From 594d70332e6d7552f1cb2180b59e1c78bea05ea1 Mon Sep 17 00:00:00 2001
From: Yves Orton <demerphq@gmail.com>
Date: Wed, 7 Feb 2007 23:53:25 +0100
Subject: Re: [PATCH - provisional] H. Merijn Brands idea of buffer numbering.
 Message-ID: <9b18b3110702071353l250d8a67x188c4e234e8905c7@mail.gmail.com>

p4raw-id: //depot/perl@30169
---
 pod/perlre.pod | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

(limited to 'pod')

diff --git a/pod/perlre.pod b/pod/perlre.pod
index d913c8074a..5287965fe3 100644
--- a/pod/perlre.pod
+++ b/pod/perlre.pod
@@ -711,6 +711,29 @@ is equivalent to the more verbose
 
     /(?:(?s-i)more.*than).*million/i
 
+=item C<(?|pattern)>
+X<(?|)> X<Branch reset>
+
+This is the "branch reset" pattern, which has the special property
+that the capture buffers are numbered from the same starting point
+in each branch. 
+
+Normally capture buffers in a pattern are number sequentially, left
+to right in the pattern. Inside of this construct this behaviour is
+overriden so that the captures buffers in each branch share the same
+numbers. The numbering in each branch will be as normal, and any 
+buffers following the use of this pattern will be numbered as though
+the construct contained only one branch, that being the one with the
+most capture buffers in it.
+
+Consider the following pattern. The numbers underneath are which
+buffer number the captured content will be stored in.
+
+
+    # before  ---------------branch-reset----------- after        
+    / ( a )  (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
+    # 1            2         2  3        2     3     4  
+
 =item Look-Around Assertions
 X<look-around assertion> X<lookaround assertion> X<look-around> X<lookaround>
 
-- 
cgit v1.2.1