1=encoding utf8 2 3=head1 NAME 4 5perlebcdic - Considerations for running Perl on EBCDIC platforms 6 7=head1 DESCRIPTION 8 9An exploration of some of the issues facing Perl programmers 10on EBCDIC based computers. We do not cover localization, 11internationalization, or multi-byte character set issues other 12than some discussion of UTF-8 and UTF-EBCDIC. 13 14Portions that are still incomplete are marked with XXX. 15 16Perl used to work on EBCDIC machines, but there are now areas of the code where 17it doesn't. If you want to use Perl on an EBCDIC machine, please let us know 18by sending mail to perlbug@perl.org 19 20=head1 COMMON CHARACTER CODE SETS 21 22=head2 ASCII 23 24The American Standard Code for Information Interchange (ASCII or US-ASCII) is a 25set of 26integers running from 0 to 127 (decimal) that imply character 27interpretation by the display and other systems of computers. 28The range 0..127 can be covered by setting the bits in a 7-bit binary 29digit, hence the set is sometimes referred to as "7-bit ASCII". 30ASCII was described by the American National Standards Institute 31document ANSI X3.4-1986. It was also described by ISO 646:1991 32(with localization for currency symbols). The full ASCII set is 33given in the table below as the first 128 elements. Languages that 34can be written adequately with the characters in ASCII include 35English, Hawaiian, Indonesian, Swahili and some Native American 36languages. 37 38There are many character sets that extend the range of integers 39from 0..2**7-1 up to 2**8-1, or 8 bit bytes (octets if you prefer). 40One common one is the ISO 8859-1 character set. 41 42=head2 ISO 8859 43 44The ISO 8859-$n are a collection of character code sets from the 45International Organization for Standardization (ISO), each of which 46adds characters to the ASCII set that are typically found in European 47languages, many of which are based on the Roman, or Latin, alphabet. 48 49=head2 Latin 1 (ISO 8859-1) 50 51A particular 8-bit extension to ASCII that includes grave and acute 52accented Latin characters. Languages that can employ ISO 8859-1 53include all the languages covered by ASCII as well as Afrikaans, 54Albanian, Basque, Catalan, Danish, Faroese, Finnish, Norwegian, 55Portuguese, Spanish, and Swedish. Dutch is covered albeit without 56the ij ligature. French is covered too but without the oe ligature. 57German can use ISO 8859-1 but must do so without German-style 58quotation marks. This set is based on Western European extensions 59to ASCII and is commonly encountered in world wide web work. 60In IBM character code set identification terminology ISO 8859-1 is 61also known as CCSID 819 (or sometimes 0819 or even 00819). 62 63=head2 EBCDIC 64 65The Extended Binary Coded Decimal Interchange Code refers to a 66large collection of single- and multi-byte coded character sets that are 67different from ASCII or ISO 8859-1 and are all slightly different from each 68other; they typically run on host computers. The EBCDIC encodings derive from 698-bit byte extensions of Hollerith punched card encodings. The layout on the 70cards was such that high bits were set for the upper and lower case alphabet 71characters [a-z] and [A-Z], but there were gaps within each Latin alphabet 72range. 73 74Some IBM EBCDIC character sets may be known by character code set 75identification numbers (CCSID numbers) or code page numbers. 76 77Perl can be compiled on platforms that run any of three commonly used EBCDIC 78character sets, listed below. 79 80=head3 The 13 variant characters 81 82Among IBM EBCDIC character code sets there are 13 characters that 83are often mapped to different integer values. Those characters 84are known as the 13 "variant" characters and are: 85 86 \ [ ] { } ^ ~ ! # | $ @ ` 87 88When Perl is compiled for a platform, it looks at some of these characters to 89guess which EBCDIC character set the platform uses, and adapts itself 90accordingly to that platform. If the platform uses a character set that is not 91one of the three Perl knows about, Perl will either fail to compile, or 92mistakenly and silently choose one of the three. 93They are: 94 95=over 96 97=item B<0037> 98 99Character code set ID 0037 is a mapping of the ASCII plus Latin-1 100characters (i.e. ISO 8859-1) to an EBCDIC set. 0037 is used 101in North American English locales on the OS/400 operating system 102that runs on AS/400 computers. CCSID 0037 differs from ISO 8859-1 103in 237 places, in other words they agree on only 19 code point values. 104 105=item B<1047> 106 107Character code set ID 1047 is also a mapping of the ASCII plus 108Latin-1 characters (i.e. ISO 8859-1) to an EBCDIC set. 1047 is 109used under Unix System Services for OS/390 or z/OS, and OpenEdition 110for VM/ESA. CCSID 1047 differs from CCSID 0037 in eight places. 111 112=item B<POSIX-BC> 113 114The EBCDIC code page in use on Siemens' BS2000 system is distinct from 1151047 and 0037. It is identified below as the POSIX-BC set. 116 117=back 118 119=head2 Unicode code points versus EBCDIC code points 120 121In Unicode terminology a I<code point> is the number assigned to a 122character: for example, in EBCDIC the character "A" is usually assigned 123the number 193. In Unicode the character "A" is assigned the number 65. 124This causes a problem with the semantics of the pack/unpack "U", which 125are supposed to pack Unicode code points to characters and back to numbers. 126The problem is: which code points to use for code points less than 256? 127(for 256 and over there's no problem: Unicode code points are used) 128In EBCDIC, for the low 256 the EBCDIC code points are used. This 129means that the equivalences 130 131 pack("U", ord($character)) eq $character 132 unpack("U", $character) == ord $character 133 134will hold. (If Unicode code points were applied consistently over 135all the possible code points, pack("U",ord("A")) would in EBCDIC 136equal I<A with acute> or chr(101), and unpack("U", "A") would equal 13765, or I<non-breaking space>, not 193, or ord "A".) 138 139=head2 Remaining Perl Unicode problems in EBCDIC 140 141=over 4 142 143=item * 144 145Many of the remaining problems seem to be related to case-insensitive matching 146 147=item * 148 149The extensions Unicode::Collate and Unicode::Normalized are not 150supported under EBCDIC, likewise for the encoding pragma. 151 152=back 153 154=head2 Unicode and UTF 155 156UTF stands for C<Unicode Transformation Format>. 157UTF-8 is an encoding of Unicode into a sequence of 8-bit byte chunks, based on 158ASCII and Latin-1. 159The length of a sequence required to represent a Unicode code point 160depends on the ordinal number of that code point, 161with larger numbers requiring more bytes. 