Commit | Line | Data |
a0d0e21e |
1 | =head1 NAME |
2 | |
3 | perltrap - Perl traps for the unwary |
4 | |
5 | =head1 DESCRIPTION |
6 | |
cb1a09d0 |
7 | The biggest trap of all is forgetting to use the B<-w> switch; see |
8 | L<perlrun>. The second biggest trap is not making your entire program |
daff0e37 |
9 | runnable under C<use strict>. The third biggest trap is not reading |
10 | the list of changes in this version of Perl; see L<perldelta>. |
a0d0e21e |
11 | |
12 | =head2 Awk Traps |
13 | |
14 | Accustomed B<awk> users should take special note of the following: |
15 | |
16 | =over 4 |
17 | |
18 | =item * |
19 | |
20 | The English module, loaded via |
21 | |
22 | use English; |
23 | |
54310121 |
24 | allows you to refer to special variables (like C<$/>) with names (like |
19799a22 |
25 | $RS), as though they were in B<awk>; see L<perlvar> for details. |
a0d0e21e |
26 | |
27 | =item * |
28 | |
29 | Semicolons are required after all simple statements in Perl (except |
30 | at the end of a block). Newline is not a statement delimiter. |
31 | |
32 | =item * |
33 | |
34 | Curly brackets are required on C<if>s and C<while>s. |
35 | |
36 | =item * |
37 | |
5db417f7 |
38 | Variables begin with "$", "@" or "%" in Perl. |
a0d0e21e |
39 | |
40 | =item * |
41 | |
42 | Arrays index from 0. Likewise string positions in substr() and |
43 | index(). |
44 | |
45 | =item * |
46 | |
47 | You have to decide whether your array has numeric or string indices. |
48 | |
49 | =item * |
50 | |
aa689395 |
51 | Hash values do not spring into existence upon mere reference. |
a0d0e21e |
52 | |
53 | =item * |
54 | |
55 | You have to decide whether you want to use string or numeric |
56 | comparisons. |
57 | |
58 | =item * |
59 | |
60 | Reading an input line does not split it for you. You get to split it |
54310121 |
61 | to an array yourself. And the split() operator has different |
62 | arguments than B<awk>'s. |
a0d0e21e |
63 | |
64 | =item * |
65 | |
66 | The current input line is normally in $_, not $0. It generally does |
67 | not have the newline stripped. ($0 is the name of the program |
68 | executed.) See L<perlvar>. |
69 | |
70 | =item * |
71 | |
c47ff5f1 |
72 | $<I<digit>> does not refer to fields--it refers to substrings matched |
8b0a4b75 |
73 | by the last match pattern. |
a0d0e21e |
74 | |
75 | =item * |
76 | |
77 | The print() statement does not add field and record separators unless |
8b0a4b75 |
78 | you set C<$,> and C<$\>. You can set $OFS and $ORS if you're using |
a0d0e21e |
79 | the English module. |
80 | |
81 | =item * |
82 | |
83 | You must open your files before you print to them. |
84 | |
85 | =item * |
86 | |
87 | The range operator is "..", not comma. The comma operator works as in |
88 | C. |
89 | |
90 | =item * |
91 | |
92 | The match operator is "=~", not "~". ("~" is the one's complement |
93 | operator, as in C.) |
94 | |
95 | =item * |
96 | |
97 | The exponentiation operator is "**", not "^". "^" is the XOR |
98 | operator, as in C. (You know, one could get the feeling that B<awk> is |
99 | basically incompatible with C.) |
100 | |
101 | =item * |
102 | |
103 | The concatenation operator is ".", not the null string. (Using the |
5f05dabc |
104 | null string would render C</pat/ /pat/> unparsable, because the third slash |
105 | would be interpreted as a division operator--the tokenizer is in fact |
c47ff5f1 |
106 | slightly context sensitive for operators like "/", "?", and ">". |
a0d0e21e |
107 | And in fact, "." itself can be the beginning of a number.) |
108 | |
109 | =item * |
110 | |
111 | The C<next>, C<exit>, and C<continue> keywords work differently. |
112 | |
113 | =item * |
114 | |
115 | |
116 | The following variables work differently: |
117 | |
118 | Awk Perl |
119 | ARGC $#ARGV or scalar @ARGV |
120 | ARGV[0] $0 |
121 | FILENAME $ARGV |
122 | FNR $. - something |
123 | FS (whatever you like) |
124 | NF $#Fld, or some such |
125 | NR $. |
126 | OFMT $# |
127 | OFS $, |
128 | ORS $\ |
129 | RLENGTH length($&) |
130 | RS $/ |
131 | RSTART length($`) |
132 | SUBSEP $; |
133 | |
134 | =item * |
135 | |
136 | You cannot set $RS to a pattern, only a string. |
137 | |
138 | =item * |
139 | |
140 | When in doubt, run the B<awk> construct through B<a2p> and see what it |
141 | gives you. |
142 | |
143 | =back |
144 | |
145 | =head2 C Traps |
146 | |
147 | Cerebral C programmers should take note of the following: |
148 | |
149 | =over 4 |
150 | |
151 | =item * |
152 | |
153 | Curly brackets are required on C<if>'s and C<while>'s. |
154 | |
155 | =item * |
156 | |
157 | You must use C<elsif> rather than C<else if>. |
158 | |
159 | =item * |
160 | |
54310121 |
161 | The C<break> and C<continue> keywords from C become in |
a0d0e21e |
162 | Perl C<last> and C<next>, respectively. |
19799a22 |
163 | Unlike in C, these do I<not> work within a C<do { } while> construct. |
a0d0e21e |
164 | |
165 | =item * |
166 | |
167 | There's no switch statement. (But it's easy to build one on the fly.) |
168 | |
169 | =item * |
170 | |
5db417f7 |
171 | Variables begin with "$", "@" or "%" in Perl. |
a0d0e21e |
172 | |
173 | =item * |
174 | |
a0d0e21e |
175 | Comments begin with "#", not "/*". |
176 | |
177 | =item * |
178 | |
179 | You can't take the address of anything, although a similar operator |
5f05dabc |
180 | in Perl is the backslash, which creates a reference. |
a0d0e21e |
181 | |
182 | =item * |
183 | |
4633a7c4 |
184 | C<ARGV> must be capitalized. C<$ARGV[0]> is C's C<argv[1]>, and C<argv[0]> |
185 | ends up in C<$0>. |
a0d0e21e |
186 | |
187 | =item * |
188 | |
189 | System calls such as link(), unlink(), rename(), etc. return nonzero for |
190 | success, not 0. |
191 | |
192 | =item * |
193 | |
194 | Signal handlers deal with signal names, not numbers. Use C<kill -l> |
195 | to find their names on your system. |
196 | |
197 | =back |
198 | |
199 | =head2 Sed Traps |
200 | |
201 | Seasoned B<sed> programmers should take note of the following: |
202 | |
203 | =over 4 |
204 | |
205 | =item * |
206 | |
207 | Backreferences in substitutions use "$" rather than "\". |
208 | |
209 | =item * |
210 | |
211 | The pattern matching metacharacters "(", ")", and "|" do not have backslashes |
212 | in front. |
213 | |
214 | =item * |
215 | |
216 | The range operator is C<...>, rather than comma. |
217 | |
218 | =back |
219 | |
