CEDET

Maindoor writes:
> (require 'semantic/bovine/c)
> (require 'semantic/bovine/gcc)
> (require 'semantic/bovine/clang)
> (require 'semantic/ia)
> (require 'semantic/decorate/include)
> (require 'semantic/analyze/debug)
> (require 'semantic/lex-spp)
> (require 'eassist)

Please get rid of those requires. I know Alex has them in his config,
but they should not be necessary. If something does not work after
removing those requires, please report it as a bug.

> I have built cscope and gtags database where the "Makefile" is present
>
> 1) I am searching for the definition of "isValidTxRate".
> There are 4 definitions of it in the project in directories ar5210,
> ar5211, ar5212, ar5216.
> But when I do a semantic-ia-fast-jump, it jumps to only 1 of the
> definitions. I want to edit
> all of the definitions. In my case the board id will dictate which one
> to use, but that's run time.
> Is there any way to see all the definitions other than doing a grep ?

You can use semantic-complete-jump (bound to C-c , J). It will use the
GNU global database and you should be able to cycle through the
different definitions by repeatedly pressing TAB. Unfortunately, it's
not working very reliably... I often wanted to fix it, but then got lost
in all those collector and displayor classes...

> 2) I wanted to semantic-symref-symbol nt_node which is part struct
> ieee80211_node_table, but with the above
> configuration (ede-cpp-root-project), I get [No Match]. What am I doing wrong ? 

Yes, that one's simply not working. It's on my ever-growing TODO
list. Calling `semantic-symref' (bound to C-c , G) should work, but it
uses the symbol under point.

-David

David Engster writes:
>> 2) I wanted to semantic-symref-symbol nt_node which is part struct
>> ieee80211_node_table, but with the above
>> configuration (ede-cpp-root-project), I get [No Match]. What am I doing wrong ? 
>
> Yes, that one's simply not working. It's on my ever-growing TODO
> list.

It should basically work now with latest CEDET from bzr. I'm not very
happy with how the completion system works, but at least you should be
able to search for tags in the current project.

-David

>It should basically work now with latest CEDET from bzr. I'm not very
>happy with how the completion system works, but at least you should be
>able to search for tags in the current project.

I tried with the latest from bzr. 
There is definite improvement over my last  experience. But....
This is my config file now:

(setq cedet-root-path (file-name-as-directory "~/cedet/"))
(load-file (concat cedet-root-path "cedet-devel-load.el"))
(add-to-list 'load-path (concat cedet-root-path "contrib"))

;; select which submodes we want to activate
(add-to-list 'semantic-default-submodes 'global-semantic-mru-bookmark-mode)
(add-to-list 'semantic-default-submodes 'global-semanticdb-minor-mode)
(add-to-list 'semantic-default-submodes 'global-semantic-idle-scheduler-mode)
(add-to-list 'semantic-default-submodes 'global-semantic-stickyfunc-mode)
(add-to-list 'semantic-default-submodes 'global-cedet-m3-minor-mode)
(add-to-list 'semantic-default-submodes 'global-semantic-highlight-func-mode)
(add-to-list 'semantic-default-submodes 'global-semanticdb-minor-mode)

;; Activate semantic
(semantic-mode 1)

;; load contrib library
(require 'eassist)

(semanticdb-enable-gnu-global-databases 'c-mode t)
(semanticdb-enable-gnu-global-databases 'c++-mode t)

(when (cedet-ectag-version-check t)
  (semantic-load-enable-primary-ectags-support))

;; SRecode
(global-srecode-minor-mode 1)

;; EDE
(global-ede-mode 1)
(ede-enable-generic-projects)
(ede-cpp-root-project "madwifi"
		      :file "~/Downloads/madwifi/Makefile"
		      :include-path '("/ath"
				      "/ath_hal"
				      "/net80211")
;;		      )

1) With the last open brace commented, I get symbol references 
(semantic-symref-symbol of nt_node only from the current directory (ie. net80211 
when searched from ieee80211_node.c). there are some references in ath directory that are not shown.

2) When the last brace is corrected. ie matched properly then I get "No references 
found". when I do the semantic-symref-symbol of nt_node from a ".h" file I get 
"Cannot analyze buffers not supported by semantic".


What am I doing wrong here ?

-Maindoor


----- Original Message -----
From: David Engster <deng@...>
To: Maindoor <sanjeevfiles@...>
Cc: "cedet-devel@..." <cedet-devel@...>
Sent: Monday, November 5, 2012 8:12 AM
Subject: Re: [CEDET-devel] Setting up to browse madwifi

David Engster writes:
>> 2) I wanted to semantic-symref-symbol nt_node which is part struct
>> ieee80211_node_table, but with the above
>> configuration (ede-cpp-root-project), I get [No Match]. What am I doing wrong ? 
>
> Yes, that one's simply not working. It's on my ever-growing TODO
> list.

It should basically work now with latest CEDET from bzr. I'm not very
happy with how the completion system works, but at least you should be
able to search for tags in the current project.

-David

Maindoor writes:
> 1) With the last open brace commented, I get symbol references 
> (semantic-symref-symbol of nt_node only from the current directory (ie. net80211 
> when searched from ieee80211_node.c). there are some references in ath directory that are not shown.
>
> 2) When the last brace is corrected. ie matched properly then I get "No references 
> found". when I do the semantic-symref-symbol of nt_node from a ".h" file I get 
> "Cannot analyze buffers not supported by semantic".

If you don't have defined a project (which is the case in 1)),
semantic-symref will not find your GNU global index and simply use
'grep' to find symbols. But since it has no project, it will only search
the current directory.

If you have a proper project, it will use GNU global, but global doesn't
know C++ very well. It does not index class members, for instance, which
is why it says it cannot find references. Just do "global -c" in the
root of your project to see the symbols it has in its index. The error
you're getting in the ".h" file is probably a bug.

It might be that another indexing tool like 'idutils' could work better
because it actually is *less* smart than global, but I'm not sure if
it's worth fiddling with this stuff. I usually just use M-x rgrep in
cases like these.

-David