Author: lwall
Date: 2010-05-17 18:08:01 +0200 (Mon, 17 May 2010)
New Revision: 30667
Modified:
docs/Perl6/Spec/S05-regex.pod
Log:
[S05] Disrequire retroactive semantics on :ii and :aa
Modified: docs/Perl6/Spec/S05-regex.pod
===================================================================
--- docs/Perl6/Spec/S05-regex.pod 2010-05-17 14:43:16 UTC (rev 30666)
+++ docs/Perl6/Spec/S05-regex.pod 2010-05-17 16:08:01 UTC (rev 30667)
@@ -16,8 +16,8 @@
Created: 24 Jun 2002
- Last Modified: 5 May 2010
- Version: 119
+ Last Modified: 17 May 2010
+ Version: 120
This document summarizes Apocalypse 5, which is about the new regex
syntax. We now try to call them I<regex> rather than "regular
@@ -4101,6 +4101,31 @@
(Here we set those explicitly using the C<< <(...)> >> pair; otherwise we
would have had to use lookbehind to match the C<$>.)
+Please note that the C<:ii>/C<:samecase> and C<:aa>/C<:sameaccent>
+switches are really two different modifiers in one, and when the compiler
desugars
+the quote-like forms it distributes semantics to both the pattern
+and the replacement. That is, C<:ii> on the replacement implies a C<:i> on the
+pattern, and C<:aa> implies C<:a>. The proper method equivalents to:
+
+ s:ii/foo/bar/
+ s:aa/boo/far/
+
+are not:
+
+ .subst(/foo/, 'bar', :ii) # WRONG
+ .subst(/boo/, 'far', :aa) # WRONG
+
+but rather:
+
+ .subst(rx:i/foo/, 'bar', :ii) # okay
+ .subst(rx:a/boo/, 'far', :aa) # okay
+
+It is specifically I<not> required of an implementation that it treat
+the regexes as generic with respect to case and accent. Retroactive
+recompilation is considered harmful. If an implement does do lazy
+generic case and accent semantics, it is erroneous and non-portable
+for a program to depend on it.
+
=head1 Positional matching, fixed width types
=over
