add three new EBCDIC Encode-ings && many tests

Message-ID: <Pine.OSF.4.10.10010301643590.67363-100000@aspara.forte.com>

p4raw-id: //depot/perl@7498

diff --git a/MANIFEST b/MANIFEST
index 4ba0a86..d75930b 100644 (file)
--- a/MANIFEST
+++ b/MANIFEST
@@ -195,6 +195,7 @@ ext/Encode/Makefile.PL              Encode extension
 ext/Encode/Todo                        Encode extension
 ext/Encode/Encode/ascii.enc    Encoding tables
 ext/Encode/Encode/big5.enc     Encoding tables
+ext/Encode/Encode/cp1047.enc   Encoding tables
 ext/Encode/Encode/cp1250.enc   Encoding tables
 ext/Encode/Encode/cp1251.enc   Encoding tables
 ext/Encode/Encode/cp1252.enc   Encoding tables
@@ -204,6 +205,7 @@ ext/Encode/Encode/cp1255.enc        Encoding tables
 ext/Encode/Encode/cp1256.enc   Encoding tables
 ext/Encode/Encode/cp1257.enc   Encoding tables
 ext/Encode/Encode/cp1258.enc   Encoding tables
+ext/Encode/Encode/cp37.enc     Encoding tables
 ext/Encode/Encode/cp437.enc    Encoding tables
 ext/Encode/Encode/cp737.enc    Encoding tables
 ext/Encode/Encode/cp775.enc    Encoding tables
@@ -260,6 +262,7 @@ ext/Encode/Encode/macRomania.enc    Encoding tables
 ext/Encode/Encode/macThai.enc          Encoding tables
 ext/Encode/Encode/macTurkish.enc       Encoding tables
 ext/Encode/Encode/macUkraine.enc       Encoding tables
+ext/Encode/Encode/posix-bc.enc         Encoding tables
 ext/Encode/Encode/shiftjis.enc         Encoding tables
 ext/Encode/Encode/symbol.enc           Encoding tables
 ext/Errno/ChangeLog            Errno perl module change log

diff --git a/ext/Encode/Encode/cp1047.enc b/ext/Encode/Encode/cp1047.enc
new file mode 100644 (file)
index 0000000..35e0ed3
--- /dev/null
+++ b/ext/Encode/Encode/cp1047.enc
@@ -0,0 +1,20 @@
+# Encoding file: cp1047 (EBCDIC), single-byte
+S
+003F 0 1
+00
+0000000100020003009C00090086007F0097008D008E000B000C000D000E000F
+0010001100120013009D000A00080087001800190092008F001C001D001E001F
+0080008100820083008400850017001B00880089008A008B008C000500060007
+0090009100160093009400950096000400980099009A009B00140015009E001A
+002000A000E200E400E000E100E300E500E700F100A2002E003C0028002B007C
+002600E900EA00EB00E800ED00EE00EF00EC00DF00210024002A0029003B005E
+002D002F00C200C400C000C100C300C500C700D100A6002C0025005F003E003F
+00F800C900CA00CB00C800CD00CE00CF00CC0060003A002300400027003D0022
+00D800610062006300640065006600670068006900AB00BB00F000FD00FE00B1
+00B0006A006B006C006D006E006F00700071007200AA00BA00E600B800C600A4
+00B5007E0073007400750076007700780079007A00A100BF00D0005B00DE00AE
+00AC00A300A500B700A900A700B600BC00BD00BE00DD00A800AF005D00B400D7
+007B00410042004300440045004600470048004900AD00F400F600F200F300F5
+007D004A004B004C004D004E004F00500051005200B900FB00FC00F900FA00FF
+005C00F70053005400550056005700580059005A00B200D400D600D200D300D5
+003000310032003300340035003600370038003900B300DB00DC00D900DA009F

diff --git a/ext/Encode/Encode/cp37.enc b/ext/Encode/Encode/cp37.enc
new file mode 100644 (file)
index 0000000..f4b575e
--- /dev/null
+++ b/ext/Encode/Encode/cp37.enc
@@ -0,0 +1,20 @@
+# Encoding file: cp37 (EBCDIC), single-byte
+S
+003F 0 1
+00
+0000000100020003009C00090086007F0097008D008E000B000C000D000E000F
+0010001100120013009D008500080087001800190092008F001C001D001E001F
+00800081008200830084000A0017001B00880089008A008B008C000500060007
+0090009100160093009400950096000400980099009A009B00140015009E001A
+002000A000E200E400E000E100E300E500E700F100A2002E003C0028002B007C
+002600E900EA00EB00E800ED00EE00EF00EC00DF00210024002A0029003B00AC
+002D002F00C200C400C000C100C300C500C700D100A6002C0025005F003E003F
+00F800C900CA00CB00C800CD00CE00CF00CC0060003A002300400027003D0022
+00D800610062006300640065006600670068006900AB00BB00F000FD00FE00B1
+00B0006A006B006C006D006E006F00700071007200AA00BA00E600B800C600A4
+00B5007E0073007400750076007700780079007A00A100BF00D000DD00DE00AE
+005E00A300A500B700A900A700B600BC00BD00BE005B005D00AF00A800B400D7
+007B00410042004300440045004600470048004900AD00F400F600F200F300F5
+007D004A004B004C004D004E004F00500051005200B900FB00FC00F900FA00FF
+005C00F70053005400550056005700580059005A00B200D400D600D200D300D5
+003000310032003300340035003600370038003900B300DB00DC00D900DA009F