162UTF-EBCDIC is like UTF-8, but based on EBCDIC. 163 164You may see the term C<invariant> character or code point. 165This simply means that the character has the same numeric 166value when encoded as when not. 167(Note that this is a very different concept from L</The 13 variant characters> 168mentioned above.) 169For example, the ordinal value of 'A' is 193 in most EBCDIC code pages, 170and also is 193 when encoded in UTF-EBCDIC. 171All variant code points occupy at least two bytes when encoded. 172In UTF-8, the code points corresponding to the lowest 128 173ordinal numbers (0 - 127: the ASCII characters) are invariant. 174In UTF-EBCDIC, there are 160 invariant characters. 175(If you care, the EBCDIC invariants are those characters 176which have ASCII equivalents, plus those that correspond to 177the C1 controls (80..9f on ASCII platforms).) 178 179A string encoded in UTF-EBCDIC may be longer (but never shorter) than 180one encoded in UTF-8. 181 182=head2 Using Encode 183 184Starting from Perl 5.8 you can use the standard new module Encode 185to translate from EBCDIC to Latin-1 code points. 186Encode knows about more EBCDIC character sets than Perl can currently 187be compiled to run on. 188 189 use Encode 'from_to'; 190 191 my %ebcdic = ( 176 => 'cp37', 95 => 'cp1047', 106 => 'posix-bc' ); 192 193 # $a is in EBCDIC code points 194 from_to($a, $ebcdic{ord '^'}, 'latin1'); 195 # $a is ISO 8859-1 code points 196 197and from Latin-1 code points to EBCDIC code points 198 199 use Encode 'from_to'; 200 201 my %ebcdic = ( 176 => 'cp37', 95 => 'cp1047', 106 => 'posix-bc' ); 202 203 # $a is ISO 8859-1 code points 204 from_to($a, 'latin1', $ebcdic{ord '^'}); 205 # $a is in EBCDIC code points 206 207For doing I/O it is suggested that you use the autotranslating features 208of PerlIO, see L<perluniintro>. 209 210Since version 5.8 Perl uses the new PerlIO I/O library. This enables 211you to use different encodings per IO channel. For example you may use 212 213 use Encode; 214 open($f, ">:encoding(ascii)", "test.ascii"); 215 print $f "Hello World!\n"; 216 open($f, ">:encoding(cp37)", "test.ebcdic"); 217 print $f "Hello World!\n"; 218 open($f, ">:encoding(latin1)", "test.latin1"); 219 print $f "Hello World!\n"; 220 open($f, ">:encoding(utf8)", "test.utf8"); 221 print $f "Hello World!\n"; 222 223to get four files containing "Hello World!\n" in ASCII, CP 0037 EBCDIC, 224ISO 8859-1 (Latin-1) (in this example identical to ASCII since only ASCII 225characters were printed), and 226UTF-EBCDIC (in this example identical to normal EBCDIC since only characters 227that don't differ between EBCDIC and UTF-EBCDIC were printed). See the 228documentation of Encode::PerlIO for details. 229 230As the PerlIO layer uses raw IO (bytes) internally, all this totally 231ignores things like the type of your filesystem (ASCII or EBCDIC). 232 233=head1 SINGLE OCTET TABLES 234 235The following tables list the ASCII and Latin 1 ordered sets including 236the subsets: C0 controls (0..31), ASCII graphics (32..7e), delete (7f), 237C1 controls (80..9f), and Latin-1 (a.k.a. ISO 8859-1) (a0..ff). In the 238table names of the Latin 1 239extensions to ASCII have been labelled with character names roughly 240corresponding to I<The Unicode Standard, Version 6.1> albeit with 241substitutions such as s/LATIN// and s/VULGAR// in all cases, s/CAPITAL 242LETTER// in some cases, and s/SMALL LETTER ([A-Z])/\l$1/ in some other 243cases. Controls are listed using their Unicode 6.2 abbreviations. 244The differences between the 0037 and 1047 sets are 245flagged with **. The differences between the 1047 and POSIX-BC sets 246are flagged with ##. All ord() numbers listed are decimal. If you 247would rather see this table listing octal values, then run the table 248(that is, the pod source text of this document, since this recipe may not 249work with a pod2_other_format translation) through: 250 251=over 4 252 253=item recipe 0 254 255=back 256 257 perl -ne 'if(/(.{29})(\d+)\s+(\d+)\s+(\d+)\s+(\d+)/)' \ 258 -e '{printf("%s%-5.03o%-5.03o%-5.03o%.03o\n",$1,$2,$3,$4,$5)}' \ 259 perlebcdic.pod 260 261If you want to retain the UTF-x code points then in script form you 262might want to write: 263 264=over 4 265 266=item recipe 1 267 268=back 269 270 open(FH,"<perlebcdic.pod") or die "Could not open perlebcdic.pod: $!"; 271 while (<FH>) { 272 if (/(.{29})(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\.?(\d*)\s+(\d+)\.?(\d*)/) 273 { 274 if ($7 ne '' && $9 ne '') { 275 printf( 276 "%s%-5.03o%-5.03o%-5.03o%-5.03o%-3o.%-5o%-3o.%.03o\n", 277 $1,$2,$3,$4,$5,$6,$7,$8,$9); 278 } 279 elsif ($7 ne '') { 280 printf("%s%-5.03o%-5.03o%-5.03o%-5.03o%-3o.%-5o%.03o\n", 281 $1,$2,$3,$4,$5,$6,$7,$8); 282 } 283 else { 284 printf("%s%-5.03o%-5.03o%-5.03o%-5.03o%-5.03o%.03o\n", 285 $1,$2,$3,$4,$5,$6,$8); 286 } 287 } 288 } 289 290If you would rather see this table listing hexadecimal values then 291run the table through: 292 293=over 4 294 295=item recipe 2 296 297=back 298 299 perl -ne 'if(/(.{29})(\d+)\s+(\d+)\s+(\d+)\s+(\d+)/)' \ 300 -e '{printf("%s%-5.02X%-5.02X%-5.02X%.02X\n",$1,$2,$3,$4,$5)}' \ 301 perlebcdic.pod 302 303Or, in order to retain the UTF-x code points in hexadecimal: 304 305=over 4 306 307=item recipe 3 308 309=back 310 311 open(FH,"<perlebcdic.pod") or die "Could not open perlebcdic.pod: $!"; 312 while (<FH>) { 313 if (/(.{29})(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\.?(\d*)\s+(\d+)\.?(\d*)/) 314 { 315 if ($7 ne '' && $9 ne '') { 316 printf( 317 "%s%-5.02X%-5.02X%-5.02X%-5.02X%-2X.%-6.02X%02X.%02X\n", 318 $1,$2,$3,$4,$5,$6,$7,$8,$9); 319 } 320 elsif ($7 ne '') { 321 printf("%s%-5.02X%-5.02X%-5.02X%-5.02X%-2X.%-6.02X%02X\n", 322 $1,$2,$3,$4,$5,$6,$7,$8); 323 } 324 else { 325 printf("%s%-5.02X%-5.02X%-5.02X%-5.02X%-5.02X%02X\n", 326 $1,$2,$3,$4,$5,$6,$8); 327 } 328 } 329 } 330 331 332 ISO 333 8859-1 POS- 334 CCSID CCSID CCSID IX- 335 chr 0819 0037 1047 BC UTF-8 UTF-EBCDIC 336 --------------------------------------------------------------------- 337 <NUL> 0 0 0 0 0 0 338 <SOH> 1 1 1 1 1 1 339 <STX> 2 2 2 2 2 2 340 <ETX> 3 3 3 3 3 3 341 <EOT> 4 55 55 55 4 55 342 <ENQ> 5 45 45 45 5 45 343 <ACK> 6 46 46 46 6 46 344 <BEL> 7 47 47 47 7 47 345 <BS> 8 22 22 22 8 22 346 <HT> 9 5 5 5 9 5 347 <LF> 10 37 21 21 10 21 ** 348 <VT> 11 11 11 11 11 11 349 <FF> 12 12 12 12 12 12 350 <CR> 13 13 13 13 13 13 351 <SO> 14 14 14 14 14 14 352 <SI> 15 15 15 15 15 15 353 <DLE> 16 16 16 16 16 16 354 <DC1> 17 17 17 17 17 17 355 <DC2> 18 18 18 18 18 18 356 <DC3> 19 19 19 19 19 19 357 <DC4> 20 60 60 60 20 60 358 <NAK> 21 61 61 61 21 61 359 <SYN> 22 50 50 50 22 50 360 <ETB> 23 38 38 38 23 38 361 <CAN> 24 24 24 24 24 24 362 <EOM> 25 25 25 25 25 25 363 <SUB> 26 63 63 63 26 63 364 <ESC> 27 39 39 39 27 39 365 <FS> 28 28 28 28 28 28 366 <GS> 29 29 29 29 29 29 367 <RS> 30 30 30 30 30 30 368 <US> 31 31 31 31 31 31 369 <SPACE> 32 64 64 64 32 64 370 ! 