version 2, including all changes.
.
Rev |
Author |
# |
Line |
1 |
perry |
1 |
PERLTRAP |
|
|
2 |
!!!PERLTRAP |
|
|
3 |
NAME |
|
|
4 |
DESCRIPTION |
|
|
5 |
---- |
|
|
6 |
!!NAME |
|
|
7 |
|
|
|
8 |
|
|
|
9 |
perltrap - Perl traps for the unwary |
|
|
10 |
!!DESCRIPTION |
|
|
11 |
|
|
|
12 |
|
|
|
13 |
The biggest trap of all is forgetting to use |
|
|
14 |
warnings or use the __-w__ switch; see perllexwarn |
|
|
15 |
and perlrun. The second biggest trap is not making your |
|
|
16 |
entire program runnable under use strict. The third |
|
|
17 |
biggest trap is not reading the list of changes in this |
|
|
18 |
version of Perl; see perldelta. |
|
|
19 |
|
|
|
20 |
|
|
|
21 |
__Awk Traps__ |
|
|
22 |
|
|
|
23 |
|
|
|
24 |
Accustomed __awk__ users should take special note of the |
|
|
25 |
following: |
|
|
26 |
|
|
|
27 |
|
|
|
28 |
The English module, loaded via |
|
|
29 |
|
|
|
30 |
|
|
|
31 |
use English; |
|
|
32 |
allows you to refer to special variables (like $/) with names (like $RS), as though they were in __awk__; see perlvar for details. |
|
|
33 |
|
|
|
34 |
|
|
|
35 |
Semicolons are required after all simple statements in Perl |
|
|
36 |
(except at the end of a block). Newline is not a statement |
|
|
37 |
delimiter. |
|
|
38 |
|
|
|
39 |
|
|
|
40 |
Curly brackets are required on ifs and |
|
|
41 |
whiles. |
|
|
42 |
|
|
|
43 |
|
|
|
44 |
Variables begin with ``$'', ``@'' or ``%'' in |
|
|
45 |
Perl. |
|
|
46 |
|
|
|
47 |
|
|
|
48 |
Arrays index from 0. Likewise string positions in |
|
|
49 |
''substr()'' and ''index()''. |
|
|
50 |
|
|
|
51 |
|
|
|
52 |
You have to decide whether your array has numeric or string |
|
|
53 |
indices. |
|
|
54 |
|
|
|
55 |
|
|
|
56 |
Hash values do not spring into existence upon mere |
|
|
57 |
reference. |
|
|
58 |
|
|
|
59 |
|
|
|
60 |
You have to decide whether you want to use string or numeric |
|
|
61 |
comparisons. |
|
|
62 |
|
|
|
63 |
|
|
|
64 |
Reading an input line does not split it for you. You get to |
|
|
65 |
split it to an array yourself. And the ''split()'' |
|
|
66 |
operator has different arguments than |
|
|
67 |
__awk__'s. |
|
|
68 |
|
|
|
69 |
|
|
|
70 |
The current input line is normally in $_, not |
|
|
71 |
$0. It generally does not have the newline |
|
|
72 |
stripped. ($0 is the name of the program executed.) See |
|
|
73 |
perlvar. |
|
|
74 |
|
|
|
75 |
|
|
|
76 |
$digit'' |
|
|
77 |
'' |
|
|
78 |
|
|
|
79 |
|
|
|
80 |
The ''print()'' statement does not add field and record |
|
|
81 |
separators unless you set $, and $\. You |
|
|
82 |
can set $OFS and $ORS if you're using the |
|
|
83 |
English module. |
|
|
84 |
|
|
|
85 |
|
|
|
86 |
You must open your files before you print to |
|
|
87 |
them. |
|
|
88 |
|
|
|
89 |
|
|
|
90 |
The range operator is ``..'', not comma. The comma operator |
|
|
91 |
works as in C. |
|
|
92 |
|
|
|
93 |
|
|
|
94 |
The match operator is ``=~'', not ``~''. (``~'' is the one's |
|
|
95 |
complement operator, as in C.) |
|
|
96 |
|
|
|
97 |
|
|
|
98 |
The exponentiation operator is ``**'', not ``^''. ``^'' is |