diff --git a/ext/Encode/Encode/posix-bc.enc b/ext/Encode/Encode/posix-bc.enc
new file mode 100644 (file)
index 0000000..d54f5bb
--- /dev/null
+++ b/ext/Encode/Encode/posix-bc.enc
@@ -0,0 +1,20 @@
+# Encoding file: posix-bc (EBCDIC), single-byte
+S
+003F 0 1
+00
+0000000100020003009C00090086007F0097008D008E000B000C000D000E000F
+0010001100120013009D000A00080087001800190092008F001C001D001E001F
+0080008100820083008400850017001B00880089008A008B008C000500060007
+0090009100160093009400950096000400980099009A009B00140015009E001A
+002000A000E200E400E000E100E300E500E700F10060002E003C0028002B007C
+002600E900EA00EB00E800ED00EE00EF00EC00DF00210024002A0029003B009F
+002D002F00C200C400C000C100C300C500C700D1005E002C0025005F003E003F
+00F800C900CA00CB00C800CD00CE00CF00CC00A8003A002300400027003D0022
+00D800610062006300640065006600670068006900AB00BB00F000FD00FE00B1
+00B0006A006B006C006D006E006F00700071007200AA00BA00E600B800C600A4
+00B500AF0073007400750076007700780079007A00A100BF00D000DD00DE00AE
+00A200A300A500B700A900A700B600BC00BD00BE00AC005B005C005D00B400D7
+00F900410042004300440045004600470048004900AD00F400F600F200F300F5
+00A6004A004B004C004D004E004F00500051005200B900FB00FC00DB00FA00FF
+00D900F70053005400550056005700580059005A00B200D400D600D200D300D5
+003000310032003300340035003600370038003900B3007B00DC007D00DA007E

diff --git a/t/lib/encode.t b/t/lib/encode.t
index 568efd3..34527d7 100644 (file)
--- a/t/lib/encode.t
+++ b/t/lib/encode.t
@@ -12,7 +12,11 @@ use Encode qw(from_to);
 use charnames qw(greek);
 my @encodings = grep(/iso8859/,Encode::encodings());
 my $n = 2;
-plan test => 13+$n*@encodings;
+my @character_set = ('0'..'9', 'A'..'Z', 'a'..'z');
+my @source = qw(ascii iso8859-1 cp1250);
+my @destiny = qw(cp1047 cp37 posix-bc);
+my @ebcdic_sets = qw(cp1047 cp37 posix-bc);
+plan test => 13+$n*@encodings + 2*@source*@destiny*@character_set + 2*@ebcdic_sets*256;
 my $str = join('',map(chr($_),0x20..0x7E));
 my $cpy = $str;
 ok(length($str),from_to($cpy,'iso8859-1','Unicode'),"Length Wrong");
@@ -27,7 +31,7 @@ ok(length($str),from_to($cpy,'iso8859-1','Unicode'),"Length Wrong");
 
 my $sym = Encode->getEncoding('symbol');
 my $uni = $sym->toUnicode('a');
-ok("\N{alpha}",substr($uni,0,1),"alpha does not map so symbol 'a'");
+ok("\N{alpha}",substr($uni,0,1),"alpha does not map to symbol 'a'");
 $str = $sym->fromUnicode("\N{Beta}");
 ok("B",substr($str,0,1),"Symbol 'B' does not map to Beta");
 
@@ -41,3 +45,49 @@ foreach my $enc (qw(symbol dingbats ascii),@encodings)
   ok($cpy,$str,"$enc mangled translating to Unicode and back");
  }
 
+# On ASCII based machines see if we can map several codepoints from
+# three distinct ASCII sets to three distinct EBCDIC coded character sets.
+# On EBCDIC machines see if we can map from three EBCDIC sets to three 
+# distinct ASCII sets.
+
+my @expectation = (240..249, 193..201,209..217,226..233, 129..137,145..153,162..169);
+if (ord('A') != 65) {
+    my @temp = @destiny;
+    @destiny = @source;
+    @source = @temp;
+    undef(@temp);
+    @expectation = (48..57, 65..90, 97..122);
+}
+
+foreach my $to (@destiny)
+ {
+  foreach my $from (@source)
+   {
+    my @expected = @expectation;
+    foreach my $chr (@character_set)
+     {
+      my $native_chr = $chr;
+      my $cpy = $chr;
+      my $rc = from_to($cpy,$from,$to);
+      ok(1,$rc,"Could not translate from $from to $to");
+      ok(ord($cpy),shift(@expected),"mangled translating $native_chr from $from to $to");
+     }
+   }
+ }
+
+# On either ASCII or EBCDIC machines ensure we can take the full one
+# byte repetoire to EBCDIC sets and back.
+
+my $enc_as = 'iso8859-1';
+foreach my $enc_eb (@ebcdic_sets)
+ {
+  foreach my $ord (0..255)
+   {
+    $str = chr($ord);
+    my $rc = from_to($str,$enc_as,$enc_eb);
+    $rc += from_to($str,$enc_eb,$enc_as);
+    ok($rc,2,"return code for $ord $enc_eb -> $enc_as -> $enc_eb was not obtained");
+    ok($ord,ord($str),"$enc_as mangled translating $ord to $enc_eb and back");
+   }
+ }
+