220 | =head2 Shell Traps |
221 | |
222 | Sharp shell programmers should take note of the following: |
223 | |
224 | =over 4 |
225 | |
226 | =item * |
227 | |
54310121 |
228 | The backtick operator does variable interpolation without regard to |
a0d0e21e |
229 | the presence of single quotes in the command. |
230 | |
231 | =item * |
232 | |
54310121 |
233 | The backtick operator does no translation of the return value, unlike B<csh>. |
a0d0e21e |
234 | |
235 | =item * |
236 | |
237 | Shells (especially B<csh>) do several levels of substitution on each |
5f05dabc |
238 | command line. Perl does substitution in only certain constructs |
54310121 |
239 | such as double quotes, backticks, angle brackets, and search patterns. |
a0d0e21e |
240 | |
241 | =item * |
242 | |
243 | Shells interpret scripts a little bit at a time. Perl compiles the |
244 | entire program before executing it (except for C<BEGIN> blocks, which |
245 | execute at compile time). |
246 | |
247 | =item * |
248 | |
249 | The arguments are available via @ARGV, not $1, $2, etc. |
250 | |
251 | =item * |
252 | |
253 | The environment is not automatically made available as separate scalar |
254 | variables. |
255 | |
256 | =back |
257 | |
258 | =head2 Perl Traps |
259 | |
260 | Practicing Perl Programmers should take note of the following: |
261 | |
262 | =over 4 |
263 | |
264 | =item * |
265 | |
266 | Remember that many operations behave differently in a list |
267 | context than they do in a scalar one. See L<perldata> for details. |
268 | |
269 | =item * |
270 | |
68dc0745 |
271 | Avoid barewords if you can, especially all lowercase ones. |
54310121 |
272 | You can't tell by just looking at it whether a bareword is |
273 | a function or a string. By using quotes on strings and |
5f05dabc |
274 | parentheses on function calls, you won't ever get them confused. |
a0d0e21e |
275 | |
276 | =item * |
277 | |
54310121 |
278 | You cannot discern from mere inspection which builtins |
279 | are unary operators (like chop() and chdir()) |
a0d0e21e |
280 | and which are list operators (like print() and unlink()). |
5f05dabc |
281 | (User-defined subroutines can be B<only> list operators, never |
a0d0e21e |
282 | unary ones.) See L<perlop>. |
283 | |
284 | =item * |
285 | |
748a9306 |
286 | People have a hard time remembering that some functions |
a0d0e21e |
287 | default to $_, or @ARGV, or whatever, but that others which |
54310121 |
288 | you might expect to do not. |
a0d0e21e |
289 | |
6dbacca0 |
290 | =item * |
a0d0e21e |
291 | |
c47ff5f1 |
292 | The <FH> construct is not the name of the filehandle, it is a readline |
5f05dabc |
293 | operation on that handle. The data read is assigned to $_ only if the |
748a9306 |
294 | file read is the sole condition in a while loop: |
295 | |
296 | while (<FH>) { } |
54310121 |
297 | while (defined($_ = <FH>)) { }.. |
748a9306 |
298 | <FH>; # data discarded! |
299 | |
6dbacca0 |
300 | =item * |
748a9306 |
301 | |
19799a22 |
302 | Remember not to use C<=> when you need C<=~>; |
a0d0e21e |
303 | these two constructs are quite different: |
304 | |
305 | $x = /foo/; |
306 | $x =~ /foo/; |
307 | |
308 | =item * |
309 | |
54310121 |
310 | The C<do {}> construct isn't a real loop that you can use |
a0d0e21e |
311 | loop control on. |
312 | |
313 | =item * |
314 | |
54310121 |
315 | Use C<my()> for local variables whenever you can get away with |
316 | it (but see L<perlform> for where you can't). |
317 | Using C<local()> actually gives a local value to a global |
a0d0e21e |
318 | variable, which leaves you open to unforeseen side-effects |
319 | of dynamic scoping. |
320 | |
c07a80fd |
321 | =item * |
322 | |
323 | If you localize an exported variable in a module, its exported value will |
324 | not change. The local name becomes an alias to a new value but the |
325 | external name is still an alias for the original. |
326 | |
a0d0e21e |
327 | =back |
328 | |
5f05dabc |
329 | =head2 Perl4 to Perl5 Traps |
a0d0e21e |
330 | |
54310121 |
331 | Practicing Perl4 Programmers should take note of the following |
6dbacca0 |
332 | Perl4-to-Perl5 specific traps. |
333 | |
334 | They're crudely ordered according to the following list: |
a0d0e21e |
335 | |
336 | =over 4 |
337 | |
6dbacca0 |
338 | =item Discontinuance, Deprecation, and BugFix traps |
a0d0e21e |
339 | |
6dbacca0 |
340 | Anything that's been fixed as a perl4 bug, removed as a perl4 feature |
341 | or deprecated as a perl4 feature with the intent to encourage usage of |
342 | some other perl5 feature. |
a0d0e21e |
343 | |
6dbacca0 |
344 | =item Parsing Traps |
748a9306 |
345 | |
6dbacca0 |
346 | Traps that appear to stem from the new parser. |
a0d0e21e |
347 | |
6dbacca0 |
348 | =item Numerical Traps |
a0d0e21e |
349 | |
6dbacca0 |
350 | Traps having to do with numerical or mathematical operators. |
a0d0e21e |
351 | |
6dbacca0 |
352 | =item General data type traps |
a0d0e21e |
353 | |
6dbacca0 |
354 | Traps involving perl standard data types. |
a0d0e21e |
355 | |
6dbacca0 |
356 | =item Context Traps - scalar, list contexts |
357 | |
358 | Traps related to context within lists, scalar statements/declarations. |
359 | |
360 | =item Precedence Traps |
361 | |
362 | Traps related to the precedence of parsing, evaluation, and execution of |
363 | code. |
364 | |
365 | =item General Regular Expression Traps using s///, etc. |
366 | |
367 | Traps related to the use of pattern matching. |
368 | |
369 | =item Subroutine, Signal, Sorting Traps |
370 | |
371 | Traps related to the use of signals and signal handlers, general subroutines, |
372 | and sorting, along with sorting subroutines. |
373 | |
374 | =item OS Traps |
375 | |
376 | OS-specific traps. |
377 | |
378 | =item DBM Traps |
379 | |
380 | Traps specific to the use of C<dbmopen()>, and specific dbm implementations. |
381 | |
382 | =item Unclassified Traps |
383 | |
384 | Everything else. |
385 | |
386 | =back |
387 | |
388 | If you find an example of a conversion trap that is not listed here, |
4375e838 |
389 | please submit it to <F<perlbug@perl.org>> for inclusion. |
9f1b1f2d |
390 | Also note that at least some of these can be caught with the |
391 | C<use warnings> pragma or the B<-w> switch. |
6dbacca0 |
392 | |
393 | =head2 Discontinuance, Deprecation, and BugFix traps |
394 | |
395 | Anything that has been discontinued, deprecated, or fixed as |
