Re: [CEDET-devel] Misc. questions about cedet setup.

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

Hello,

Anders Rønningen wrote:
> Hey.
> 
> I've just installed cedet following the docs on the webpage and the 
> "Gentle introduction". All the functionality seems to work, but I have 
> some questions on further customizing.
> 
> 1: I have added this to my cedet-setup:
> 
> (local-set-key "." 'semantic-complete-self-insert)
> (local-set-key ">" 'semantic-complete-self-insert)
> 
> Is it possible to make these completions appear in a popup instead of in 
> a dedicated buffer? I use "C-c ?" to get 
> semantic-ia-complete-symbol-menu which gives me what I want, but it 
> would be nice to have the popup.

If you customize semantic-complete-inline-analyzer-displayor-class, you 
can choose different ways to display completions.

> 2: How can I control what semantic indexes? I have generated a list of 
> all the c/h files in my project, and I would like to index only the 
> files there. Alternatively, I would like to tell semantic about a top 
> directory, and have it find all the c-files recursively from there. What 
> is it that tells semantic which c-files to scan? To use the symrefs and 
> tags functionality I have to know/control which files it have indexed. I 
> have generated a cscope.out that I tried to use, but emacs hangs 
> everytime I tell semantic to use it (this is on windows, so I have 
> generated the file using 'cscope -bR' in a terminal, since the 
> cscope-indexer isn't available).

You can use cedet-cscope-version-check to figure out basic cscope 
problems.  It might be you need to customize cedet-cscope-command.

For C, if you use ede-cpp-root project type to mark the top of your 
project, then you shouldn't need to ask it to index anything.  See the 
commentary section of ede-cpp-root.el for config details, or 
common/cedet.texi section on C++ coding for basic setup.

The idea is that Semantic will dynamically parse only what is needed 
from your includes to do its job.  If you want to use symref type tools, 
then it uses grep, cscope or other external tool to do the work where 
only a name is needed, and not real parsing.

There are some tools that do need semantic to have parsed everything. 
You could use semanticdb.sh and pass it a list of files to parse when 
your regular emacs is not running.  (I haven't tried this in quite some 
time.)  Otherwise, visit each file you care about.  Once you do that, it 
will be indexed and you don't have to visit it again unless you delete 
your ~/.semanticdb cache directory.

Good Luck
Eric