Hi,
You have a good idea. This wasn't reasonable in the original way
the stuff in CEDET was packaged and distributed. It will probably
make more sense now. I know for C/C++, there are 4 ede files, 4
semantic files, and 2 template files, plus some misc settings embedded
elsewhere. I will pose this to the Emacs maintainers.
Thanks
Eric
>>> Ming-Wei Chang <mchang21@...> seems to think that:
>
>Hi Eric,
>
> First of all, thanks for the wonderful CEDET!. Here are my opinions.
> I only do small scale projects with C and C++, so my opinion can
> be biased.
>
> 1) I do not use semanticdb-create-system-database at all.
>
> 2) I am not so familiar with the key binding in CEDET, so I will skip
> this point. However, I think it would be good if CEDET provide
> default key binding for intellisense. (It can be different for
> each language.)
>
> I do have one suggestion. Is it possible to re-factor the directory
> structure of CEDET so that all language-specific files for a specific
> language are in a single directory? That is, we will have a directory
> for C support and another directory for java support. The code for
> project management in C and the parsers for C language are in the
> directory for C language. I think this is important for several
> reasons.
>
> a) If someone want to add support for some language, this would make
> him/her life much easier. It is also make it possible to control the
> version for each language support.
>
> b) For a beginner like me, it is very hard to understand the
> relationship of wisent-c.wy, semantic-c-by.el and semantic-c.el.
> If we put them in a single directory, I believe the relationships
> between them will be more clear. Moreover, it might be possible to
> remove some redundant codes (if any) by doing this.
>
> c) For many languages, some other tools which are not written in
> emacs lisp might be needed. For example, we might want to use rope
> (written in python) as the backend for the python language
> support. If we can re-factor the data structure, I believe that it
> would be easier for the developers and users to manage the additional
> tools.
>
> d) In this way, I feel the interface part and the backend part are
> more independent to each other.
>
> Any comment?
>
>Best,
>
>Ming-Wei
>
>
>"Eric M. Ludlam" <eric@...> writes:
>
>> Hi all,
>>
>> I have some more poll questions for everyone. These are related to
>> integrating CEDET into Emacs.
>>
>> 1) Is anyone still using `semanticdb-create-system-database' for anything?
>>
>> This feature doesn't add much anymore, and I'm thinking of removing
>> it. The last bit for me was adding the work timer that will go out
>> and build databases in long-idle time. This should help remove the
>> need for this pre-setup step. (The secondary work timer is in the CVS
>> version of CEDET.)
>>
>>
>> 2) Which Semantic related keybindings do you use, or setup yourself?
>>
>> Senator, due to its use of advice and the way it overrides core
>> Emacs functions, will not be integrated into Emacs directly. (ie, the
>> maintainers don't like it.) As a side-effect I'm looking to
>> understand how it is used beyond what I do myself.
>>
>> For example, Senator "fixes" beginning/end of defun and related core
>> Emacs keybindings. It adds next/previous navigation, and has bindings
>> for jumping to tags, and doing different kinds of searches.
>>
>> If there is some function you wish was in Semantic (and within its
>> capabilities, of course) that would be good to know too. ie, is there
>> a combination of tools you use that would be better served as a single
>> slightly smarter tool?
>>
>> Since Senator has acted as a catch-all for anything interesting,
>> it's time to prune the tree to the useful items.
>>
>> Thanks for any and all input.
>> Eric
>
--
Eric Ludlam: eric@...
Siege: http://www.siege-engine.com Emacs: http://cedet.sourceforge.net
|