cpp header organizer

[Qin An]

Hi,

Dear friend, it is so busy now... 
So I am sorry that I can not help you...


best wishing,

-Qin An

	

======= 2005-01-07 04:01:04 =======

>
>Hi,
>
>Well.. I guess someone should answer, but I'm afraid I can't offer much 
>insight..
>
>On Sunday 02 January 2005 17.24, Felix Berger wrote:
>> Hi,
>>
>> I started to do some work on a cpp header includer, that is invoked with a
>> shortcut or from a menu and starts parsing the current file to find out
>> what classes and functions are used in the file but are not declared there.
>>
>> The ultimate goal would be to discriminate between cases where it suffices
>> to use forward declaration and cases where we need to include the header
>> file. Furthermore one could try to minimize the number of included files if
>> we knew what they themselves include transitively (Calculate transitive
>> closure).
>>
>> I imagine the user interface to ressemble Eclipse's "Organize Imports"
>> interface for java imports. It shows a dialog in case the import is
>> ambiguous.
>>
>> First question: Is someone else already working on this?
>
>Sounds like an interesting thing to have, and no, as far as I know no one is 
>working on it.
>
>
>> What I've done so far is create a class and provide an action in the edit
>> menu in the cppsupportpart. When the action is activated I fetch the
>> translation unit from the backgroundparser and start traversing it.
>>
>> To find out if a foward declaration suffices I have to go down to
>> statements to see if a pointer to a class is dereferenced somewhere. I
>> extended TreeParser to do this. My problem now is, that the statement lists
>> in the function ASTs seem to be empty. Is this a bug in my code or does the
>> parser not provide the information? Is it ok to ask the backgroundparser,
>> or should I create my own KDevDriver instance?
>
>Here's where it gets tricky. The only one who truly understands the parser 
>code is Roberto, the guy who wrote it, and he isn't around much these days. 
>Unless someone else knows I think the code is your only option to figure out 
>the answer..
>
>
>jd
>
>_______________________________________________
>KDevelop-devel mailing list
>KDevelop-devel at barney.cs.uni-potsdam.de
>http://barney.cs.uni-potsdam.de/mailman/listinfo/kdevelop-devel
>