Re: [cedet-semantic] Semantic issue and questions
Brought to you by:
zappo
From: Eric M. L. <er...@si...> - 2008-02-06 17:56:29
|
>>> Michael Reiher <re...@gm...> seems to think that: >Hi > >After there have been a number of improvements recently I decided to Cedet >another spin. And after playing around a bit it seems indeed to work a lot >better than last time I tried. Graet work!! Thanks! The answer to your questions below is basically that: The functions in 'semantic-ia' are all lowest order interfaces to various semantic features. Each function is pretty short, and I maintain them as teaching aids for application development and testing. I should probably update the documentation about these items. More detailed answers below: >There is an issue however I came across. When calling >semantic-ia-complete-symbol it nicely shows the list of completions in the >other buffer. But when I select an entry, it adds the _full_ function header >in the source buffer. So when it offers "void foo ( int bar )" it inserts >exactly that, so when I'm at "xyz.f<semantic-ia-complete-symbol>" I get >"xyz.fvoid foo ( int bar )" ... Using semantic-ia-complete-symbol-menu, it >seems to work fine, it adds just "oo(". I'm using XEmacs 21.4.17. The basic problem here is that the completion buffer is filled with useful info abut the possible names (like return type, etc). If you click on a completion in that buffer, it inserts the entirety of that text. I'm not that familiar with that part of Emacs to know what to do about it. If anyone out there knows how to fix stuff from the completion buffer, let me know. >And I have two questions: >Is it somehow possible to cycle through the list of completions, using TAB or >whatever? Because the popup menu (semantic-ia-complete-symbol-menu) is not >really usable in XEmacs as it doesn't get keyboard focus and the list in the >buffer (semantic-ia-complete-symbol) is IMHO also a bit inefficent to use. >Currently I use dabbrev-expand on TAB, which works not too bad actually. It >would be even better if I could first go through symbols offered by semantic, >and only fall back to general completion by dabbrev-expand if there are none. You might want to use `semantic-complete-analyze-inline' for your completion needs. It is a big complicated fancy tooltip driven UI for this sort of thing. There is also speedbar's analyzer mode, ECB's analyzer completions panel, and I think out on Emacs Wiki is "company-mode" that comes with a pretty cool semantic smart completions thingy. (The web video/demo was pretty cool. I haven't gotten around to trying it myself though.) >If I use semantic-ia-fast-jump, is there a command to let me jump back to >where I came from? I tried pop-tag-mark, but obviously it uses a different >mechanism. [ ... ] That function calls `push-mark' which ought to enable the mark stack pop-global-mark (C-x C-SPC) which is different from the tag stack. Alternately, give `semantic-mru-bookmark-mode' a try. Hmmm. I need to get that into senator's menu. Eric -- Eric Ludlam: er...@si... Siege: www.siege-engine.com Emacs: http://cedet.sourceforge.net |