33 90 90 90 33 90 371 " 34 127 127 127 34 127 372 # 35 123 123 123 35 123 373 $ 36 91 91 91 36 91 374 % 37 108 108 108 37 108 375 & 38 80 80 80 38 80 376 ' 39 125 125 125 39 125 377 ( 40 77 77 77 40 77 378 ) 41 93 93 93 41 93 379 * 42 92 92 92 42 92 380 + 43 78 78 78 43 78 381 , 44 107 107 107 44 107 382 - 45 96 96 96 45 96 383 . 46 75 75 75 46 75 384 / 47 97 97 97 47 97 385 0 48 240 240 240 48 240 386 1 49 241 241 241 49 241 387 2 50 242 242 242 50 242 388 3 51 243 243 243 51 243 389 4 52 244 244 244 52 244 390 5 53 245 245 245 53 245 391 6 54 246 246 246 54 246 392 7 55 247 247 247 55 247 393 8 56 248 248 248 56 248 394 9 57 249 249 249 57 249 395 : 58 122 122 122 58 122 396 ; 59 94 94 94 59 94 397 < 60 76 76 76 60 76 398 = 61 126 126 126 61 126 399 > 62 110 110 110 62 110 400 ? 63 111 111 111 63 111 401 @ 64 124 124 124 64 124 402 A 65 193 193 193 65 193 403 B 66 194 194 194 66 194 404 C 67 195 195 195 67 195 405 D 68 196 196 196 68 196 406 E 69 197 197 197 69 197 407 F 70 198 198 198 70 198 408 G 71 199 199 199 71 199 409 H 72 200 200 200 72 200 410 I 73 201 201 201 73 201 411 J 74 209 209 209 74 209 412 K 75 210 210 210 75 210 413 L 76 211 211 211 76 211 414 M 77 212 212 212 77 212 415 N 78 213 213 213 78 213 416 O 79 214 214 214 79 214 417 P 80 215 215 215 80 215 418 Q 81 216 216 216 81 216 419 R 82 217 217 217 82 217 420 S 83 226 226 226 83 226 421 T 84 227 227 227 84 227 422 U 85 228 228 228 85 228 423 V 86 229 229 229 86 229 424 W 87 230 230 230 87 230 425 X 88 231 231 231 88 231 426 Y 89 232 232 232 89 232 427 Z 90 233 233 233 90 233 428 [ 91 186 173 187 91 173 ** ## 429 \ 92 224 224 188 92 224 ## 430 ] 93 187 189 189 93 189 ** 431 ^ 94 176 95 106 94 95 ** ## 432 _ 95 109 109 109 95 109 433 ` 96 121 121 74 96 121 ## 434 a 97 129 129 129 97 129 435 b 98 130 130 130 98 130 436 c 99 131 131 131 99 131 437 d 100 132 132 132 100 132 438 e 101 133 133 133 101 133 439 f 102 134 134 134 102 134 440 g 103 135 135 135 103 135 441 h 104 136 136 136 104 136 442 i 105 137 137 137 105 137 443 j 106 145 145 145 106 145 444 k 107 146 146 146 107 146 445 l 108 147 147 147 108 147 446 m 109 148 148 148 109 148 447 n 110 149 149 149 110 149 448 o 111 150 150 150 111 150 449 p 112 151 151 151 112 151 450 q 113 152 152 152 113 152 451 r 114 153 153 153 114 153 452 s 115 162 162 162 115 162 453 t 116 163 163 163 116 163 454 u 117 164 164 164 117 164 455 v 118 165 165 165 118 165 456 w 119 166 166 166 119 166 457 x 120 167 167 167 120 167 458 y 121 168 168 168 121 168 459 z 122 169 169 169 122 169 460 { 123 192 192 251 123 192 ## 461 | 124 79 79 79 124 79 462 } 125 208 208 253 125 208 ## 463 ~ 126 161 161 255 126 161 ## 464 <DEL> 127 7 7 7 127 7 465 <PAD> 128 32 32 32 194.128 32 466 <HOP> 129 33 33 33 194.129 33 467 <BPH> 130 34 34 34 194.130 34 468 <NBH> 131 35 35 35 194.131 35 469 <IND> 132 36 36 36 194.132 36 470 <NEL> 133 21 37 37 194.133 37 ** 471 <SSA> 134 6 6 6 194.134 6 472 <ESA> 135 23 23 23 194.135 23 473 <HTS> 136 40 40 40 194.136 40 474 <HTJ> 137 41 41 41 194.137 41 475 <VTS> 138 42 42 42 194.138 42 476 <PLD> 139 43 43 43 194.139 43 477 <PLU> 140 44 44 44 194.140 44 478 <RI> 141 9 9 9 194.141 9 479 <SS2> 142 10 10 10 194.142 10 480 <SS3> 143 27 27 27 194.143 27 481 <DCS> 144 48 48 48 194.144 48 482 <PU1> 145 49 49 49 194.145 49 483 <PU2> 146 26 26 26 194.146 26 484 <STS> 147 51 51 51 194.147 51 485 <CCH> 148 52 52 52 194.148 52 486 <MW> 149 53 53 53 194.149 53 487 <SPA> 150 54 54 54 194.150 54 488 <EPA> 151 8 8 8 194.151 8 489 <SOS> 152 56 56 56 194.152 56 490 <SGC> 153 57 57 57 194.153 57 491 <SCI> 154 58 58 58 194.154 58 492 <CSI> 155 59 59 59 194.155 59 493 <ST> 156 4 4 4 194.156 4 494 <OSC> 157 20 20 20 194.157 20 495 <PM> 158 62 62 62 194.158 62 496 <APC> 159 255 255 95 194.159 255 ## 497 <NON-BREAKING SPACE> 160 65 65 65 194.160 128.65 498 <INVERTED "!" > 161 170 170 170 194.161 128.66 499 <CENT SIGN> 162 74 74 176 194.162 128.67 ## 500 <POUND SIGN> 163 177 177 177 194.163 128.68 501 <CURRENCY SIGN> 164 159 159 159 194.164 128.69 502 <YEN SIGN> 165 178 178 178 194.165 128.70 503 <BROKEN BAR> 166 106 106 208 194.166 128.71 ## 504 <SECTION SIGN> 167 181 181 181 194.167 128.72 505 <DIAERESIS> 168 189 187 121 194.168 128.73 ** ## 506 <COPYRIGHT SIGN> 169 180 180 180 194.169 128.74 507 <FEMININE ORDINAL> 170 154 154 154 194.170 128.81 508 <LEFT POINTING GUILLEMET> 171 138 138 138 194.171 128.82 509 <NOT SIGN> 172 95 176 186 194.172 128.83 ** ## 510 <SOFT HYPHEN> 173 202 202 202 194.173 128.84 511 <REGISTERED TRADE MARK> 174 175 175 175 194.174 128.85 512 <MACRON> 175 188 188 161 194.175 128.86 ## 513 <DEGREE SIGN> 176 144 144 144 194.176 128.87 514 <PLUS-OR-MINUS SIGN> 177 143 143 143 194.177 128.88 515 <SUPERSCRIPT TWO> 178 234 234 234 194.178 128.89 516 <SUPERSCRIPT THREE> 179 250 250 250 194.179 128.98 517 <ACUTE ACCENT> 180 190 190 190 194.180 128.99 518 <MICRO SIGN> 181 160 160 160 194.181 128.100 519 <PARAGRAPH SIGN> 182 182 182 182 194.182 128.101 520 <MIDDLE DOT> 183 179 179 179 194.183 128.102 521 <CEDILLA> 184 157 157 157 194.184 128.103 522 <SUPERSCRIPT ONE> 185 218 218 218 194.185 128.104 523 <MASC. ORDINAL INDICATOR> 186 155 155 155 194.186 128.105 524 <RIGHT POINTING GUILLEMET> 187 139 139 139 194.187 128.106 525 <FRACTION ONE QUARTER> 188 183 183 183 194.188 128.112 526 <FRACTION ONE HALF> 189 184 184 184 194.189 128.113 527 <FRACTION THREE QUARTERS> 190 185 185 185 194.190 128.114 528 <INVERTED QUESTION MARK> 191 171 171 171 194.191 128.115 529 <A WITH GRAVE> 192 100 100 100 195.128 138.65 530 <A WITH ACUTE> 193 101 101 101 195.129 138.66 531 <A WITH CIRCUMFLEX> 194 98 98 98 195.130 138.67 532 <A WITH TILDE> 195 102 102 102 195.131 138.68 533 <A WITH DIAERESIS> 196 99 99 99 195.132 138.69 534 <A WITH RING ABOVE> 197 103 103 103 195.133 138.70 535 <CAPITAL LIGATURE AE> 198 158 158 158 195.134 138.71 536 <C WITH CEDILLA> 199 104 104 104 195.135 138.72 537 <E WITH GRAVE> 200 116 116 116 195.136 138.73 538 <E WITH ACUTE> 201 113 113 113 195.137 138.74 539 <E WITH CIRCUMFLEX> 202 114 114 114 195.138 138.81 540 <E WITH DIAERESIS> 203 115 115 115 195.139 138.82 541 <I WITH GRAVE> 204 120 120 120 195.140 138.83 542 <I WITH ACUTE> 205 117 117 117 195.141 138.84 543 <I WITH CIRCUMFLEX> 206 118 118 118 195.142 138.85 544 <I WITH DIAERESIS> 207 119 119 119 195.143 138.86 545 <CAPITAL LETTER ETH> 208 172 172 172 