Re: [cedet-semantic] Strange completion

[Eric M. L. <er...@si...>]

Hi,

  It appears you are not subscribed to the cedet-semantic mailing
  list.  Non-subscriber email is currently blocked due to all the spam
  that gets through source-forge's filters.  Your mail will go through
  faster if you subscribe.  Thanks.

  Answer below.

>>> =?ISO-8859-15?Q?Nathan_H=FCsken?= <ls...@lo...> seems to think that:
>Hello,
>
>I found somthing, which semantic completes very strangly:
>
>#include <stdio.h>
>
>class A
>{
>public:
>    void Function(int a,int b);
>};
>
>int func(a,b)
>{
>    A inst;
>    binst.
>}
>
>
>Please note, that I made a type error on purpose by writing "binst" 
>instead of "inst". In this case, semantic gives me the following 
>completions:
>
><The first completion is empty>
>inst
>min
  [ ... ]

  When semantic fails to identify the prefix, which is "binst" in your
case, it goes off an pretends that it isn't there, thus providing lots
of other completions you might not expect.

  It sounds like what you'd really want is for the user-interface to
take the unknown prefix, and do an ispell-like thing, and suggest
items that are spelled similarly and correctly.

>
>Is this fixable?
  [ ... ]

  We'd have to define what "fixed" is first.

Thanks for the good example.

Eric

-- 
          Eric Ludlam:                 za...@gn..., er...@si...
   Home: http://www.ludlam.net            Siege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: www.gnu.org