On Sun, Jul 03, 2005 at 11:08:56AM -0400, Will Coleda wrote: > When I was compiling the rules, they weren't going where I thought > they were going. This was because I stole code from compilers/pge/ > demo.pir improperly; there is a gname variable in there that's a PMC.
No, it's a string: [EMAIL PROTECTED] pge]$ grep gname demo.pir .local string gname p6rule_compile($S2, gname, $S1) print gname gname = $P9 [EMAIL PROTECTED] pge]$ > (My clue was your sample, which was calling the p6rule_compile with a > literal string to describe the parent grammar, not a PMC as in > demo.pir). > > Of course, now I don't know why demo.pir works, but it does. =-) Well, it works because (AFAICT) gname is in fact a string. :-) Maybe there's something else bizarre going on here? > >As a side note, it'd be nice if the code that demo.pir uses to read > >in a perl6 rules file was part of PGE's interface. I considered adding this at one point and then decided it might be a little too much creeping featurism. But I can see how it'd be very useful, so I'll see about re-incorporating it. Pm