195.144 138.87 546 <N WITH TILDE> 209 105 105 105 195.145 138.88 547 <O WITH GRAVE> 210 237 237 237 195.146 138.89 548 <O WITH ACUTE> 211 238 238 238 195.147 138.98 549 <O WITH CIRCUMFLEX> 212 235 235 235 195.148 138.99 550 <O WITH TILDE> 213 239 239 239 195.149 138.100 551 <O WITH DIAERESIS> 214 236 236 236 195.150 138.101 552 <MULTIPLICATION SIGN> 215 191 191 191 195.151 138.102 553 <O WITH STROKE> 216 128 128 128 195.152 138.103 554 <U WITH GRAVE> 217 253 253 224 195.153 138.104 ## 555 <U WITH ACUTE> 218 254 254 254 195.154 138.105 556 <U WITH CIRCUMFLEX> 219 251 251 221 195.155 138.106 ## 557 <U WITH DIAERESIS> 220 252 252 252 195.156 138.112 558 <Y WITH ACUTE> 221 173 186 173 195.157 138.113 ** ## 559 <CAPITAL LETTER THORN> 222 174 174 174 195.158 138.114 560 <SMALL LETTER SHARP S> 223 89 89 89 195.159 138.115 561 <a WITH GRAVE> 224 68 68 68 195.160 139.65 562 <a WITH ACUTE> 225 69 69 69 195.161 139.66 563 <a WITH CIRCUMFLEX> 226 66 66 66 195.162 139.67 564 <a WITH TILDE> 227 70 70 70 195.163 139.68 565 <a WITH DIAERESIS> 228 67 67 67 195.164 139.69 566 <a WITH RING ABOVE> 229 71 71 71 195.165 139.70 567 <SMALL LIGATURE ae> 230 156 156 156 195.166 139.71 568 <c WITH CEDILLA> 231 72 72 72 195.167 139.72 569 <e WITH GRAVE> 232 84 84 84 195.168 139.73 570 <e WITH ACUTE> 233 81 81 81 195.169 139.74 571 <e WITH CIRCUMFLEX> 234 82 82 82 195.170 139.81 572 <e WITH DIAERESIS> 235 83 83 83 195.171 139.82 573 <i WITH GRAVE> 236 88 88 88 195.172 139.83 574 <i WITH ACUTE> 237 85 85 85 195.173 139.84 575 <i WITH CIRCUMFLEX> 238 86 86 86 195.174 139.85 576 <i WITH DIAERESIS> 239 87 87 87 195.175 139.86 577 <SMALL LETTER eth> 240 140 140 140 195.176 139.87 578 <n WITH TILDE> 241 73 73 73 195.177 139.88 579 <o WITH GRAVE> 242 205 205 205 195.178 139.89 580 <o WITH ACUTE> 243 206 206 206 195.179 139.98 581 <o WITH CIRCUMFLEX> 244 203 203 203 195.180 139.99 582 <o WITH TILDE> 245 207 207 207 195.181 139.100 583 <o WITH DIAERESIS> 246 204 204 204 195.182 139.101 584 <DIVISION SIGN> 247 225 225 225 195.183 139.102 585 <o WITH STROKE> 248 112 112 112 195.184 139.103 586 <u WITH GRAVE> 249 221 221 192 195.185 139.104 ## 587 <u WITH ACUTE> 250 222 222 222 195.186 139.105 588 <u WITH CIRCUMFLEX> 251 219 219 219 195.187 139.106 589 <u WITH DIAERESIS> 252 220 220 220 195.188 139.112 590 <y WITH ACUTE> 253 141 141 141 195.189 139.113 591 <SMALL LETTER thorn> 254 142 142 142 195.190 139.114 592 <y WITH DIAERESIS> 255 223 223 223 195.191 139.115 593 594If you would rather see the above table in CCSID 0037 order rather than 595ASCII + Latin-1 order then run the table through: 596 597=over 4 598 599=item recipe 4 600 601=back 602 603 perl \ 604 -ne 'if(/.{29}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}/)'\ 605 -e '{push(@l,$_)}' \ 606 -e 'END{print map{$_->[0]}' \ 607 -e ' sort{$a->[1] <=> $b->[1]}' \ 608 -e ' map{[$_,substr($_,34,3)]}@l;}' perlebcdic.pod 609 610If you would rather see it in CCSID 1047 order then change the number 61134 in the last line to 39, like this: 612 613=over 4 614 615=item recipe 5 616 617=back 618 619 perl \ 620 -ne 'if(/.{29}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}/)'\ 621 -e '{push(@l,$_)}' \ 622 -e 'END{print map{$_->[0]}' \ 623 -e ' sort{$a->[1] <=> $b->[1]}' \ 624 -e ' map{[$_,substr($_,39,3)]}@l;}' perlebcdic.pod 625 626If you would rather see it in POSIX-BC order then change the number 62739 in the last line to 44, like this: 628 629=over 4 630 631=item recipe 6 632 633=back 634 635 perl \ 636 -ne 'if(/.{29}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}\s{2,4}\d{1,3}/)'\ 637 -e '{push(@l,$_)}' \ 638 -e 'END{print map{$_->[0]}' \ 639 -e ' sort{$a->[1] <=> $b->[1]}' \ 640 -e ' map{[$_,substr($_,44,3)]}@l;}' perlebcdic.pod 641 642 643=head1 IDENTIFYING CHARACTER CODE SETS 644 645To determine the character set you are running under from perl one 646could use the return value of ord() or chr() to test one or more 647character values. For example: 648 649 $is_ascii = "A" eq chr(65); 650 $is_ebcdic = "A" eq chr(193); 651 652Also, "\t" is a C<HORIZONTAL TABULATION> character so that: 653 654 $is_ascii = ord("\t") == 9; 655 $is_ebcdic = ord("\t") == 5; 656 657To distinguish EBCDIC code pages try looking at one or more of 658the characters that differ between them. For example: 659 660 $is_ebcdic_37 = "\n" eq chr(37); 661 $is_ebcdic_1047 = "\n" eq chr(21); 662 663Or better still choose a character that is uniquely encoded in any 664of the code sets, e.g.: 665 666 $is_ascii = ord('[') == 91; 667 $is_ebcdic_37 = ord('[') == 186; 668 $is_ebcdic_1047 = ord('[') == 173; 669 $is_ebcdic_POSIX_BC = ord('[') == 187; 670 671However, it would be unwise to write tests such as: 672 673 $is_ascii = "\r" ne chr(13); # WRONG 674 $is_ascii = "\n" ne chr(10); # ILL ADVISED 675 676Obviously the first of these will fail to distinguish most ASCII platforms 677from either a CCSID 0037, a 1047, or a POSIX-BC EBCDIC platform since "\r" eq 678chr(13) under all of those coded character sets. But note too that 679because "\n" is chr(13) and "\r" is chr(10) on the Macintosh (which is an 680ASCII platform) the second C<$is_ascii> test will lead to trouble there. 681 682To determine whether or not perl was built under an EBCDIC 683code page you can use the Config module like so: 684 685 use Config; 686 $is_ebcdic = $Config{'ebcdic'} eq 'define'; 687 688=head1 CONVERSIONS 689 690=head2 C<utf8::unicode_to_native()> and C<utf8::native_to_unicode()> 691 692These functions take an input numeric code point in one encoding and 693return what its equivalent value is in the other. 694 695=head2 tr/// 696 697In order to convert a string of characters from one character set to 698another a simple list of numbers, such as in the right columns in the 699above table, along with perl's tr/// operator is all that is needed. 700The data in the table are in ASCII/Latin1 order, hence the EBCDIC columns 701provide easy-to-use ASCII/Latin1 to EBCDIC operations that are also easily 702reversed. 703 704For example, to convert ASCII/Latin1 to code page 037 take the output of the 705second numbers column from the output of recipe 2 (modified to add '\' 706characters), and use it in tr/// like so: 707 708 $cp_037 = 709 '\x00\x01\x02\x03\x37\x2D\x2E\x2F\x16\x05\x25\x0B\x0C\x0D\x0E\x0F' . 710 '\x10\x11\x12\x13\x3C\x3D\x32\x26\x18\x19\x3F\x27\x1C\x1D\x1E\x1F' . 711 '\x40\x5A\x7F\x7B\x5B\x6C\x50\x7D\x4D\x5D\x5C\x4E\x6B\x60\x4B\x61' . 712 '\xF0\xF1\xF2\xF3\xF4\xF5\xF6\xF7\xF8\xF9\x7A\x5E\x4C\x7E\x6E\x6F' . 713 '\x7C\xC1\xC2\xC3\xC4\xC5\xC6\xC7\xC8\xC9\xD1\xD2\xD3\xD4\xD5\xD6' . 714 '\xD7\xD8\xD9\xE2\xE3\xE4\xE5\xE6\xE7\xE8\xE9\xBA\xE0\xBB\xB0\x6D' . 