Re: [anjuta-devel] RFC: Symbol-db future



Hi!

If you think this filtering is fast enough we can do that. But the file
filter list will have 50-300 entries.


yes but if you use the 'packages' grouping you would not require to pass 
all those entries to a single query.

Yes, but it wouldn't work as good as single file filtering. And in
addition what would we do with files that aren't part of any package
because they are system headers or just part of a library that's not
using pkg-config?

As said before, I don't see much benefit for the package filtering in
general.

well, nope.
But I'm not sure that a recursive query that searches into parents scope 
can be written. That's why the cxxparser was written, to give an help in 
these kind of situations.

@IANJUTA_SYMBOL_QUERY_SEARCH_PARENT_SCOPE: Query to get the parent scope 
of a symbol.

This query isn't really easy to calculate, and thinking to put the 
procedures to search the parents recursively is hard.

cxxparser is great to detect situation like object->method() but it
doesn't help us with scoping in the global database. But we would really
like to have autocompletion for this case:

Class MyButton : Gtk::Button
{
        
}

MyButton::do_something()
{
        set_label ("Something"); // this is a method of Gtk::Button
        // but Gtk::Label might also have a similar method that we don't
        // want
}

Of course we can do something like that in the language-support plugin
(sync pseudo code):

for (scope = current_scope; scope != NULL; scope =
scope_get_parent_scope (scope)
{
        symbols = g_list_concat (symbols, find_symbols_in_scope (scope));
}

But I though that there might be a faster way using database queries. If
there isn't we have to do it this way.

Regards,
Johannes

Attachment: signature.asc
Description: This is a digitally signed message part



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]