This module provides a simple debugger for grammars. Just use it:
use Grammar::Debugger;
And any grammar in the lexical scope of the use satement will automatically have debugging enabled. The debugger will break execution when you first enter the grammar, and provide a prompt. Type "h" for a list of commands.
If you are debugging a grammar and want to set up breakpoints in code rather than entering them manually at the debug prompt, you can apply the breakpoint trait to any rule:
token name is breakpoint { \w+ [\h+ \w+]* }
If you want to conditionally break, you can also do something like:
token name will break { $^m eq 'Russia' } { \w+ [\h+ \w+]* }
Which will only break after the name rule has matched "Russia".
This gives similar output to Grammar::Debugger, but just runs through the whole grammar without stopping until it is successful or fails. Once again, after a use:
use Grammar::Tracer;
It will apply to any grammars in the lexical scope of the use statement.
On Windows you might find yourself annoyed by some gibberish like
←[1mTOP←[0m
| ←[1mfoo←[0m
| * ←[37;42mMATCH←[0m←[37m "x"←[0m
* ←[37;42mMATCH←[0m←[37m "x"←[0m
That's because the plain ol' cmd.exe doesn't understand ANSI escape codes which are to colour the output. In that case your options are:
-
Accept it and try to become Neo, he who "sees through"...
-
Wait for a perl6 equivalent of Win32::Console::ANSI
-
Wait for "colour-stripped" versions of Debugger and Tracer
-
Or, instead of ol' cmd.exe, use a console that understands the codes (ref #6)
-
Or you can use perl5 to get proper output, with
- either
perl6 MyGrammar.pm | perl -e "use Win32::Console::ANSI; print while (<>)"(have perl5's Win32::Console::ANSI handle it) - or
perl6 MyGrammar.pm | perl -e "print s/\e\[[0-9;]+m//gr while (<>)"(just strip 'em off)
- either
Please file them in GitHub issues.