[SVN] r4777 (OCaml, OCamlCodeCompletion)

[David Powers]

I like being able to fold functions out of the way - but I don't love it
more than life itself.  If it annoys others it is fine to remove.

I'll look at the tuple matching problem if I get a chance.  Same with class
parameters since those should probably be a variable of some sort.

constant.language.boolean would be the appropriate scope
(constant.booleanisn't allowed)


On 8/15/06, William D. Neumann <wneumann at cs.unm.edu> wrote:
>
> FIXED:
> Moved the class type part of the grammar above the class section to keep
> class types
> from being hijacked.
>
> Added ':' as a possible end marker for methods so things are handled
> properly in class
> type definitions. Note, however that this causes the method name in the
> type definition
> to also appear in the symbol list, but I'm not sure that's bad.
>
> Corrected the spelling of tuple.
>
> Changed the end pattern for meta.object.ocaml to allow for optional ;;
> after the 'end'.
>
> Changed tag on the input and really_input snippets to inp and rinp
> respectively to
> avoid a collision with the trigger for stdin.
>
> TODO (anyone , feel free to take these as I may be a while getting around
> to them):
>
> Need to fix the tuple definition, as it is catching constructs like:
> let f (a : int) b = ...
> and labeling the (a : int) as a tuple.  Also, the int is being considered
> a parameter
> of the function and not a type, as it should be.  This might be fixed by
> matching on,
> say (\()([^,],)+(\)), but match patterns can't have sub-patterns, so
> something
> slightly different is needed.
>
> Folding markers are messed up for classes and modules... this needs fixing
> if possible.
>
> Class parameters (e.g. the xx, yy, and zz in class blah xx yy zz = ...)
> aren't given
> any scope other than source.ocaml.  This may or may not be worth fixing.
>
> QUESTIONS:
> Folding markers for let function bindings are also pretty silly.  I'm
> wondering if
> these shouldn't just be removed, as they are next to useless, and do
> little more than
> add extraneous fold start markers to the window.  Does anyone else agree?
>
> Why are true and false labeled as "constant.language.pseudo-variable.ocaml
> "?
> Wouldn't constant.boolean.ocaml be better?
>
> William D. Neumann
>
> Changed:
> U   trunk/Bundles/OCaml.tmbundle/Syntaxes/OCaml.plist
> U   trunk/Bundles/OCamlCodeCompletion.tmbundle/Snippets/input.tmSnippet
> U
> trunk/Bundles/OCamlCodeCompletion.tmbundle/Snippets/really_input.tmSnippet
> U   trunk/Themes/Slush & Poppies.tmTheme
>
> _______________________________________________
> textmate-dev mailing list
> textmate-dev at lists.macromates.com
> http://lists.macromates.com/mailman/listinfo/textmate-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.macromates.com/textmate-dev/attachments/20060815/bd259feb/attachment.html>