>>> Hannu Koivisto <azure@...> seems to think that:
>"Eric M. Ludlam" <eric@...> writes:
>> It thinks that you are calling "member", and that member's first
>> argument is of type Xy. Az Xy is not defined anywhere, it can't find
>> any data types to match, thus no completions.
>> If you use the senator completion, it will use a simplified completer
>> as a backup, and likely hit your case.
>> This is a class of problem I've had my eye on fixing for a while. The
>Thanks for the explanation. Though I thought senator's completer
>_is_ the simplified completer (which has no backup) and semantic's
>completer (semantic-ia-complete-symbol at least) uses it as a
Oi, I think you are right. There are a lot of relation ships between
some of these functions and it is easy to get confused.
>I have actually disabled semantic-ia-complete-symbol's use of
>senator completion. I tend to think consistency and precision is
>the reason for semantic-ia-complete-symbol and I'd rather it gives
>up than does guesswork. I prefer to use dynamic abbrevs for that.
>Sometimes, when making a bug report of incorrect completion
>behaviour, I didn't realize that it was senator completer in action
>and that made my attempts to come up with a test case a nightmare
>-- I thought it was semantic's completion behaving in a hard to
>What would think of a patch to make use of senator's completer in
>semantic's completion interface a customizable option?
That seems like a fine idea. Let me tell you my longer-term plan
because, well, I like to chat. :)
My short term goal is to get a stable CEDET 1.0 that can stand the
test of time. That will set things up so I can do some refactoring.
My next project is to do some serious work on my catapults, which are
long neglected, and not really related to CEDET which has been taking
all my free time.
After that, I'd like to take a high-level look at CEDET and all the
options. Right now it is a mix of different tools and packages, and
things like completion have evolved as ideas came up, and was not
really designed. I'd like to design a CEDET set of interfaces and
commands with some sort of rational consistency. If possible, offload
to some pre-existing tools.
Within that space, I'll need a ton of help figuring out what that
looks like, how it lays out on the keyboard and a users mental space.
Hopefully you will enjoy participating in that. I hope that is a
project more folks can and would participate in writing code to help
improve the interactive pieces.
Eric Ludlam: eric@...
Siege: http://www.siege-engine.com Emacs: http://cedet.sourceforge.net