Details - The "fine print" for Marpa::R3
This section contains the "fine print". It describes various details, and issues some caveats that apply to Marpa::R3 in a general way. Everyone writing an Marpa::R3 app should skim this document.
Marpa::R3 expects all strings passed to it to be valid UTF-8. If a string is passed to Marpa::R3 which does not look like valid UTF-8, a fatal error results.
All ASCII-7 strings are valid UTF-8 strings, so apps which restrict themselves to the ASCII 7 characters will have no problem.
Pedantically, Marpa::R3 is "outside" of Perl, so that strings must be encoded before being passed to Marpa::R3, and decoded after being returned from Marpa::R3. By default and in practice, Perl uses only valid UTF-8 characters so that it will usually not be necessary to encode/decode.
Unfortunately, Perl's default is "utf8", which is different from the official Unicode UTF-8. Perl's "utf8" is a superset of "UTF-8" -- it allows additional characters not valid according to the Unicode standard. Most Perl applications do not use UTF-8-invalid characters.
Perl applications that do not use UTF-8-invalid characters do not need to decode/encode when communicating with Marpa::R3. However, if one of the Perl "utf8" characters that is not a valid UTF-8 character is passed to Marpa::R3, a fatal error will result. To fix the problem, an application will have to encode the Perl "utf8" string into "UTF-8".
Marpa::R3 exists to allow its input to alter execution in flexible and powerful ways. Marpa should not be used with untrusted input. In Perl' s taint mode, it is a fatal error to use Marpa's SLIF interface with a tainted grammar, a tainted input string, or tainted token values.
Perl interpreter-based threads are now officially "discouraged". Nonetheless, Marpa::R3 supports them, with one restriction: A Marpa grammar object, and any recognizer or other object created from that grammar, must be used within a single thread.
This restriction is less severe than it may sound: Marpa grammars are relatively inexpensive to create, and several Marpa grammar objects can be created from the same SLIF source.
Marpa:: top-level namespace is reserved. For extensions to Marpa, one appropriate place is the
MarpaX:: namespace. This practice helps avoid namespace collisions, and follows a CPAN standard, as exemplified by the
MooseX:: which are for extensions of, respectively, DBI, LWP and Moose.
Marpa::R3 is Copyright (C) 2017, Jeffrey Kegler. This module is free software; you can redistribute it and/or modify it under the same terms as Perl 5.10.1. For more details, see the full text of the licenses in the directory LICENSES. This program is distributed in the hope that it will be useful, but without any warranty; without even the implied warranty of merchantability or fitness for a particular purpose.