54310121 |
396 | a bug from perl4. |
a0d0e21e |
397 | |
6dbacca0 |
398 | =over 4 |
399 | |
54310121 |
400 | =item * Discontinuance |
6dbacca0 |
401 | |
402 | Symbols starting with "_" are no longer forced into package main, except |
403 | for C<$_> itself (and C<@_>, etc.). |
404 | |
405 | package test; |
406 | $_legacy = 1; |
cb1a09d0 |
407 | |
6dbacca0 |
408 | package main; |
409 | print "\$_legacy is ",$_legacy,"\n"; |
54310121 |
410 | |
6dbacca0 |
411 | # perl4 prints: $_legacy is 1 |
412 | # perl5 prints: $_legacy is |
413 | |
54310121 |
414 | =item * Deprecation |
6dbacca0 |
415 | |
416 | Double-colon is now a valid package separator in a variable name. Thus these |
5f05dabc |
417 | behave differently in perl4 vs. perl5, because the packages don't exist. |
6dbacca0 |
418 | |
419 | $a=1;$b=2;$c=3;$var=4; |
420 | print "$a::$b::$c "; |
cb1a09d0 |
421 | print "$var::abc::xyz\n"; |
c47ff5f1 |
422 | |
6dbacca0 |
423 | # perl4 prints: 1::2::3 4::abc::xyz |
424 | # perl5 prints: 3 |
cb1a09d0 |
425 | |
6dbacca0 |
426 | Given that C<::> is now the preferred package delimiter, it is debatable |
427 | whether this should be classed as a bug or not. |
428 | (The older package delimiter, ' ,is used here) |
cb1a09d0 |
429 | |
6dbacca0 |
430 | $x = 10 ; |
431 | print "x=${'x}\n" ; |
54310121 |
432 | |
6dbacca0 |
433 | # perl4 prints: x=10 |
434 | # perl5 prints: Can't find string terminator "'" anywhere before EOF |
a0d0e21e |
435 | |
5e77893f |
436 | You can avoid this problem, and remain compatible with perl4, if you |
437 | always explicitly include the package name: |
438 | |
439 | $x = 10 ; |
440 | print "x=${main'x}\n" ; |
441 | |
54310121 |
442 | Also see precedence traps, for parsing C<$:>. |
a0d0e21e |
443 | |
6dbacca0 |
444 | =item * BugFix |
a0d0e21e |
445 | |
6dbacca0 |
446 | The second and third arguments of C<splice()> are now evaluated in scalar |
447 | context (as the Camel says) rather than list context. |
a0d0e21e |
448 | |
1d2dff63 |
449 | sub sub1{return(0,2) } # return a 2-element list |
450 | sub sub2{ return(1,2,3)} # return a 3-element list |
54310121 |
451 | @a1 = ("a","b","c","d","e"); |
6dbacca0 |
452 | @a2 = splice(@a1,&sub1,&sub2); |
453 | print join(' ',@a2),"\n"; |
54310121 |
454 | |
6dbacca0 |
455 | # perl4 prints: a b |
54310121 |
456 | # perl5 prints: c d e |
a0d0e21e |
457 | |
54310121 |
458 | =item * Discontinuance |
a0d0e21e |
459 | |
6dbacca0 |
460 | You can't do a C<goto> into a block that is optimized away. Darn. |
a0d0e21e |
461 | |
6dbacca0 |
462 | goto marker1; |
a0d0e21e |
463 | |
54310121 |
464 | for(1){ |
6dbacca0 |
465 | marker1: |
466 | print "Here I is!\n"; |
54310121 |
467 | } |
468 | |
6dbacca0 |
469 | # perl4 prints: Here I is! |
470 | # perl5 dumps core (SEGV) |
471 | |
54310121 |
472 | =item * Discontinuance |
6dbacca0 |
473 | |
474 | It is no longer syntactically legal to use whitespace as the name |
475 | of a variable, or as a delimiter for any kind of quote construct. |
54310121 |
476 | Double darn. |
6dbacca0 |
477 | |
478 | $a = ("foo bar"); |
479 | $b = q baz ; |
480 | print "a is $a, b is $b\n"; |
54310121 |
481 | |
6dbacca0 |
482 | # perl4 prints: a is foo bar, b is baz |
54310121 |
483 | # perl5 errors: Bareword found where operator expected |
5e378fdf |
484 | |
6dbacca0 |
485 | =item * Discontinuance |
486 | |
487 | The archaic while/if BLOCK BLOCK syntax is no longer supported. |
488 | |
489 | if { 1 } { |
490 | print "True!"; |
491 | } |
492 | else { |
493 | print "False!"; |
494 | } |
54310121 |
495 | |
6dbacca0 |
496 | # perl4 prints: True! |
497 | # perl5 errors: syntax error at test.pl line 1, near "if {" |
498 | |
499 | =item * BugFix |
500 | |
501 | The C<**> operator now binds more tightly than unary minus. |
502 | It was documented to work this way before, but didn't. |
503 | |
504 | print -4**2,"\n"; |
54310121 |
505 | |
6dbacca0 |
506 | # perl4 prints: 16 |
507 | # perl5 prints: -16 |
508 | |
54310121 |
509 | =item * Discontinuance |
6dbacca0 |
510 | |
511 | The meaning of C<foreach{}> has changed slightly when it is iterating over a |
512 | list which is not an array. This used to assign the list to a |
513 | temporary array, but no longer does so (for efficiency). This means |
514 | that you'll now be iterating over the actual values, not over copies of |
515 | the values. Modifications to the loop variable can change the original |
516 | values. |
517 | |
518 | @list = ('ab','abc','bcd','def'); |
519 | foreach $var (grep(/ab/,@list)){ |
520 | $var = 1; |
521 | } |
522 | print (join(':',@list)); |
54310121 |
523 | |
6dbacca0 |
524 | # perl4 prints: ab:abc:bcd:def |
525 | # perl5 prints: 1:1:bcd:def |
526 | |
527 | To retain Perl4 semantics you need to assign your list |
54310121 |
528 | explicitly to a temporary array and then iterate over that. For |
6dbacca0 |
529 | example, you might need to change |
530 | |
531 | foreach $var (grep(/ab/,@list)){ |
532 | |
533 | to |
534 | |
535 | foreach $var (@tmp = grep(/ab/,@list)){ |
536 | |
537 | Otherwise changing $var will clobber the values of @list. (This most often |
538 | happens when you use C<$_> for the loop variable, and call subroutines in |
539 | the loop that don't properly localize C<$_>.) |
540 | |
5e378fdf |
541 | =item * Discontinuance |
542 | |
543 | C<split> with no arguments now behaves like C<split ' '> (which doesn't |
544 | return an initial null field if $_ starts with whitespace), it used to |
545 | behave like C<split /\s+/> (which does). |
546 | |
547 | $_ = ' hi mom'; |
548 | print join(':', split); |
549 | |
550 | # perl4 prints: :hi:mom |
551 | # perl5 prints: hi:mom |
552 | |
55497cff |
553 | =item * BugFix |
554 | |
9607fc9c |
555 | Perl 4 would ignore any text which was attached to an B<-e> switch, |
55497cff |
556 | always taking the code snippet from the following arg. Additionally, it |
9607fc9c |
557 | would silently accept an B<-e> switch without a following arg. Both of |
55497cff |
558 | these behaviors have been fixed. |
559 | |
560 | perl -e'print "attached to -e"' 'print "separate arg"' |
54310121 |
561 | |
55497cff |
562 | # perl4 prints: separate arg |
563 | # perl5 prints: attached to -e |
54310121 |
564 | |
55497cff |
565 | perl -e |
566 | |
567 | # perl4 prints: |
568 | # perl5 dies: No code specified for -e. |
569 | |
570 | =item * Discontinuance |
