EBCDIC: the toke.c wariness about UTF-8 (really, UTF-EBCDIC)

scripts does not apply any more.  (The utf8.t needed a little
bit of relaxing to work in EBCDIC, since *both* the bytes in
subtest #30 are illegal UTF-EBCDIC, and the error message
matched is slightly different.)  The utf8.t now passes.

p4raw-id: //depot/perl@16550

diff --git a/lib/utf8.t b/lib/utf8.t
index acd64bf..e11d856 100644 (file)
--- a/lib/utf8.t
+++ b/lib/utf8.t
@@ -166,7 +166,7 @@ plan tests => 94;
     use utf8; %a = ("\xE1\xA0"=>"sterling");
     print 'start'; printf '%x,', ord \$_ foreach keys %a; print "end\n";
 BANG
-             qr/^Malformed UTF-8 character \(2 bytes, need 3.+\).*start\d+,end$/s
+             qr/^Malformed UTF-8 character \(\d bytes?, need \d, .+\).*start\d+,end$/sm
             ],
             );
     foreach (@tests) {

diff --git a/toke.c b/toke.c
index 5635e2c..7265fb0 100644 (file)
--- a/toke.c
+++ b/toke.c
@@ -41,11 +41,7 @@ static I32 utf16rev_textfilter(pTHX_ int idx, SV *sv, int maxlen);
 #ifdef USE_UTF8_SCRIPTS
 #   define UTF (!IN_BYTES)
 #else
-#   ifdef EBCDIC /* For now 'use utf8' does not affect tokenizer on EBCDIC */
-#       define UTF (PL_linestr && DO_UTF8(PL_linestr))
-#   else
-#       define UTF ((PL_linestr && DO_UTF8(PL_linestr)) || (PL_hints & HINT_UTF8))
-#   endif
+#   define UTF ((PL_linestr && DO_UTF8(PL_linestr)) || (PL_hints & HINT_UTF8))
 #endif
 
 /* In variables named $^X, these are the legal values for X.