On Mon, Feb 27, 2012 at 8:34 AM, Lex Trotman elextr@gmail.com wrote:
Whese settings won't affect tag parsing AFAIK, as I said before I think tag parsers are mostly hard coded. The use of the dot is usually part of the language syntax and not considered part of an identifier. What language are you making tags for?
I mentioned that: Lua and Python Where are they hard-coded as only place where word characters are defined is 'editor.h:30' which should be overwritten with provided setting for "wordcharacter" property
I doubt it is that simple, as I said above most tag parsers don't consider dot to be an identifier character so you won't get tags with dots in them. The changes you have made to date should let Geany decide that the thing it wants to autocomplete has a dot in it, but that doesn't matter if the tag parser doesn't consider a dot to be a word character. It won't output tags with dots in them and so looking up tags with dots will never find anything.
As I said before the "right" way might be to figure out how to use the scope feature of tag files. But there are bugs in the scope lookup anyway and I don't think it works well outside of the C filetype for which it was invented.
It should be easy to set, IMHO, and I don't know if it's easy to be set in Geany.
I would not expect same calltip for:
numpy.core.load(file) Wrapper around cPickle.load which accepts either a file-like object ora filename json.load(file) Deserialize fp to a Python object.
which according Geany login should be set as:
load||(file) Wrapper around cPickle.load which accepts either a file-like object ora filename load||(file) Deserialize fp to a Python object.
or similar.
And it's not just that, but I'd like more to solve this problem instead opening discussion