571 | |
572 | In Perl 4 the return value of C<push> was undocumented, but it was |
573 | actually the last value being pushed onto the target list. In Perl 5 |
574 | the return value of C<push> is documented, but has changed, it is the |
575 | number of elements in the resulting list. |
576 | |
577 | @x = ('existing'); |
578 | print push(@x, 'first new', 'second new'); |
54310121 |
579 | |
55497cff |
580 | # perl4 prints: second new |
581 | # perl5 prints: 3 |
582 | |
6dbacca0 |
583 | =item * Deprecation |
584 | |
585 | Some error messages will be different. |
586 | |
54310121 |
587 | =item * Discontinuance |
6dbacca0 |
588 | |
589 | Some bugs may have been inadvertently removed. :-) |
590 | |
591 | =back |
592 | |
593 | =head2 Parsing Traps |
594 | |
595 | Perl4-to-Perl5 traps from having to do with parsing. |
596 | |
597 | =over 4 |
598 | |
599 | =item * Parsing |
600 | |
601 | Note the space between . and = |
602 | |
603 | $string . = "more string"; |
604 | print $string; |
54310121 |
605 | |
6dbacca0 |
606 | # perl4 prints: more string |
607 | # perl5 prints: syntax error at - line 1, near ". =" |
608 | |
609 | =item * Parsing |
610 | |
611 | Better parsing in perl 5 |
612 | |
613 | sub foo {} |
614 | &foo |
615 | print("hello, world\n"); |
54310121 |
616 | |
6dbacca0 |
617 | # perl4 prints: hello, world |
618 | # perl5 prints: syntax error |
619 | |
620 | =item * Parsing |
621 | |
622 | "if it looks like a function, it is a function" rule. |
623 | |
624 | print |
625 | ($foo == 1) ? "is one\n" : "is zero\n"; |
54310121 |
626 | |
6dbacca0 |
627 | # perl4 prints: is zero |
628 | # perl5 warns: "Useless use of a constant in void context" if using -w |
629 | |
c12982c8 |
630 | =item * Parsing |
631 | |
632 | String interpolation of the C<$#array> construct differs when braces |
633 | are to used around the name. |
634 | |
635 | @ = (1..3); |
636 | print "${#a}"; |
637 | |
638 | # perl4 prints: 2 |
639 | # perl5 fails with syntax error |
640 | |
641 | @ = (1..3); |
642 | print "$#{a}"; |
643 | |
644 | # perl4 prints: {a} |
645 | # perl5 prints: 2 |
646 | |
6dbacca0 |
647 | =back |
648 | |
649 | =head2 Numerical Traps |
650 | |
651 | Perl4-to-Perl5 traps having to do with numerical operators, |
652 | operands, or output from same. |
653 | |
654 | =over 5 |
655 | |
656 | =item * Numerical |
657 | |
658 | Formatted output and significant digits |
659 | |
660 | print 7.373504 - 0, "\n"; |
54310121 |
661 | printf "%20.18f\n", 7.373504 - 0; |
662 | |
6dbacca0 |
663 | # Perl4 prints: |
664 | 7.375039999999996141 |
665 | 7.37503999999999614 |
54310121 |
666 | |
6dbacca0 |
667 | # Perl5 prints: |
668 | 7.373504 |
669 | 7.37503999999999614 |
670 | |
671 | =item * Numerical |
672 | |
5f05dabc |
673 | This specific item has been deleted. It demonstrated how the auto-increment |
5e378fdf |
674 | operator would not catch when a number went over the signed int limit. Fixed |
a6006777 |
675 | in version 5.003_04. But always be wary when using large integers. |
676 | If in doubt: |
6dbacca0 |
677 | |
5e378fdf |
678 | use Math::BigInt; |
6dbacca0 |
679 | |
54310121 |
680 | =item * Numerical |
6dbacca0 |
681 | |
682 | Assignment of return values from numeric equality tests |
683 | does not work in perl5 when the test evaluates to false (0). |
684 | Logical tests now return an null, instead of 0 |
a6006777 |
685 | |
6dbacca0 |
686 | $p = ($test == 1); |
687 | print $p,"\n"; |
a6006777 |
688 | |
6dbacca0 |
689 | # perl4 prints: 0 |
690 | # perl5 prints: |
691 | |
dc848c6f |
692 | Also see L<"General Regular Expression Traps using s///, etc."> |
693 | for another example of this new feature... |
6dbacca0 |
694 | |
651ad3b1 |
695 | =item * Bitwise string ops |
696 | |
697 | When bitwise operators which can operate upon either numbers or |
698 | strings (C<& | ^ ~>) are given only strings as arguments, perl4 would |
699 | treat the operands as bitstrings so long as the program contained a call |
700 | to the C<vec()> function. perl5 treats the string operands as bitstrings. |
701 | (See L<perlop/Bitwise String Operators> for more details.) |
702 | |
703 | $fred = "10"; |
704 | $barney = "12"; |
705 | $betty = $fred & $barney; |
706 | print "$betty\n"; |
707 | # Uncomment the next line to change perl4's behavior |
708 | # ($dummy) = vec("dummy", 0, 0); |
709 | |
710 | # Perl4 prints: |
711 | 8 |
712 | |
713 | # Perl5 prints: |
714 | 10 |
715 | |
716 | # If vec() is used anywhere in the program, both print: |
717 | 10 |
718 | |
6dbacca0 |
719 | =back |
720 | |
721 | =head2 General data type traps |
722 | |
723 | Perl4-to-Perl5 traps involving most data-types, and their usage |
724 | within certain expressions and/or context. |
725 | |
726 | =over 5 |
727 | |
728 | =item * (Arrays) |
729 | |
730 | Negative array subscripts now count from the end of the array. |
731 | |
732 | @a = (1, 2, 3, 4, 5); |
733 | print "The third element of the array is $a[3] also expressed as $a[-2] \n"; |
54310121 |
734 | |
6dbacca0 |
735 | # perl4 prints: The third element of the array is 4 also expressed as |
736 | # perl5 prints: The third element of the array is 4 also expressed as 4 |
737 | |
738 | =item * (Arrays) |
739 | |
740 | Setting C<$#array> lower now discards array elements, and makes them |
741 | impossible to recover. |
742 | |
54310121 |
743 | @a = (a,b,c,d,e); |
6dbacca0 |
744 | print "Before: ",join('',@a); |
54310121 |
745 | $#a =1; |
6dbacca0 |
746 | print ", After: ",join('',@a); |
747 | $#a =3; |
748 | print ", Recovered: ",join('',@a),"\n"; |
54310121 |
749 | |
6dbacca0 |
750 | # perl4 prints: Before: abcde, After: ab, Recovered: abcd |
751 | # perl5 prints: Before: abcde, After: ab, Recovered: ab |
752 | |
753 | =item * (Hashes) |
754 | |
755 | Hashes get defined before use |
756 | |
54310121 |
757 | local($s,@a,%h); |
6dbacca0 |
758 | die "scalar \$s defined" if defined($s); |
759 | die "array \@a defined" if defined(@a); |
760 | die "hash \%h defined" if defined(%h); |
54310121 |
761 | |
6dbacca0 |
762 | # perl4 prints: |
763 | # perl5 dies: hash %h defined |
764 | |
475342a6 |
765 | Perl will now generate a warning when it sees defined(@a) and |
766 | defined(%h). |
767 | |
6dbacca0 |
768 | =item * (Globs) |
769 | |
770 | glob assignment from variable to variable will fail if the assigned |
771 | variable is localized subsequent to the assignment |
