Remove the 'asciir' re subpragma. Should instead implement

the 'physical vs logical' range scheme:

\xAA-\xCC is a native physical range, you want that range of
codepoints in your native encoding.  In EBCDIC the codepoints
in the gaps (between i-j and r-s) should be included.

\x{AA}-\x{CC} is a physical Unicode range, you want that range of
codepoints in Unicode.

a-z is a logical range, you want that range of 'logical' codepoints
in your native encoding.  In EBCDIC the codepoints in the gaps
(between i-j and r-s) should not be included.

Mixed cases (a-\xAA, etc) should either be errors, or maybe
the 'logical' endpoints should be converted to native/Unicode
codepoints, and the range handled as a physical range.

'Logical endpoints' are to be recognized only in the A-Z, a-z,
and 0-9 ranges.  Probably a warning should be given for mixed
cases like A-z or a-9 (since such expressions are encoding
dependent), with a recommendation to use physical ranges.

p4raw-id: //depot/perl@10085

2 files changed, 0 insertions, 4 deletions

diff --git a/t/op/pat.t b/t/op/pat.t
index 1be72346f8..0df4d786ee 100755
--- a/t/op/pat.t
+++ b/t/op/pat.t
@@ -13,8 +13,6 @@ BEGIN {
     @INC = '../lib';
 }
 
-use re 'asciirange'; # Compute ranges in ASCII space
-
 eval 'use Config';          #  Defaults assumed if this fails
 
 $x = "abc\ndef\n";
diff --git a/t/op/regexp.t b/t/op/regexp.t
index 0b81e714a9..4a4d42fd98 100755
--- a/t/op/regexp.t
+++ b/t/op/regexp.t
@@ -36,8 +36,6 @@ BEGIN {
     @INC = '../lib';
 }
 
-use re 'asciirange'; # ranges are computed in ASCII
-
 $iters = shift || 1;		# Poor man performance suite, 10000 is OK.
 
 open(TESTS,'op/re_tests') || open(TESTS,'t/op/re_tests') ||