|
|
99 |
the XOR operator, as in C. (You know, one |
|
|
100 |
could get the feeling that __awk__ is basically |
|
|
101 |
incompatible with C.) |
|
|
102 |
|
|
|
103 |
|
|
|
104 |
The concatenation operator is ``.'', not the null string. |
|
|
105 |
(Using the null string would render /pat/ /pat/ |
|
|
106 |
unparsable, because the third slash would be interpreted as |
|
|
107 |
a division operator--the tokenizer is in fact slightly |
|
|
108 |
context sensitive for operators like ``/'', ``?'', and |
|
|
109 |
`` |
|
|
110 |
|
|
|
111 |
|
|
|
112 |
The next, exit, and continue |
|
|
113 |
keywords work differently. |
|
|
114 |
|
|
|
115 |
|
|
|
116 |
The following variables work differently: |
|
|
117 |
|
|
|
118 |
|
|
|
119 |
Awk Perl |
|
|
120 |
ARGC scalar @ARGV (compare with $#ARGV) |
|
|
121 |
ARGV[[0] $0 |
|
|
122 |
FILENAME $ARGV |
|
|
123 |
FNR $. - something |
|
|
124 |
FS (whatever you like) |
|
|
125 |
NF $#Fld, or some such |
|
|
126 |
NR $. |
|
|
127 |
OFMT $# |
|
|
128 |
OFS $, |
|
|
129 |
ORS $\ |
|
|
130 |
RLENGTH length($ |
|
|
131 |
|
|
|
132 |
|
|
|
133 |
You cannot set $RS to a pattern, only a |
|
|
134 |
string. |
|
|
135 |
|
|
|
136 |
|
|
|
137 |
When in doubt, run the __awk__ construct through |
|
|
138 |
__a2p__ and see what it gives you. |
|
|
139 |
|
|
|
140 |
|
|
|
141 |
__C Traps__ |
|
|
142 |
|
|
|
143 |
|
|
|
144 |
Cerebral C programmers should take note of the |
|
|
145 |
following: |
|
|
146 |
|
|
|
147 |
|
|
|
148 |
Curly brackets are required on if's and |
|
|
149 |
while's. |
|
|
150 |
|
|
|
151 |
|
|
|
152 |
You must use elsif rather than else |
|
|
153 |
if. |
|
|
154 |
|
|
|
155 |
|
|
|
156 |
The break and continue keywords from C |
|
|
157 |
become in Perl last and next, |
|
|
158 |
respectively. Unlike in C, these do ''not'' work within a |
|
|
159 |
do { } while construct. |
|
|
160 |
|
|
|
161 |
|
|
|
162 |
There's no switch statement. (But it's easy to build one on |
|
|
163 |
the fly.) |
|
|
164 |
|
|
|
165 |
|
|
|
166 |
Variables begin with ``$'', ``@'' or ``%'' in |
|
|
167 |
Perl. |
|
|
168 |
|
|
|
169 |
|
|
|
170 |
Comments begin with ``#'', not ``/*''. |
|
|
171 |
|
|
|
172 |
|
|
|
173 |
You can't take the address of anything, although a similar |
|
|
174 |
operator in Perl is the backslash, which creates a |
|
|
175 |
reference. |
|
|
176 |
|
|
|
177 |
|
|
|
178 |
ARGV must be capitalized. $ARGV[[0] is C's |
|
|
179 |
argv[[1], and argv[[0] ends up in |
|
|
180 |
$0. |
|
|
181 |
|
|
|
182 |
|
|
|
183 |
System calls such as ''link()'', ''unlink()'', |
|
|
184 |
''rename()'', etc. return nonzero for success, not 0. |
|
|
185 |
(''system()'', however, returns zero for |
|
|
186 |
success.) |
|
|
187 |
|
|
|
188 |
|
|
|
189 |
Signal handlers deal with signal names, not numbers. Use |
|
|
190 |
kill -l to find their names on your |
|
|
191 |
system. |
|
|
192 |
|
|
|
193 |
|
|
|
194 |
__Sed Traps__ |
|
|
195 |
|
|
|
196 |
|
|
|
197 |
Seasoned __sed__ programmers should take note of the |
|
|
198 |
following: |
|
|
199 |
|
|
|
200 |
|
|
|
201 |