772 | |
773 | @a = ("This is Perl 4"); |
774 | *b = *a; |
775 | local(@a); |
776 | print @b,"\n"; |
54310121 |
777 | |
6dbacca0 |
778 | # perl4 prints: This is Perl 4 |
779 | # perl5 prints: |
54310121 |
780 | |
a3cb178b |
781 | =item * (Globs) |
54310121 |
782 | |
a3cb178b |
783 | Assigning C<undef> to a glob has no effect in Perl 5. In Perl 4 |
784 | it undefines the associated scalar (but may have other side effects |
785 | including SEGVs). |
5e378fdf |
786 | |
6dbacca0 |
787 | =item * (Scalar String) |
788 | |
789 | Changes in unary negation (of strings) |
790 | This change effects both the return value and what it |
791 | does to auto(magic)increment. |
792 | |
793 | $x = "aaa"; |
794 | print ++$x," : "; |
795 | print -$x," : "; |
796 | print ++$x,"\n"; |
54310121 |
797 | |
6dbacca0 |
798 | # perl4 prints: aab : -0 : 1 |
799 | # perl5 prints: aab : -aab : aac |
800 | |
801 | =item * (Constants) |
802 | |
803 | perl 4 lets you modify constants: |
804 | |
805 | $foo = "x"; |
806 | &mod($foo); |
807 | for ($x = 0; $x < 3; $x++) { |
808 | &mod("a"); |
809 | } |
810 | sub mod { |
811 | print "before: $_[0]"; |
812 | $_[0] = "m"; |
813 | print " after: $_[0]\n"; |
814 | } |
54310121 |
815 | |
6dbacca0 |
816 | # perl4: |
817 | # before: x after: m |
818 | # before: a after: m |
819 | # before: m after: m |
820 | # before: m after: m |
54310121 |
821 | |
6dbacca0 |
822 | # Perl5: |
823 | # before: x after: m |
824 | # Modification of a read-only value attempted at foo.pl line 12. |
825 | # before: a |
826 | |
827 | =item * (Scalars) |
828 | |
829 | The behavior is slightly different for: |
830 | |
831 | print "$x", defined $x |
54310121 |
832 | |
6dbacca0 |
833 | # perl 4: 1 |
834 | # perl 5: <no output, $x is not called into existence> |
835 | |
836 | =item * (Variable Suicide) |
837 | |
838 | Variable suicide behavior is more consistent under Perl 5. |
aa689395 |
839 | Perl5 exhibits the same behavior for hashes and scalars, |
5f05dabc |
840 | that perl4 exhibits for only scalars. |
6dbacca0 |
841 | |
842 | $aGlobal{ "aKey" } = "global value"; |
843 | print "MAIN:", $aGlobal{"aKey"}, "\n"; |
844 | $GlobalLevel = 0; |
845 | &test( *aGlobal ); |
846 | |
847 | sub test { |
848 | local( *theArgument ) = @_; |
849 | local( %aNewLocal ); # perl 4 != 5.001l,m |
54310121 |
850 | $aNewLocal{"aKey"} = "this should never appear"; |
6dbacca0 |
851 | print "SUB: ", $theArgument{"aKey"}, "\n"; |
852 | $aNewLocal{"aKey"} = "level $GlobalLevel"; # what should print |
853 | $GlobalLevel++; |
854 | if( $GlobalLevel<4 ) { |
855 | &test( *aNewLocal ); |
856 | } |
857 | } |
54310121 |
858 | |
6dbacca0 |
859 | # Perl4: |
860 | # MAIN:global value |
861 | # SUB: global value |
862 | # SUB: level 0 |
863 | # SUB: level 1 |
864 | # SUB: level 2 |
54310121 |
865 | |
6dbacca0 |
866 | # Perl5: |
867 | # MAIN:global value |
868 | # SUB: global value |
869 | # SUB: this should never appear |
870 | # SUB: this should never appear |
871 | # SUB: this should never appear |
872 | |
84dc3c4d |
873 | =back |
6dbacca0 |
874 | |
875 | =head2 Context Traps - scalar, list contexts |
876 | |
877 | =over 5 |
878 | |
879 | =item * (list context) |
880 | |
881 | The elements of argument lists for formats are now evaluated in list |
882 | context. This means you can interpolate list values now. |
883 | |
884 | @fmt = ("foo","bar","baz"); |
885 | format STDOUT= |
886 | @<<<<< @||||| @>>>>> |
887 | @fmt; |
888 | . |
54310121 |
889 | write; |
890 | |
6dbacca0 |
891 | # perl4 errors: Please use commas to separate fields in file |
892 | # perl5 prints: foo bar baz |
893 | |
894 | =item * (scalar context) |
895 | |
54310121 |
896 | The C<caller()> function now returns a false value in a scalar context |
897 | if there is no caller. This lets library files determine if they're |
6dbacca0 |
898 | being required. |
899 | |
900 | caller() ? (print "You rang?\n") : (print "Got a 0\n"); |
54310121 |
901 | |
6dbacca0 |
902 | # perl4 errors: There is no caller |
903 | # perl5 prints: Got a 0 |
5e378fdf |
904 | |
6dbacca0 |
905 | =item * (scalar context) |
906 | |
907 | The comma operator in a scalar context is now guaranteed to give a |
908 | scalar context to its arguments. |
909 | |
910 | @y= ('a','b','c'); |
911 | $x = (1, 2, @y); |
912 | print "x = $x\n"; |
54310121 |
913 | |
6dbacca0 |
914 | # Perl4 prints: x = c # Thinks list context interpolates list |
915 | # Perl5 prints: x = 3 # Knows scalar uses length of list |
916 | |
917 | =item * (list, builtin) |
918 | |
919 | C<sprintf()> funkiness (array argument converted to scalar array count) |
920 | This test could be added to t/op/sprintf.t |
921 | |
922 | @z = ('%s%s', 'foo', 'bar'); |
923 | $x = sprintf(@z); |
924 | if ($x eq 'foobar') {print "ok 2\n";} else {print "not ok 2 '$x'\n";} |
54310121 |
925 | |
6dbacca0 |
926 | # perl4 prints: ok 2 |
927 | # perl5 prints: not ok 2 |
928 | |
929 | C<printf()> works fine, though: |
930 | |
931 | printf STDOUT (@z); |
54310121 |
932 | print "\n"; |
933 | |
6dbacca0 |
934 | # perl4 prints: foobar |
935 | # perl5 prints: foobar |
936 | |
937 | Probably a bug. |
938 | |
939 | =back |
940 | |
941 | =head2 Precedence Traps |
942 | |
943 | Perl4-to-Perl5 traps involving precedence order. |
944 | |
f4b17341 |
945 | Perl 4 has almost the same precedence rules as Perl 5 for the operators |
946 | that they both have. Perl 4 however, seems to have had some |
947 | inconsistencies that made the behavior differ from what was documented. |
948 | |
84dc3c4d |
949 | =over 5 |
950 | |
5e378fdf |
951 | =item * Precedence |
952 | |
8dbef698 |
953 | LHS vs. RHS of any assignment operator. LHS is evaluated first |
954 | in perl4, second in perl5; this can affect the relationship |
955 | between side-effects in sub-expressions. |
5e378fdf |
956 | |
957 | @arr = ( 'left', 'right' ); |
958 | $a{shift @arr} = shift @arr; |
959 | print join( ' ', keys %a ); |
960 | |
961 | # perl4 prints: left |
962 | # perl5 prints: right |
963 | |
964 | =item * Precedence |
6dbacca0 |
965 | |
966 | These are now semantic errors because of precedence: |
967 | |
968 | @list = (1,2,3,4,5); |
969 | %map = ("a",1,"b",2,"c",3,"d",4); |
970 | $n = shift @list + 2; # first item in list plus 2 |
971 | print "n is $n, "; |
972 | $m = keys %map + 2; # number of items in hash plus 2 |
973 | print "m is $m\n"; |
