having built something similar for antlr in the old days, I can also point out that it’s extremely challenging deciding whether to loop/recurse when faced with a decision. It’s hard to determine what the right choices to generate reasonably sized input. I ended up with extremely short or extremely long input sometimes.
Ter On Sep 5, 2014, at 12:06 PM, mathew palsberg <matpalsb...@gmail.com> wrote: > Thanks Lukas, and agree generating strings considering lookahead and ! should > be very hard! > I dont know how one could generate valid string according to NOT and > Lookaheads. > cheers. > > > On Fri, Sep 5, 2014 at 7:23 AM, Lukas Renggli <reng...@gmail.com> wrote: > The PetitParser UI [1] has the capability to display random examples of a > selected production. Unfortunately the UI (and the example generating code) > is only available in Smalltalk and hasn't been ported to other host > languages. However, you can look at the code here [2] and play with the tool > by downloading the one-click distribution [3]. > > The example generating code is really short and simple (less than 30 lines, > the rest is markup from the export). There are a few (hardcoded) heuristics > to make it produce a short and readable example. > > Note that the code ignores logical predicates (! and & expressions). I am not > sure, if it is even possible to make them work in a generic case? Either way, > I found that ignoring the predicates isn't as bad as it sounds. Predicates > are rarely used in practice and do have a big effect on the resulting output > (as long as you are aware of it). > > To summarise, I find the examples quite useful while developing new grammars. > It allows one to quickly spot errors or edge-cases that might get missed > otherwise. > > Cheers, > Lukas > > [1] http://scg.unibe.ch/archive/papers/Reng10cDynamicGrammars.pdf (Section > 3.5) > [2] > http://www.lukas-renggli.ch/dropbox/petitparser/libraries/PetitParser-Example.st > [3] http://source.lukas-renggli.ch/built/oneclick/PetitParser-OneClick.zip > > > On 4 September 2014 19:54, mathew palsberg <matpalsb...@gmail.com> wrote: > Hello, > I recently built a small language with PEG. I now want to test my language > with random generated(not necessarily meaningful) programs. Does anybody know > a tool that given a PEG grammar, it generates random sequences that satisfy > the given grammar? > > Thanks; > > > _______________________________________________ > PEG mailing list > PEG@lists.csail.mit.edu > https://lists.csail.mit.edu/mailman/listinfo/peg > > > > > -- > Lukas Renggli > www.lukas-renggli.ch > > _______________________________________________ > PEG mailing list > PEG@lists.csail.mit.edu > https://lists.csail.mit.edu/mailman/listinfo/peg _______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
