Fixed an incompatibility with the old Parser.Pike.
Why isn't pike_tokenizer.h 100% compatible with Parser.Pike's tokenizer?
Actually, most bugs were not mine but yours. :-)
The only serious bugs that was fixed was #"", the handling of zero-length strings as input and an UNSET_ONERROR() that was missing.
I only feel responsible for the last one. :-)
I'll happily take the blame for not knowing how to properly push empty strings and arrays (which by the way is a bit inconsistent if you compare the two).
When it comes to missing Pike constructs it would be great to have a better acid test. It would be great to generated syntactically correct code from the BNF or somthing similar.
pike-devel@lists.lysator.liu.se
-
Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum -
Martin Nilsson (DivX Inc.) ＠ Pike (-) developers forum
-
Per Hedbor () ＠ Pike (-) developers forum