Re: HEAD b0rked

Sebastian Bergmann wrote:

Sebastian Bergmann wrote:

conflicts: 21 shift/reduce, 15 reduce/reduce

I get this same message with bison 1.75 only that this version does not abort. Maybe some bison parameter(s) changed?

yes, with older bison versions you can declare how many shift/reduce conflicts you expect to have in your grammer and it terminates if the actual number of shift/reduce conflicts does not match bison 1.875 also requires that there are *no* reduce/reduce conflicts when testing the number of shift/reduce conflicts with %expect this is generally a good thing as you don't want to have reduce/reduce conflicts in your grammer, but in the special case of parsedate.y i did not care about the reduce/reduce conflicts as the results are still ok. the datetime parsing grammer can live with reduce/reduce ambiguities as the same result can be reached on different pathes here ... looks like i have to re-think the patch again :( -- Hartmut Holzgraefe <hartmut@php.net>