54310121 |
974 | |
6dbacca0 |
975 | # perl4 prints: n is 3, m is 6 |
976 | # perl5 errors and fails to compile |
977 | |
978 | =item * Precedence |
a0d0e21e |
979 | |
4633a7c4 |
980 | The precedence of assignment operators is now the same as the precedence |
981 | of assignment. Perl 4 mistakenly gave them the precedence of the associated |
982 | operator. So you now must parenthesize them in expressions like |
983 | |
984 | /foo/ ? ($a += 2) : ($a -= 2); |
a6006777 |
985 | |
4633a7c4 |
986 | Otherwise |
987 | |
6dbacca0 |
988 | /foo/ ? $a += 2 : $a -= 2 |
4633a7c4 |
989 | |
990 | would be erroneously parsed as |
991 | |
992 | (/foo/ ? $a += 2 : $a) -= 2; |
993 | |
994 | On the other hand, |
995 | |
54310121 |
996 | $a += /foo/ ? 1 : 2; |
4633a7c4 |
997 | |
998 | now works as a C programmer would expect. |
999 | |
6dbacca0 |
1000 | =item * Precedence |
4633a7c4 |
1001 | |
6dbacca0 |
1002 | open FOO || die; |
a0d0e21e |
1003 | |
5f05dabc |
1004 | is now incorrect. You need parentheses around the filehandle. |
1005 | Otherwise, perl5 leaves the statement as its default precedence: |
a0d0e21e |
1006 | |
6dbacca0 |
1007 | open(FOO || die); |
54310121 |
1008 | |
6dbacca0 |
1009 | # perl4 opens or dies |
1010 | # perl5 errors: Precedence problem: open FOO should be open(FOO) |
a0d0e21e |
1011 | |
6dbacca0 |
1012 | =item * Precedence |
a0d0e21e |
1013 | |
6dbacca0 |
1014 | perl4 gives the special variable, C<$:> precedence, where perl5 |
1015 | treats C<$::> as main C<package> |
a0d0e21e |
1016 | |
6dbacca0 |
1017 | $a = "x"; print "$::a"; |
54310121 |
1018 | |
6dbacca0 |
1019 | # perl 4 prints: -:a |
1020 | # perl 5 prints: x |
5e378fdf |
1021 | |
6dbacca0 |
1022 | =item * Precedence |
a0d0e21e |
1023 | |
f4b17341 |
1024 | perl4 had buggy precedence for the file test operators vis-a-vis |
1025 | the assignment operators. Thus, although the precedence table |
1026 | for perl4 leads one to believe C<-e $foo .= "q"> should parse as |
1027 | C<((-e $foo) .= "q")>, it actually parses as C<(-e ($foo .= "q"))>. |
1028 | In perl5, the precedence is as documented. |
54310121 |
1029 | |
1030 | -e $foo .= "q" |
a0d0e21e |
1031 | |
6dbacca0 |
1032 | # perl4 prints: no output |
1033 | # perl5 prints: Can't modify -e in concatenation |
a0d0e21e |
1034 | |
f4b17341 |
1035 | =item * Precedence |
1036 | |
1037 | In perl4, keys(), each() and values() were special high-precedence operators |
1038 | that operated on a single hash, but in perl5, they are regular named unary |
1039 | operators. As documented, named unary operators have lower precedence |
1040 | than the arithmetic and concatenation operators C<+ - .>, but the perl4 |
1041 | variants of these operators actually bind tighter than C<+ - .>. |
1042 | Thus, for: |
1043 | |
1044 | %foo = 1..10; |
1045 | print keys %foo - 1 |
1046 | |
1047 | # perl4 prints: 4 |
1048 | # perl5 prints: Type of arg 1 to keys must be hash (not subtraction) |
1049 | |
1050 | The perl4 behavior was probably more useful, if less consistent. |
1051 | |
6dbacca0 |
1052 | =back |
1053 | |
1054 | =head2 General Regular Expression Traps using s///, etc. |
1055 | |
1056 | All types of RE traps. |
1057 | |
1058 | =over 5 |
1059 | |
1060 | =item * Regular Expression |
1061 | |
1062 | C<s'$lhs'$rhs'> now does no interpolation on either side. It used to |
19799a22 |
1063 | interpolate $lhs but not $rhs. (And still does not match a literal |
6dbacca0 |
1064 | '$' in string) |
1065 | |
1066 | $a=1;$b=2; |
1067 | $string = '1 2 $a $b'; |
1068 | $string =~ s'$a'$b'; |
1069 | print $string,"\n"; |
54310121 |
1070 | |
6dbacca0 |
1071 | # perl4 prints: $b 2 $a $b |
1072 | # perl5 prints: 1 2 $a $b |
1073 | |
1074 | =item * Regular Expression |
a0d0e21e |
1075 | |
1076 | C<m//g> now attaches its state to the searched string rather than the |
6dbacca0 |
1077 | regular expression. (Once the scope of a block is left for the sub, the |
1078 | state of the searched string is lost) |
1079 | |
1080 | $_ = "ababab"; |
1081 | while(m/ab/g){ |
1082 | &doit("blah"); |
1083 | } |
1084 | sub doit{local($_) = shift; print "Got $_ "} |
54310121 |
1085 | |
6dbacca0 |
1086 | # perl4 prints: blah blah blah |
1087 | # perl5 prints: infinite loop blah... |
1088 | |
1089 | =item * Regular Expression |
1090 | |
68dc0745 |
1091 | Currently, if you use the C<m//o> qualifier on a regular expression |
1092 | within an anonymous sub, I<all> closures generated from that anonymous |
1093 | sub will use the regular expression as it was compiled when it was used |
1094 | the very first time in any such closure. For instance, if you say |
1095 | |
1096 | sub build_match { |
1097 | my($left,$right) = @_; |
1098 | return sub { $_[0] =~ /$left stuff $right/o; }; |
1099 | } |
1100 | |
1101 | build_match() will always return a sub which matches the contents of |
19799a22 |
1102 | $left and $right as they were the I<first> time that build_match() |
68dc0745 |
1103 | was called, not as they are in the current call. |
1104 | |
1105 | This is probably a bug, and may change in future versions of Perl. |
1106 | |
1107 | =item * Regular Expression |
1108 | |
6dbacca0 |
1109 | If no parentheses are used in a match, Perl4 sets C<$+> to |
1110 | the whole match, just like C<$&>. Perl5 does not. |
1111 | |
1112 | "abcdef" =~ /b.*e/; |
1113 | print "\$+ = $+\n"; |
54310121 |
1114 | |
6dbacca0 |
1115 | # perl4 prints: bcde |
1116 | # perl5 prints: |
1117 | |
1118 | =item * Regular Expression |
1119 | |
1120 | substitution now returns the null string if it fails |
1121 | |
1122 | $string = "test"; |
1123 | $value = ($string =~ s/foo//); |
1124 | print $value, "\n"; |
54310121 |
1125 | |
6dbacca0 |
1126 | # perl4 prints: 0 |
1127 | # perl5 prints: |
1128 | |
1129 | Also see L<Numerical Traps> for another example of this new feature. |
1130 | |
1131 | =item * Regular Expression |
1132 | |
54310121 |
1133 | C<s`lhs`rhs`> (using backticks) is now a normal substitution, with no |
1134 | backtick expansion |
6dbacca0 |
1135 | |
1136 | $string = ""; |
1137 | $string =~ s`^`hostname`; |
1138 | print $string, "\n"; |
54310121 |
1139 | |
6dbacca0 |
1140 | # perl4 prints: <the local hostname> |
1141 | # perl5 prints: hostname |
1142 | |
1143 | =item * Regular Expression |
1144 | |
1145 | Stricter parsing of variables used in regular expressions |
1146 | |
