Hi, On Tue, 1 Jul 2014, Simon Marlow wrote:
The s/r conflicts can also be a problem, depending on what you're trying to parse. It's generally a good idea to get rid of them if you can, but at the least you should understand why they exist (use happy --info) and document them in Parser.y.pp.
OK I've uploaded the code to wip/T9023 and the happy --info report to http://unsafePerform.IO/files/Parser.y.happy-info.gz Loads of shift/reduce conflicts come from state 214. Looking at them, I would have thought that the solution would be to parse pattern synonym definitions like we parse data constructors as types: by parsing it as a single pattern and then splitting it. A quick shortcut to this should be to just disable the parsing rule for infix pattern synonyms; so I tried that. As expected, this cuts down on the number of shift/reduce conflicts in that state considerably. The rest then seems to be coming from state 570: patsyn_context -> forall . (rule 157) patsyn_context -> forall . context '=>' (rule 158) Unfortunately, I don't have a quick workaround for that one yet. Thanks, Gergo