Re: From etags to ctags?

[Eric M. Ludlam]

On 06/04/2014 10:23 AM, Danil Orlov wrote:
> Emacs now support only etags format directly. But this format does not support 
> storage of extended information about symbols, like scoping and inheritance 
> data. But Exuberant Ctags can gather such data.
>
> And if such information will be presented in tags, it will be possible make 
> smarter completion for company-mode and auto-complete-mode. Without creating a 
> parser in elisp, like in js2-mode.
>
> How do you think, does it make sense to create direct support of ctags format, 
> like it done for etags now?

The command "semantic-load-enable-secondary-ectags-support" will enable 
use of Exuberent CTags to be used as a parser using Semantic for C/C++. 
 The extended parsing information can be used for C/C++ so that files 
not in buffers are parsed without loading them into Emacs.   That, in 
turn, can be used with the smart completion engine available via the 
CEDET/semantic system.

I tend not to use it because ctags parser isn't very good, and it fails 
to parse information about templates and other language features needed 
in more complex applications.

There is also "semantic-load-enable-primary-ectags-support" to add 
semantic support via ctags for 8 other modes.

The core parse for ctags isn't too complex so of someone wanted to use 
it to do something else, that probably wouldn't be too hard.  As a 
bonus, said code would then also work with other kinds of CEDET/semantic 
parsers.

Eric