1147 | s/^([^$grpc]*$grpc[$opt$plus$rep]?)//o; |
54310121 |
1148 | |
6dbacca0 |
1149 | # perl4: compiles w/o error |
1150 | # perl5: with Scalar found where operator expected ..., near "$opt$plus" |
1151 | |
1152 | an added component of this example, apparently from the same script, is |
1153 | the actual value of the s'd string after the substitution. |
1154 | C<[$opt]> is a character class in perl4 and an array subscript in perl5 |
1155 | |
54310121 |
1156 | $grpc = 'a'; |
6dbacca0 |
1157 | $opt = 'r'; |
1158 | $_ = 'bar'; |
1159 | s/^([^$grpc]*$grpc[$opt]?)/foo/; |
1160 | print ; |
54310121 |
1161 | |
6dbacca0 |
1162 | # perl4 prints: foo |
1163 | # perl5 prints: foobar |
1164 | |
1165 | =item * Regular Expression |
1166 | |
1167 | Under perl5, C<m?x?> matches only once, like C<?x?>. Under perl4, it matched |
1168 | repeatedly, like C</x/> or C<m!x!>. |
1169 | |
1170 | $test = "once"; |
1171 | sub match { $test =~ m?once?; } |
1172 | &match(); |
1173 | if( &match() ) { |
1174 | # m?x? matches more then once |
1175 | print "perl4\n"; |
54310121 |
1176 | } else { |
6dbacca0 |
1177 | # m?x? matches only once |
54310121 |
1178 | print "perl5\n"; |
6dbacca0 |
1179 | } |
54310121 |
1180 | |
6dbacca0 |
1181 | # perl4 prints: perl4 |
1182 | # perl5 prints: perl5 |
a0d0e21e |
1183 | |
a0d0e21e |
1184 | |
6dbacca0 |
1185 | =back |
1186 | |
1187 | =head2 Subroutine, Signal, Sorting Traps |
a0d0e21e |
1188 | |
6dbacca0 |
1189 | The general group of Perl4-to-Perl5 traps having to do with |
1190 | Signals, Sorting, and their related subroutines, as well as |
1191 | general subroutine traps. Includes some OS-Specific traps. |
a0d0e21e |
1192 | |
6dbacca0 |
1193 | =over 5 |
a0d0e21e |
1194 | |
6dbacca0 |
1195 | =item * (Signals) |
a0d0e21e |
1196 | |
6dbacca0 |
1197 | Barewords that used to look like strings to Perl will now look like subroutine |
1198 | calls if a subroutine by that name is defined before the compiler sees them. |
a0d0e21e |
1199 | |
6dbacca0 |
1200 | sub SeeYa { warn"Hasta la vista, baby!" } |
1201 | $SIG{'TERM'} = SeeYa; |
1202 | print "SIGTERM is now $SIG{'TERM'}\n"; |
54310121 |
1203 | |
6dbacca0 |
1204 | # perl4 prints: SIGTERM is main'SeeYa |
1205 | # perl5 prints: SIGTERM is now main::1 |
a0d0e21e |
1206 | |
6dbacca0 |
1207 | Use B<-w> to catch this one |
a0d0e21e |
1208 | |
6dbacca0 |
1209 | =item * (Sort Subroutine) |
a0d0e21e |
1210 | |
6dbacca0 |
1211 | reverse is no longer allowed as the name of a sort subroutine. |
a0d0e21e |
1212 | |
6dbacca0 |
1213 | sub reverse{ print "yup "; $a <=> $b } |
54310121 |
1214 | print sort reverse a,b,c; |
1215 | |
6dbacca0 |
1216 | # perl4 prints: yup yup yup yup abc |
54310121 |
1217 | # perl5 prints: abc |
a0d0e21e |
1218 | |
b996531f |
1219 | =item * warn() won't let you specify a filehandle. |
1220 | |
1221 | Although it _always_ printed to STDERR, warn() would let you specify a |
1222 | filehandle in perl4. With perl5 it does not. |
5e378fdf |
1223 | |
1224 | warn STDERR "Foo!"; |
1225 | |
1226 | # perl4 prints: Foo! |
54310121 |
1227 | # perl5 prints: String found where operator expected |
5e378fdf |
1228 | |
6dbacca0 |
1229 | =back |
a0d0e21e |
1230 | |
6dbacca0 |
1231 | =head2 OS Traps |
1232 | |
1233 | =over 5 |
1234 | |
1235 | =item * (SysV) |
1236 | |
54310121 |
1237 | Under HPUX, and some other SysV OSes, one had to reset any signal handler, |
1238 | within the signal handler function, each time a signal was handled with |
1239 | perl4. With perl5, the reset is now done correctly. Any code relying |
6dbacca0 |
1240 | on the handler _not_ being reset will have to be reworked. |
1241 | |
a6006777 |
1242 | Since version 5.002, Perl uses sigaction() under SysV. |
6dbacca0 |
1243 | |
1244 | sub gotit { |
54310121 |
1245 | print "Got @_... "; |
1246 | } |
6dbacca0 |
1247 | $SIG{'INT'} = 'gotit'; |
54310121 |
1248 | |
6dbacca0 |
1249 | $| = 1; |
1250 | $pid = fork; |
1251 | if ($pid) { |
1252 | kill('INT', $pid); |
1253 | sleep(1); |
1254 | kill('INT', $pid); |
54310121 |
1255 | } else { |
6dbacca0 |
1256 | while (1) {sleep(10);} |
54310121 |
1257 | } |
1258 | |
6dbacca0 |
1259 | # perl4 (HPUX) prints: Got INT... |
1260 | # perl5 (HPUX) prints: Got INT... Got INT... |
1261 | |
1262 | =item * (SysV) |
1263 | |
c47ff5f1 |
1264 | Under SysV OSes, C<seek()> on a file opened to append C<<< >> >>> now does |
54310121 |
1265 | the right thing w.r.t. the fopen() manpage. e.g., - When a file is opened |
6dbacca0 |
1266 | for append, it is impossible to overwrite information already in |
1267 | the file. |
1268 | |
1269 | open(TEST,">>seek.test"); |
54310121 |
1270 | $start = tell TEST ; |
6dbacca0 |
1271 | foreach(1 .. 9){ |
1272 | print TEST "$_ "; |
1273 | } |
1274 | $end = tell TEST ; |
1275 | seek(TEST,$start,0); |
1276 | print TEST "18 characters here"; |
54310121 |
1277 | |
6dbacca0 |
1278 | # perl4 (solaris) seek.test has: 18 characters here |
1279 | # perl5 (solaris) seek.test has: 1 2 3 4 5 6 7 8 9 18 characters here |
a0d0e21e |
1280 | |
a0d0e21e |
1281 | |
a0d0e21e |
1282 | |
6dbacca0 |
1283 | =back |
a0d0e21e |
1284 | |
6dbacca0 |
1285 | =head2 Interpolation Traps |
a0d0e21e |
1286 | |
8b0a4b75 |
1287 | Perl4-to-Perl5 traps having to do with how things get interpolated |
1288 | within certain expressions, statements, contexts, or whatever. |
1289 | |
6dbacca0 |
1290 | =over 5 |
a0d0e21e |
1291 | |
6dbacca0 |
1292 | =item * Interpolation |
a0d0e21e |
1293 | |
6dbacca0 |
1294 | @ now always interpolates an array in double-quotish strings. |
1295 | |
54310121 |
1296 | print "To: someone@somewhere.com\n"; |
1297 | |
6dbacca0 |
1298 | # perl4 prints: To:someone@somewhere.com |
8593bda5 |
1299 | # perl < 5.6.1, error : In string, @somewhere now must be written as \@somewhere |
1300 | # perl >= 5.6.1, warning : Possible unintended interpolation of @somewhere in string |
6dbacca0 |
1301 | |
1302 | =item * Interpolation |
1303 | |
6dbacca0 |
1304 | Double-quoted strings may no longer end with an unescaped $ or @. |
1305 | |
1306 | $foo = "foo$"; |
1307 | $bar = "bar@"; |
1308 | print "foo is $foo, bar is $bar\n"; |
54310121 |
1309 | |
6dbacca0 |
1310 | # perl4 prints: foo is foo$, bar is bar@ |
1311 | # perl5 errors: Final $ should be \$ or $name |
1312 | |
1313 | Note: perl5 DOES NOT error on the terminating @ in $bar |
1314 | |
1315 | =item * Interpolation |
a0d0e21e |
1316 | |
8b0a4b75 |