715 '\x79\x81\x82\x83\x84\x85\x86\x87\x88\x89\x91\x92\x93\x94\x95\x96' . 716 '\x97\x98\x99\xA2\xA3\xA4\xA5\xA6\xA7\xA8\xA9\xC0\x4F\xD0\xA1\x07' . 717 '\x20\x21\x22\x23\x24\x15\x06\x17\x28\x29\x2A\x2B\x2C\x09\x0A\x1B' . 718 '\x30\x31\x1A\x33\x34\x35\x36\x08\x38\x39\x3A\x3B\x04\x14\x3E\xFF' . 719 '\x41\xAA\x4A\xB1\x9F\xB2\x6A\xB5\xBD\xB4\x9A\x8A\x5F\xCA\xAF\xBC' . 720 '\x90\x8F\xEA\xFA\xBE\xA0\xB6\xB3\x9D\xDA\x9B\x8B\xB7\xB8\xB9\xAB' . 721 '\x64\x65\x62\x66\x63\x67\x9E\x68\x74\x71\x72\x73\x78\x75\x76\x77' . 722 '\xAC\x69\xED\xEE\xEB\xEF\xEC\xBF\x80\xFD\xFE\xFB\xFC\xAD\xAE\x59' . 723 '\x44\x45\x42\x46\x43\x47\x9C\x48\x54\x51\x52\x53\x58\x55\x56\x57' . 724 '\x8C\x49\xCD\xCE\xCB\xCF\xCC\xE1\x70\xDD\xDE\xDB\xDC\x8D\x8E\xDF'; 725 726 my $ebcdic_string = $ascii_string; 727 eval '$ebcdic_string =~ tr/\000-\377/' . $cp_037 . '/'; 728 729To convert from EBCDIC 037 to ASCII just reverse the order of the tr/// 730arguments like so: 731 732 my $ascii_string = $ebcdic_string; 733 eval '$ascii_string =~ tr/' . $cp_037 . '/\000-\377/'; 734 735Similarly one could take the output of the third numbers column from recipe 2 736to obtain a C<$cp_1047> table. The fourth numbers column of the output from 737recipe 2 could provide a C<$cp_posix_bc> table suitable for transcoding as 738well. 739 740If you wanted to see the inverse tables, you would first have to sort on the 741desired numbers column as in recipes 4, 5 or 6, then take the output of the 742first numbers column. 743 744=head2 iconv 745 746XPG operability often implies the presence of an I<iconv> utility 747available from the shell or from the C library. Consult your system's 748documentation for information on iconv. 749 750On OS/390 or z/OS see the iconv(1) manpage. One way to invoke the iconv 751shell utility from within perl would be to: 752 753 # OS/390 or z/OS example 754 $ascii_data = `echo '$ebcdic_data'| iconv -f IBM-1047 -t ISO8859-1` 755 756or the inverse map: 757 758 # OS/390 or z/OS example 759 $ebcdic_data = `echo '$ascii_data'| iconv -f ISO8859-1 -t IBM-1047` 760 761For other perl-based conversion options see the Convert::* modules on CPAN. 762 763=head2 C RTL 764 765The OS/390 and z/OS C run-time libraries provide _atoe() and _etoa() functions. 766 767=head1 OPERATOR DIFFERENCES 768 769The C<..> range operator treats certain character ranges with 770care on EBCDIC platforms. For example the following array 771will have twenty six elements on either an EBCDIC platform 772or an ASCII platform: 773 774 @alphabet = ('A'..'Z'); # $#alphabet == 25 775 776The bitwise operators such as & ^ | may return different results 777when operating on string or character data in a perl program running 778on an EBCDIC platform than when run on an ASCII platform. Here is 779an example adapted from the one in L<perlop>: 780 781 # EBCDIC-based examples 782 print "j p \n" ^ " a h"; # prints "JAPH\n" 783 print "JA" | " ph\n"; # prints "japh\n" 784 print "JAPH\nJunk" & "\277\277\277\277\277"; # prints "japh\n"; 785 print 'p N$' ^ " E<H\n"; # prints "Perl\n"; 786 787An interesting property of the 32 C0 control characters 788in the ASCII table is that they can "literally" be constructed 789as control characters in perl, e.g. C<(chr(0)> eq C<\c@>)> 790C<(chr(1)> eq C<\cA>)>, and so on. Perl on EBCDIC platforms has been 791ported to take C<\c@> to chr(0) and C<\cA> to chr(1), etc. as well, but the 792characters that result depend on which code page you are 793using. The table below uses the standard acronyms for the controls. 794The POSIX-BC and 1047 sets are 795identical throughout this range and differ from the 0037 set at only 796one spot (21 decimal). Note that the C<LINE FEED> character 797may be generated by C<\cJ> on ASCII platforms but by C<\cU> on 1047 or POSIX-BC 798platforms and cannot be generated as a C<"\c.letter."> control character on 7990037 platforms. Note also that C<\c\> cannot be the final element in a string 800or regex, as it will absorb the terminator. But C<\c\I<X>> is a C<FILE 801SEPARATOR> concatenated with I<X> for all I<X>. 802The outlier C<\c?> on ASCII, which yields a non-C0 control C<DEL>, 803yields the outlier control C<APC> on EBCDIC, the one that isn't in the 804block of contiguous controls. 805 806 chr ord 8859-1 0037 1047 && POSIX-BC 807 ----------------------------------------------------------------------- 808 \c@ 0 <NUL> <NUL> <NUL> 809 \cA 1 <SOH> <SOH> <SOH> 810 \cB 2 <STX> <STX> <STX> 811 \cC 3 <ETX> <ETX> <ETX> 812 \cD 4 <EOT> <ST> <ST> 813 \cE 5 <ENQ> <HT> <HT> 814 \cF 6 <ACK> <SSA> <SSA> 815 \cG 7 <BEL> <DEL> <DEL> 816 \cH 8 <BS> <EPA> <EPA> 817 \cI 9 <HT> <RI> <RI> 818 \cJ 10 <LF> <SS2> <SS2> 819 \cK 11 <VT> <VT> <VT> 820 \cL 12 <FF> <FF> <FF> 821 \cM 13 <CR> <CR> <CR> 822 \cN 14 <SO> <SO> <SO> 823 \cO 15 <SI> <SI> <SI> 824 \cP 16 <DLE> <DLE> <DLE> 825 \cQ 17 <DC1> <DC1> <DC1> 826 \cR 18 <DC2> <DC2> <DC2> 827 \cS 19 <DC3> <DC3> <DC3> 828 \cT 20 <DC4> <OSC> <OSC> 829 \cU 21 <NAK> <NEL> <LF> ** 830 \cV 22 <SYN> <BS> <BS> 831 \cW 23 <ETB> <ESA> <ESA> 832 \cX 24 <CAN> <CAN> <CAN> 833 \cY 25 <EOM> <EOM> <EOM> 834 \cZ 26 <SUB> <PU2> <PU2> 835 \c[ 27 <ESC> <SS3> <SS3> 836 \c\X 28 <FS>X <FS>X <FS>X 837 \c] 29 <GS> <GS> <GS> 838 \c^ 30 <RS> <RS> <RS> 839 \c_ 31 <US> <US> <US> 840 \c? * <DEL> <APC> <APC> 841 842C<*> Note: C<\c?> maps to ordinal 127 (C<DEL>) on ASCII platforms, but 843since ordinal 127 is a not a control character on EBCDIC machines, 844C<\c?> instead maps to C<APC>, which is 255 in 0037 and 1047, and 95 in 845POSIX-BC. 846 847=head1 FUNCTION DIFFERENCES 848 849=over 8 850 851=item chr() 852 853chr() must be given an EBCDIC code number argument to yield a desired 854character return value on an EBCDIC platform. For example: 855 856 $CAPITAL_LETTER_A = chr(193); 857 858=item ord() 859 860ord() will return EBCDIC code number values on an EBCDIC platform. 861For example: 862 863 $the_number_193 = ord("A"); 864 865=item pack() 866 867The c and C templates for pack() are dependent upon character set 868encoding. Examples of usage on EBCDIC include: 869 870 $foo = pack("CCCC",193,194,195,196); 871 # $foo eq "ABCD" 872 $foo = pack("C4",193,194,195,196); 873 # same thing 874 875 $foo = pack("ccxxcc",193,194,195,196); 876 # $foo eq "AB\0\0CD" 877 878=item print() 879 880One must be careful with scalars and strings that are passed to 881print that contain ASCII encodings. One common place 882for this to occur is in the output of the MIME type header for 883CGI script writing. For example, many perl programming guides 884recommend something similar to: 885 886 print "Content-type:\ttext/html\015\012\015\012"; 887 # this may be wrong on EBCDIC 888 889Under the IBM OS/390 USS Web Server or WebSphere on z/OS for example 890you should instead write that as: 891 892 print "Content-type:\ttext/html\r\n\r\n"; # OK for DGW et al 893 894That is because the translation from EBCDIC to ASCII is done 895by the web server in this case (such code will not be appropriate for 896the Macintosh however). Consult your web server's documentation for 897further details. 