[PATCH] revset: don't error out if tokens parse as existing symbols

[Jordi Gutiérrez Hermoso]

On Sat, 2015-05-02 at 19:51 +0900, Yuya Nishihara wrote:
> On Sat, 02 May 2015 00:50:58 -0400, Jordi Gutiérrez Hermoso wrote:
> > @@ -291,6 +291,11 @@ quoting needed
> >    $ log '"date"'
> >    abort: unknown revision 'date'!
> >    [255]
> > +  $ hg book date -r 4
> > +  $ log 'date'
> > +  4
> 
> It doesn't test the revset function because old-style revision can be caught
> by scmutil.revrange().

I beg your pardon?

Here the revset is "date", not "4", unless I am much mistaken. This
test certainly failed before this patch, so I am almost sure it's
going through the symbolset function.

As to your other concern whether symbolset can be removed, maybe, but
it's still being called by the test above