1317 | Perl now sometimes evaluates arbitrary expressions inside braces that occur |
1318 | within double quotes (usually when the opening brace is preceded by C<$> |
1319 | or C<@>). |
1320 | |
1321 | @www = "buz"; |
1322 | $foo = "foo"; |
1323 | $bar = "bar"; |
1324 | sub foo { return "bar" }; |
1325 | print "|@{w.w.w}|${main'foo}|"; |
1326 | |
1327 | # perl4 prints: |@{w.w.w}|foo| |
1328 | # perl5 prints: |buz|bar| |
1329 | |
1330 | Note that you can C<use strict;> to ward off such trappiness under perl5. |
1331 | |
1332 | =item * Interpolation |
1333 | |
748a9306 |
1334 | The construct "this is $$x" used to interpolate the pid at that |
19799a22 |
1335 | point, but now apparently tries to dereference $x. C<$$> by itself still |
748a9306 |
1336 | works fine, however. |
1337 | |
6dbacca0 |
1338 | print "this is $$x\n"; |
748a9306 |
1339 | |
6dbacca0 |
1340 | # perl4 prints: this is XXXx (XXX is the current pid) |
1341 | # perl5 prints: this is |
1342 | |
1343 | =item * Interpolation |
1344 | |
54310121 |
1345 | Creation of hashes on the fly with C<eval "EXPR"> now requires either both |
1346 | C<$>'s to be protected in the specification of the hash name, or both curlies |
6dbacca0 |
1347 | to be protected. If both curlies are protected, the result will be compatible |
1348 | with perl4 and perl5. This is a very common practice, and should be changed |
1349 | to use the block form of C<eval{}> if possible. |
c07a80fd |
1350 | |
6dbacca0 |
1351 | $hashname = "foobar"; |
1352 | $key = "baz"; |
1353 | $value = 1234; |
1354 | eval "\$$hashname{'$key'} = q|$value|"; |
1355 | (defined($foobar{'baz'})) ? (print "Yup") : (print "Nope"); |
1356 | |
1357 | # perl4 prints: Yup |
1358 | # perl5 prints: Nope |
1359 | |
1360 | Changing |
1361 | |
1362 | eval "\$$hashname{'$key'} = q|$value|"; |
c07a80fd |
1363 | |
1364 | to |
1365 | |
6dbacca0 |
1366 | eval "\$\$hashname{'$key'} = q|$value|"; |
c07a80fd |
1367 | |
6dbacca0 |
1368 | causes the following result: |
c07a80fd |
1369 | |
6dbacca0 |
1370 | # perl4 prints: Nope |
1371 | # perl5 prints: Yup |
c07a80fd |
1372 | |
6dbacca0 |
1373 | or, changing to |
a0d0e21e |
1374 | |
6dbacca0 |
1375 | eval "\$$hashname\{'$key'\} = q|$value|"; |
1376 | |
1377 | causes the following result: |
1378 | |
1379 | # perl4 prints: Yup |
1380 | # perl5 prints: Yup |
1381 | # and is compatible for both versions |
1382 | |
1383 | |
1384 | =item * Interpolation |
1385 | |
1386 | perl4 programs which unconsciously rely on the bugs in earlier perl versions. |
1387 | |
1388 | perl -e '$bar=q/not/; print "This is $foo{$bar} perl5"' |
54310121 |
1389 | |
6dbacca0 |
1390 | # perl4 prints: This is not perl5 |
1391 | # perl5 prints: This is perl5 |
1392 | |
1393 | =item * Interpolation |
1394 | |
54310121 |
1395 | You also have to be careful about array references. |
6dbacca0 |
1396 | |
1397 | print "$foo{" |
1398 | |
1399 | perl 4 prints: { |
1400 | perl 5 prints: syntax error |
1401 | |
1402 | =item * Interpolation |
1403 | |
1404 | Similarly, watch out for: |
1405 | |
1406 | $foo = "array"; |
1407 | print "\$$foo{bar}\n"; |
54310121 |
1408 | |
6dbacca0 |
1409 | # perl4 prints: $array{bar} |
1410 | # perl5 prints: $ |
1411 | |
1412 | Perl 5 is looking for C<$array{bar}> which doesn't exist, but perl 4 is |
1413 | happy just to expand $foo to "array" by itself. Watch out for this |
1414 | especially in C<eval>'s. |
1415 | |
1416 | =item * Interpolation |
1417 | |
1418 | C<qq()> string passed to C<eval> |
1419 | |
1420 | eval qq( |
1421 | foreach \$y (keys %\$x\) { |
1422 | \$count++; |
1423 | } |
1424 | ); |
54310121 |
1425 | |
6dbacca0 |
1426 | # perl4 runs this ok |
54310121 |
1427 | # perl5 prints: Can't find string terminator ")" |
a0d0e21e |
1428 | |
6dbacca0 |
1429 | =back |
1430 | |
1431 | =head2 DBM Traps |
1432 | |
1433 | General DBM traps. |
1434 | |
1435 | =over 5 |
1436 | |
1437 | =item * DBM |
1438 | |
1439 | Existing dbm databases created under perl4 (or any other dbm/ndbm tool) |
1440 | may cause the same script, run under perl5, to fail. The build of perl5 |
1441 | must have been linked with the same dbm/ndbm as the default for C<dbmopen()> |
1442 | to function properly without C<tie>'ing to an extension dbm implementation. |
1443 | |
1444 | dbmopen (%dbm, "file", undef); |
1445 | print "ok\n"; |
1446 | |
1447 | # perl4 prints: ok |
1448 | # perl5 prints: ok (IFF linked with -ldbm or -lndbm) |
1449 | |
1450 | |
1451 | =item * DBM |
1452 | |
1453 | Existing dbm databases created under perl4 (or any other dbm/ndbm tool) |
1454 | may cause the same script, run under perl5, to fail. The error generated |
1455 | when exceeding the limit on the key/value size will cause perl5 to exit |
1456 | immediately. |
1457 | |
1458 | dbmopen(DB, "testdb",0600) || die "couldn't open db! $!"; |
1459 | $DB{'trap'} = "x" x 1024; # value too large for most dbm/ndbm |
1460 | print "YUP\n"; |
1461 | |
1462 | # perl4 prints: |
1463 | dbm store returned -1, errno 28, key "trap" at - line 3. |
1464 | YUP |
1465 | |
1466 | # perl5 prints: |
1467 | dbm store returned -1, errno 28, key "trap" at - line 3. |
a0d0e21e |
1468 | |
1469 | =back |
6dbacca0 |
1470 | |
1471 | =head2 Unclassified Traps |
1472 | |
1473 | Everything else. |
1474 | |
84dc3c4d |
1475 | =over 5 |
1476 | |
5db417f7 |
1477 | =item * C<require>/C<do> trap using returned value |
6dbacca0 |
1478 | |
1479 | If the file doit.pl has: |
1480 | |
1481 | sub foo { |
1482 | $rc = do "./do.pl"; |
1483 | return 8; |
54310121 |
1484 | } |
6dbacca0 |
1485 | print &foo, "\n"; |
1486 | |
1487 | And the do.pl file has the following single line: |
1488 | |
1489 | return 3; |
1490 | |
1491 | Running doit.pl gives the following: |
1492 | |
1493 | # perl 4 prints: 3 (aborts the subroutine early) |
54310121 |
1494 | # perl 5 prints: 8 |
6dbacca0 |
1495 | |
1496 | Same behavior if you replace C<do> with C<require>. |
1497 | |
5db417f7 |
1498 | =item * C<split> on empty string with LIMIT specified |
1499 | |
1500 | $string = ''; |
1501 | @list = split(/foo/, $string, 2) |
1502 | |
1503 | Perl4 returns a one element list containing the empty string but Perl5 |
1504 | returns an empty list. |
1505 | |
6dbacca0 |
1506 | =back |
1507 | |
54310121 |
1508 | As always, if any of these are ever officially declared as bugs, |
6dbacca0 |
1509 | they'll be fixed and removed. |
1510 | |