898 899=item printf() 900 901The formats that can convert characters to numbers and vice versa 902will be different from their ASCII counterparts when executed 903on an EBCDIC platform. Examples include: 904 905 printf("%c%c%c",193,194,195); # prints ABC 906 907=item sort() 908 909EBCDIC sort results may differ from ASCII sort results especially for 910mixed case strings. This is discussed in more detail below. 911 912=item sprintf() 913 914See the discussion of printf() above. An example of the use 915of sprintf would be: 916 917 $CAPITAL_LETTER_A = sprintf("%c",193); 918 919=item unpack() 920 921See the discussion of pack() above. 922 923=back 924 925=head1 REGULAR EXPRESSION DIFFERENCES 926 927As of perl 5.005_03 the letter range regular expressions such as 928[A-Z] and [a-z] have been especially coded to not pick up gap 929characters. For example, characters such as E<ocirc> C<o WITH CIRCUMFLEX> 930that lie between I and J would not be matched by the 931regular expression range C</[H-K]/>. This works in 932the other direction, too, if either of the range end points is 933explicitly numeric: C<[\x89-\x91]> will match C<\x8e>, even 934though C<\x89> is C<i> and C<\x91 > is C<j>, and C<\x8e> 935is a gap character from the alphabetic viewpoint. 936 937If you do want to match the alphabet gap characters in a single octet 938regular expression try matching the hex or octal code such 939as C</\313/> on EBCDIC or C</\364/> on ASCII platforms to 940have your regular expression match C<o WITH CIRCUMFLEX>. 941 942Another construct to be wary of is the inappropriate use of hex or 943octal constants in regular expressions. Consider the following 944set of subs: 945 946 sub is_c0 { 947 my $char = substr(shift,0,1); 948 $char =~ /[\000-\037]/; 949 } 950 951 sub is_print_ascii { 952 my $char = substr(shift,0,1); 953 $char =~ /[\040-\176]/; 954 } 955 956 sub is_delete { 957 my $char = substr(shift,0,1); 958 $char eq "\177"; 959 } 960 961 sub is_c1 { 962 my $char = substr(shift,0,1); 963 $char =~ /[\200-\237]/; 964 } 965 966 sub is_latin_1 { # But not ASCII; not C1 967 my $char = substr(shift,0,1); 968 $char =~ /[\240-\377]/; 969 } 970 971These are valid only on ASCII platforms, but can be easily rewritten to 972work on any platform as follows: 973 974 sub Is_c0 { 975 my $char = substr(shift,0,1); 976 return $char =~ /[[:cntrl:]]/ 977 && $char =~ /[[:ascii:]]/ 978 && ! Is_delete($char); 979 } 980 981 sub Is_print_ascii { 982 my $char = substr(shift,0,1); 983 984 return $char =~ /[[:print:]]/ && $char =~ /[[:ascii:]]/; 985 986 # Alternatively: 987 # return $char 988 # =~ /[ !"\#\$%&'()*+,\-.\/0-9:;<=>?\@A-Z[\\\]^_`a-z{|}~]/; 989 } 990 991 sub Is_delete { 992 my $char = substr(shift,0,1); 993 return utf8::native_to_unicode(ord $char) == 0x7F; 994 } 995 996 sub Is_c1 { 997 use feature 'unicode_strings'; 998 my $char = substr(shift,0,1); 999 return $char =~ /[[:cntrl:]]/ && $char !~ /[[:ascii:]]/; 1000 } 1001 1002 sub Is_latin_1 { # But not ASCII; not C1 1003 use feature 'unicode_strings'; 1004 my $char = substr(shift,0,1); 1005 return ord($char) < 256 1006 && $char !~ [[:ascii:]] 1007 && $char !~ [[:cntrl:]]; 1008 } 1009 1010Another way to write C<Is_latin_1()> would be 1011to use the characters in the range explicitly: 1012 1013 sub Is_latin_1 { 1014 my $char = substr(shift,0,1); 1015 $char =~ /[ ¡¢£¤¥¦§¨©ª«¬®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ]/; 1016 } 1017 1018Although that form may run into trouble in network transit (due to the 1019presence of 8 bit characters) or on non ISO-Latin character sets. 1020 1021=head1 SOCKETS 1022 1023Most socket programming assumes ASCII character encodings in network 1024byte order. Exceptions can include CGI script writing under a 1025host web server where the server may take care of translation for you. 1026Most host web servers convert EBCDIC data to ISO-8859-1 or Unicode on 1027output. 1028 1029=head1 SORTING 1030 1031One big difference between ASCII-based character sets and EBCDIC ones 1032are the relative positions of upper and lower case letters and the 1033letters compared to the digits. If sorted on an ASCII-based platform the 1034two-letter abbreviation for a physician comes before the two letter 1035abbreviation for drive; that is: 1036 1037 @sorted = sort(qw(Dr. dr.)); # @sorted holds ('Dr.','dr.') on ASCII, 1038 # but ('dr.','Dr.') on EBCDIC 1039 1040The property of lowercase before uppercase letters in EBCDIC is 1041even carried to the Latin 1 EBCDIC pages such as 0037 and 1047. 1042An example would be that E<Euml> C<E WITH DIAERESIS> (203) comes 1043before E<euml> C<e WITH DIAERESIS> (235) on an ASCII platform, but 1044the latter (83) comes before the former (115) on an EBCDIC platform. 1045(Astute readers will note that the uppercase version of E<szlig> 1046C<SMALL LETTER SHARP S> is simply "SS" and that the upper case version of 1047E<yuml> C<y WITH DIAERESIS> is not in the 0..255 range but it is 1048at U+x0178 in Unicode, or C<"\x{178}"> in a Unicode enabled Perl). 1049 1050The sort order will cause differences between results obtained on 1051ASCII platforms versus EBCDIC platforms. What follows are some suggestions 1052on how to deal with these differences. 1053 1054=head2 Ignore ASCII vs. EBCDIC sort differences. 1055 1056This is the least computationally expensive strategy. It may require 1057some user education. 1058 1059=head2 MONO CASE then sort data. 1060 1061In order to minimize the expense of mono casing mixed-case text, try to 1062C<tr///> towards the character set case most employed within the data. 1063If the data are primarily UPPERCASE non Latin 1 then apply tr/[a-z]/[A-Z]/ 1064then sort(). If the data are primarily lowercase non Latin 1 then 1065apply tr/[A-Z]/[a-z]/ before sorting. If the data are primarily UPPERCASE 1066and include Latin-1 characters then apply: 1067 1068 tr/[a-z]/[A-Z]/; 1069 tr/[àáâãäåæçèéêëìíîïðñòóôõöøùúûüýþ]/[ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞ/; 1070 s/ß/SS/g; 1071 1072then sort(). Do note however that such Latin-1 manipulation does not 1073address the E<yuml> C<y WITH DIAERESIS> character that will remain at 1074code point 255 on ASCII platforms, but 223 on most EBCDIC platforms 1075where it will sort to a place less than the EBCDIC numerals. With a 1076Unicode-enabled Perl you might try: 1077 1078 tr/^?