
        cursor

        Then I type meta
       
        the system proposes metaclassDefinedSelector
        I type enter

        and I get

        metaclassDefinedSelector foo

Where is my bar!!!

Sean

I just dug up MacVim to double-check...

I also always found not adding a space after completion to flow better than having to delete the space at the end of the word, even if it may require a few extra key strokes over the course of a line; because I don't have to do a mental double hop to 'correct the correction' and my finger is right on the spacebar anyway.

Otherwise, I agree it works awesomely well :)

Sean

Thanks all for the feedback!

Finding time is always hard, but I'll have a look at all your suggestions.

Cheers,
        Romain

On Apr 25, 2010, at 2:07 PM, Sean P. DeNigris wrote:

--
Romain Robbes
http://romain.robb.es
[hidden email]


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project