Subject: [String-SIG]  \ escapes in re.sub
From: "Andrew M. Kuchling" <amk@mira.idsonline.com>
To: string-sig@python.org
Date: Sun, 22 Mar 1998 22:48:27 -0500 (EST)

The patch below changes the way escapes are handled in the replacement
string of re.sub and re.subn.  An unknown escape like \s is currently
treated as just an 's', which is different from the way unknown
escapes are usually treated in Python's string literals.  This patch
modifies the behaviour so that the \ is preserved, and \s results in
'\s'.

	This was reported as a bug a little while back.  I'm not sure
if it really is; keeping just the 's' is inconsistent with Python
string literals, but it's consistent with re patterns.  For example,
r'\j' matches just 'j'.  It can be argued that the behaviour should be
left as it is.  So it's a stylistic question: which consistency is
better?  Which inconsistency is worse?

	Some of the tests in test_re.py will break as a result of this
change; I'm going to take care of that in a separate patch in a little
bit, if it's decided that the change should be accepted.

	Once this is settled, I'll be issuing a new PCRE release that
implements \g<1> in replacement strings, as a synonym for \1 that
isn't vulnerable to ambiguities as in \10\1.  No patch will be made
for that, because it's a new feature and not a bug.


A.M. Kuchling				http://starship.skyport.net/crew/amk/
There was not a lot that could be done to make Morpork a worse place. A direct
hit by a meteorite, for example, would count as gentrification.
	-- Terry Pratchett, _Pyramids_


*** pcremodule.c	1998/03/11 04:49:44	1.9
--- Modules/pcremodule.c	1998/03/19 03:55:57
***************
*** 73,77 ****
  #define BEGINNING_OF_BUFFER	7
  #define END_OF_BUFFER		8
! 
  
  static PcreObject *
--- 73,77 ----
  #define BEGINNING_OF_BUFFER	7
  #define END_OF_BUFFER		8
! #define STRING                  9
  
  static PcreObject *
***************
*** 284,287 ****
--- 284,291 ----
  		return Py_BuildValue("c", (char)8);
  		break;
+ 	case('\\'):
+ 		*indexptr=index;
+ 		return Py_BuildValue("c", '\\');
+ 		break;
  
  	case('x'):
***************
*** 450,455 ****
  
  	default:
  		*indexptr = index;
! 		return Py_BuildValue("c", c);
  		break;
  	}
--- 454,462 ----
  
  	default:
+ 	  /* It's some unknown escape like \s, so return a string containing
+ 	     \s */
+ 		*typeptr = STRING;
  		*indexptr = index;
! 		return Py_BuildValue("s#", pattern+index-2, 2);
  		break;
  	}
***************
*** 543,546 ****
--- 550,559 ----
  			}
  			break;
+ 			case(STRING):
+ 			  {
+ 			    PyList_Append(results, value);
+ 			    total_len += PyString_Size(value);
+ 			    break;
+ 			  }
  			default:
  				Py_DECREF(results);

------------------------------------------------------
String-SIG maillist  -  String-SIG@python.org
http://www.python.org/mailman/listinfo/string-sig