/\x{178}/; 1079 1080The strategy of mono casing data before sorting does not preserve the case 1081of the data and may not be acceptable for that reason. 1082 1083=head2 Convert, sort data, then re convert. 1084 1085This is the most expensive proposition that does not employ a network 1086connection. 1087 1088=head2 Perform sorting on one type of platform only. 1089 1090This strategy can employ a network connection. As such 1091it would be computationally expensive. 1092 1093=head1 TRANSFORMATION FORMATS 1094 1095There are a variety of ways of transforming data with an intra character set 1096mapping that serve a variety of purposes. Sorting was discussed in the 1097previous section and a few of the other more popular mapping techniques are 1098discussed next. 1099 1100=head2 URL decoding and encoding 1101 1102Note that some URLs have hexadecimal ASCII code points in them in an 1103attempt to overcome character or protocol limitation issues. For example 1104the tilde character is not on every keyboard hence a URL of the form: 1105 1106 http://www.pvhp.com/~pvhp/ 1107 1108may also be expressed as either of: 1109 1110 http://www.pvhp.com/%7Epvhp/ 1111 1112 http://www.pvhp.com/%7epvhp/ 1113 1114where 7E is the hexadecimal ASCII code point for '~'. Here is an example 1115of decoding such a URL under CCSID 1047: 1116 1117 $url = 'http://www.pvhp.com/%7Epvhp/'; 1118 # this array assumes code page 1047 1119 my @a2e_1047 = ( 1120 0, 1, 2, 3, 55, 45, 46, 47, 22, 5, 21, 11, 12, 13, 14, 15, 1121 16, 17, 18, 19, 60, 61, 50, 38, 24, 25, 63, 39, 28, 29, 30, 31, 1122 64, 90,127,123, 91,108, 80,125, 77, 93, 92, 78,107, 96, 75, 97, 1123 240,241,242,243,244,245,246,247,248,249,122, 94, 76,126,110,111, 1124 124,193,194,195,196,197,198,199,200,201,209,210,211,212,213,214, 1125 215,216,217,226,227,228,229,230,231,232,233,173,224,189, 95,109, 1126 121,129,130,131,132,133,134,135,136,137,145,146,147,148,149,150, 1127 151,152,153,162,163,164,165,166,167,168,169,192, 79,208,161, 7, 1128 32, 33, 34, 35, 36, 37, 6, 23, 40, 41, 42, 43, 44, 9, 10, 27, 1129 48, 49, 26, 51, 52, 53, 54, 8, 56, 57, 58, 59, 4, 20, 62,255, 1130 65,170, 74,177,159,178,106,181,187,180,154,138,176,202,175,188, 1131 144,143,234,250,190,160,182,179,157,218,155,139,183,184,185,171, 1132 100,101, 98,102, 99,103,158,104,116,113,114,115,120,117,118,119, 1133 172,105,237,238,235,239,236,191,128,253,254,251,252,186,174, 89, 1134 68, 69, 66, 70, 67, 71,156, 72, 84, 81, 82, 83, 88, 85, 86, 87, 1135 140, 73,205,206,203,207,204,225,112,221,222,219,220,141,142,223 1136 ); 1137 $url =~ s/%([0-9a-fA-F]{2})/pack("c",$a2e_1047[hex($1)])/ge; 1138 1139Conversely, here is a partial solution for the task of encoding such 1140a URL under the 1047 code page: 1141 1142 $url = 'http://www.pvhp.com/~pvhp/'; 1143 # this array assumes code page 1047 1144 my @e2a_1047 = ( 1145 0, 1, 2, 3,156, 9,134,127,151,141,142, 11, 12, 13, 14, 15, 1146 16, 17, 18, 19,157, 10, 8,135, 24, 25,146,143, 28, 29, 30, 31, 1147 128,129,130,131,132,133, 23, 27,136,137,138,139,140, 5, 6, 7, 1148 144,145, 22,147,148,149,150, 4,152,153,154,155, 20, 21,158, 26, 1149 32,160,226,228,224,225,227,229,231,241,162, 46, 60, 40, 43,124, 1150 38,233,234,235,232,237,238,239,236,223, 33, 36, 42, 41, 59, 94, 1151 45, 47,194,196,192,193,195,197,199,209,166, 44, 37, 95, 62, 63, 1152 248,201,202,203,200,205,206,207,204, 96, 58, 35, 64, 39, 61, 34, 1153 216, 97, 98, 99,100,101,102,103,104,105,171,187,240,253,254,177, 1154 176,106,107,108,109,110,111,112,113,114,170,186,230,184,198,164, 1155 181,126,115,116,117,118,119,120,121,122,161,191,208, 91,222,174, 1156 172,163,165,183,169,167,182,188,189,190,221,168,175, 93,180,215, 1157 123, 65, 66, 67, 68, 69, 70, 71, 72, 73,173,244,246,242,243,245, 1158 125, 74, 75, 76, 77, 78, 79, 80, 81, 82,185,251,252,249,250,255, 1159 92,247, 83, 84, 85, 86, 87, 88, 89, 90,178,212,214,210,211,213, 1160 48, 49, 50, 51, 52, 53, 54, 55, 56, 57,179,219,220,217,218,159 1161 ); 1162 # The following regular expression does not address the 1163 # mappings for: ('.' => '%2E', '/' => '%2F', ':' => '%3A') 1164 $url =~ s/([\t "#%&\(\),;<=>\?\@\[\\\]^`{|}~])/ 1165 sprintf("%%%02X",$e2a_1047[ord($1)])/xge; 1166 1167where a more complete solution would split the URL into components 1168and apply a full s/// substitution only to the appropriate parts. 1169 1170In the remaining examples a @e2a or @a2e array may be employed 1171but the assignment will not be shown explicitly. For code page 1047 1172you could use the @a2e_1047 or @e2a_1047 arrays just shown. 1173 1174=head2 uu encoding and decoding 1175 1176The C<u> template to pack() or unpack() will render EBCDIC data in EBCDIC 1177characters equivalent to their ASCII counterparts. For example, the 1178following will print "Yes indeed\n" on either an ASCII or EBCDIC computer: 1179 1180 $all_byte_chrs = ''; 1181 for (0..255) { $all_byte_chrs .= chr($_); } 1182 $uuencode_byte_chrs = pack('u', $all_byte_chrs); 1183 ($uu = <<'ENDOFHEREDOC') =~ s/^\s*//gm; 1184 M``$"`P0%!@<("0H+#`T.#Q`1$A,4%187&!D:&QP='A\@(2(C)"4F)R@I*BLL 1185 M+2XO,#$R,S0U-C<X.3H[/#T^/T!!0D-$149'2$E*2TQ-3D]045)35%565UA9 1186 M6EM<75Y?8&%B8V1E9F=H:6IK;&UN;W!Q<G-T=79W>'EZ>WQ]?G^`@8*#A(6& 1187 MAXB)BHN,C8Z/D)&2DY25EI>8F9J;G)V>GZ"AHJ.DI::GJ*FJJZRMKJ^PL;*S 1188 MM+6VM[BYNKN\O;Z_P,'"P\3%QL?(R<K+S,W.S]#1TM/4U=;7V-G:V]S=WM_@ 1189 ?X>+CY.7FY^CIZNOL[>[O\/'R\_3U]O?X^?K[_/W^_P`` 1190 ENDOFHEREDOC 1191 if ($uuencode_byte_chrs eq $uu) { 1192 print "Yes "; 1193 } 1194 $uudecode_byte_chrs = unpack('u', $uuencode_byte_chrs); 1195 if ($uudecode_byte_chrs eq $all_byte_chrs) { 1196 print "indeed\n"; 1197 } 1198 1199Here is a very spartan uudecoder that will work on EBCDIC provided 1200that the @e2a array is filled in appropriately: 1201 1202 #!/usr/local/bin/perl 1203 @e2a = ( # this must be filled in 1204 ); 1205 $_ = <> until ($mode,$file) = /^begin\s*(\d*)\s*(\S*)/; 1206 open(OUT, "> $file") if $file ne ""; 1207 while(<>) { 1208 last if /^end/; 1209 next if /[a-z]/; 1210 next unless int(((($e2a[ord()] - 32 ) & 077) + 2) / 3) == 1211 int(length() / 4); 1212 print OUT unpack("u", $_); 1213 } 1214 close(OUT); 1215 chmod oct($mode), $file; 1216 1217 1218=head2 Quoted-Printable encoding and decoding 1219 1220On ASCII-encoded platforms it is possible to strip characters outside of 1221the printable set using: 1222 1223 # This QP encoder works on ASCII only 1224 $qp_string =~ s/([=\x00-\x1F\x80-\xFF])/sprintf("=%02X",ord($1))/ge; 1225 1226Whereas a QP encoder that works on both ASCII and EBCDIC platforms 1227would look somewhat like the following (where the EBCDIC branch @e2a 1228array is omitted for brevity): 1229 1230 if (ord('A') == 65) { # ASCII 1231 $delete = "\x7F"; # ASCII 1232 @e2a = (0 .. 