On Wed, Feb 25, 2009 at 6:09 PM, Eric Northup <[email protected]> wrote: > MarkM beat me to the suggestion. We looked for and failed to find an > off-the-shelf parser generator that could be persuaded to do mixfix for > the Dyna language, and we ended up writing our own.
That seems to be a common curse. > I think mixfix could be regarded as syntactic sugar to be possibly added > later, guided by experience using the non-s-expression surface syntax. I tend to agree. And perhaps this means that I should proceed with a conventional grammar using conventional operator notation, with the plan that the expression parser may be shifted later. But I either need to build the parser by hand, which risks unrevealed discrepancies between the language specification and the parser, or I need to build a parser generator. If I'm building a generator anyway, the only requirement for mixfix support is to allow for a user-specified precedence resolver, so it seems a little silly not to do that much. shap _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
