Re: Empty classdef files and shift/reduce conflicts in the parser

[Michael Goffioul]

On Thu, Jan 30, 2014 at 9:57 PM, Michael Goffioul <address@hidden> wrote:

Hi,

I'm trying to make the parser accept empty classdef body or empty methods/properties/... blocks. I've tried a first approach as in patch classdef-empty-defs, which works fine. But I don't find it very clean. So I tried another approach as in patch try2.patch (just a proof-of-concept). But I then got an increase in shift/reduce conflicts (going from 14 to 23). I'm not really a bison expert, so I'm wondering whether it's possible to avoid those conflicts, or whether the first approach is ok.

Just to close the issue/thread, I've submitted this change:

http://hg.savannah.gnu.org/hgweb/octave/rev/9154dc252f47

Michael.