255) # ASCII to ASCII identity map 1233 } 1234 else { # EBCDIC 1235 $delete = "\x07"; # EBCDIC 1236 @e2a = # EBCDIC to ASCII map (as shown above) 1237 } 1238 $qp_string =~ 1239 s/([^ !"\#\$%&'()*+,\-.\/0-9:;<>?\@A-Z[\\\]^_`a-z{|}~$delete])/ 1240 sprintf("=%02X",$e2a[ord($1)])/xge; 1241 1242(although in production code the substitutions might be done 1243in the EBCDIC branch with the @e2a array and separately in the 1244ASCII branch without the expense of the identity map). 1245 1246Such QP strings can be decoded with: 1247 1248 # This QP decoder is limited to ASCII only 1249 $string =~ s/=([0-9A-Fa-f][0-9A-Fa-f])/chr hex $1/ge; 1250 $string =~ s/=[\n\r]+$//; 1251 1252Whereas a QP decoder that works on both ASCII and EBCDIC platforms 1253would look somewhat like the following (where the @a2e array is 1254omitted for brevity): 1255 1256 $string =~ s/=([0-9A-Fa-f][0-9A-Fa-f])/chr $a2e[hex $1]/ge; 1257 $string =~ s/=[\n\r]+$//; 1258 1259=head2 Caesarean ciphers 1260 1261The practice of shifting an alphabet one or more characters for encipherment 1262dates back thousands of years and was explicitly detailed by Gaius Julius 1263Caesar in his B<Gallic Wars> text. A single alphabet shift is sometimes 1264referred to as a rotation and the shift amount is given as a number $n after 1265the string 'rot' or "rot$n". Rot0 and rot26 would designate identity maps 1266on the 26-letter English version of the Latin alphabet. Rot13 has the 1267interesting property that alternate subsequent invocations are identity maps 1268(thus rot13 is its own non-trivial inverse in the group of 26 alphabet 1269rotations). Hence the following is a rot13 encoder and decoder that will 1270work on ASCII and EBCDIC platforms: 1271 1272 #!/usr/local/bin/perl 1273 1274 while(<>){ 1275 tr/n-za-mN-ZA-M/a-zA-Z/; 1276 print; 1277 } 1278 1279In one-liner form: 1280 1281 perl -ne 'tr/n-za-mN-ZA-M/a-zA-Z/;print' 1282 1283 1284=head1 Hashing order and checksums 1285 1286To the extent that it is possible to write code that depends on 1287hashing order there may be differences between hashes as stored 1288on an ASCII-based platform and hashes stored on an EBCDIC-based platform. 1289XXX 1290 1291=head1 I18N AND L10N 1292 1293Internationalization (I18N) and localization (L10N) are supported at least 1294in principle even on EBCDIC platforms. The details are system-dependent 1295and discussed under the L<perlebcdic/OS ISSUES> section below. 1296 1297=head1 MULTI-OCTET CHARACTER SETS 1298 1299Perl may work with an internal UTF-EBCDIC encoding form for wide characters 1300on EBCDIC platforms in a manner analogous to the way that it works with 1301the UTF-8 internal encoding form on ASCII based platforms. 1302 1303Legacy multi byte EBCDIC code pages XXX. 1304 1305=head1 OS ISSUES 1306 1307There may be a few system-dependent issues 1308of concern to EBCDIC Perl programmers. 1309 1310=head2 OS/400 1311 1312=over 8 1313 1314=item PASE 1315 1316The PASE environment is a runtime environment for OS/400 that can run 1317executables built for PowerPC AIX in OS/400; see L<perlos400>. PASE 1318is ASCII-based, not EBCDIC-based as the ILE. 1319 1320=item IFS access 1321 1322XXX. 1323 1324=back 1325 1326=head2 OS/390, z/OS 1327 1328Perl runs under Unix Systems Services or USS. 1329 1330=over 8 1331 1332=item chcp 1333 1334B<chcp> is supported as a shell utility for displaying and changing 1335one's code page. See also L<chcp(1)>. 1336 1337=item dataset access 1338 1339For sequential data set access try: 1340 1341 my @ds_records = `cat //DSNAME`; 1342 1343or: 1344 1345 my @ds_records = `cat //'HLQ.DSNAME'`; 1346 1347See also the OS390::Stdio module on CPAN. 1348 1349=item OS/390, z/OS iconv 1350 1351B<iconv> is supported as both a shell utility and a C RTL routine. 1352See also the iconv(1) and iconv(3) manual pages. 1353 1354=item locales 1355 1356On OS/390 or z/OS see L<locale> for information on locales. The L10N files 1357are in F</usr/nls/locale>. $Config{d_setlocale} is 'define' on OS/390 1358or z/OS. 1359 1360=back 1361 1362=head2 POSIX-BC? 1363 1364XXX. 1365 1366=head1 BUGS 1367 1368This pod document contains literal Latin 1 characters and may encounter 1369translation difficulties. In particular one popular nroff implementation 1370was known to strip accented characters to their unaccented counterparts 1371while attempting to view this document through the B<pod2man> program 1372(for example, you may see a plain C<y> rather than one with a diaeresis 1373as in E<yuml>). Another nroff truncated the resultant manpage at 1374the first occurrence of 8 bit characters. 1375 1376Not all shells will allow multiple C<-e> string arguments to perl to 1377be concatenated together properly as recipes 0, 2, 4, 5, and 6 might 1378seem to imply. 1379 1380=head1 SEE ALSO 1381 1382L<perllocale>, L<perlfunc>, L<perlunicode>, L<utf8>. 1383 1384=head1 REFERENCES 1385 1386L<http://anubis.dkuug.dk/i18n/charmaps> 1387 1388L<http://www.unicode.org/> 1389 1390L<http://www.unicode.org/unicode/reports/tr16/> 1391 1392L<http://www.wps.com/projects/codes/> 1393B<ASCII: American Standard Code for Information Infiltration> Tom Jennings, 1394September 1999. 1395 1396B<The Unicode Standard, Version 3.0> The Unicode Consortium, Lisa Moore ed., 1397ISBN 0-201-61633-5, Addison Wesley Developers Press, February 2000. 1398 1399B<CDRA: IBM - Character Data Representation Architecture - 1400Reference and Registry>, IBM SC09-2190-00, December 1996. 1401 1402"Demystifying Character Sets", Andrea Vine, Multilingual Computing 1403& Technology, B<#26 Vol. 10 Issue 4>, August/September 1999; 1404ISSN 1523-0309; Multilingual Computing Inc. Sandpoint ID, USA. 1405 1406B<Codes, Ciphers, and Other Cryptic and Clandestine Communication> 1407Fred B. Wrixon, ISBN 1-57912-040-7, Black Dog & Leventhal Publishers, 14081998. 1409 1410L<http://www.bobbemer.com/P-BIT.HTM> 1411B<IBM - EBCDIC and the P-bit; The biggest Computer Goof Ever> Robert Bemer. 1412 1413=head1 HISTORY 1414 141515 April 2001: added UTF-8 and UTF-EBCDIC to main table, pvhp. 1416 1417=head1 AUTHOR 1418 1419Peter Prymmer pvhp@best.com wrote this in 1999 and 2000 1420with CCSID 0819 and 0037 help from Chris Leach and 1421AndrE<eacute> Pirard A.Pirard@ulg.ac.be as well as POSIX-BC 1422help from Thomas Dorner Thomas.Dorner@start.de. 1423Thanks also to Vickie Cooper, Philip Newton, William Raffloer, and 1424Joe Smith. Trademarks, registered trademarks, service marks and 1425registered service marks used in this document are the property of 1426their respective owners. 1427