Backreferences in substitutions use ``$'' rather than |
|
|
202 |
``''. |
|
|
203 |
|
|
|
204 |
|
|
|
205 |
The pattern matching metacharacters ``('', ``)'', and ``'' |
|
|
206 |
do not have backslashes in front. |
|
|
207 |
|
|
|
208 |
|
|
|
209 |
The range operator is ..., rather than |
|
|
210 |
comma. |
|
|
211 |
|
|
|
212 |
|
|
|
213 |
__Shell Traps__ |
|
|
214 |
|
|
|
215 |
|
|
|
216 |
Sharp shell programmers should take note of the |
|
|
217 |
following: |
|
|
218 |
|
|
|
219 |
|
|
|
220 |
The backtick operator does variable interpolation without |
|
|
221 |
regard to the presence of single quotes in the |
|
|
222 |
command. |
|
|
223 |
|
|
|
224 |
|
|
|
225 |
The backtick operator does no translation of the return |
|
|
226 |
value, unlike __csh__. |
|
|
227 |
|
|
|
228 |
|
|
|
229 |
Shells (especially __csh__) do several levels of |
|
|
230 |
substitution on each command line. Perl does substitution in |
|
|
231 |
only certain constructs such as double quotes, backticks, |
|
|
232 |
angle brackets, and search patterns. |
|
|
233 |
|
|
|
234 |
|
|
|
235 |
Shells interpret scripts a little bit at a time. Perl |
|
|
236 |
compiles the entire program before executing it (except for |
|
|
237 |
BEGIN blocks, which execute at compile |
|
|
238 |
time). |
|
|
239 |
|
|
|
240 |
|
|
|
241 |
The arguments are available via @ARGV, not |
|
|
242 |
$1, $2, etc. |
|
|
243 |
|
|
|
244 |
|
|
|
245 |
The environment is not automatically made available as |
|
|
246 |
separate scalar variables. |
|
|
247 |
|
|
|
248 |
|
|
|
249 |
__Perl Traps__ |
|
|
250 |
|
|
|
251 |
|
|
|
252 |
Practicing Perl Programmers should take note of the |
|
|
253 |
following: |
|
|
254 |
|
|
|
255 |
|
|
|
256 |
Remember that many operations behave differently in a list |
|
|
257 |
context than they do in a scalar one. See perldata for |
|
|
258 |
details. |
|
|
259 |
|
|
|
260 |
|
|
|
261 |
Avoid barewords if you can, especially all lowercase ones. |
|
|
262 |
You can't tell by just looking at it whether a bareword is a |
|
|
263 |
function or a string. By using quotes on strings and |
|
|
264 |
parentheses on function calls, you won't ever get them |
|
|
265 |
confused. |
|
|
266 |
|
|
|
267 |
|
|
|
268 |
You cannot discern from mere inspection which builtins are |
|
|
269 |
unary operators (like ''chop()'' and ''chdir()'') and |
|
|
270 |
which are list operators (like ''print()'' and |
|
|
271 |
''unlink()''). (Unless prototyped, user-defined |
|
|
272 |
subroutines can __only__ be list operators, never unary |
|
|
273 |
ones.) See perlop and perlsub. |
|
|
274 |
|
|
|
275 |
|
|
|
276 |
People have a hard time remembering that some functions |
|
|
277 |
default to $_, or @ARGV, or whatever, but |
|
|
278 |
that others which you might expect to do not. |
|
|
279 |
|
|
|
280 |
|
|
|
281 |
The FH |
|
|
282 |
$_ only if the file |
|
|
283 |
read is the sole condition in a while loop: |
|
|
284 |
|
|
|
285 |
|
|
|
286 |
while ( |
|
|
287 |
|
|
|
288 |
|
|
|
289 |
Remember not to use = when you need =~; |
|
|
290 |
these two constructs are quite different: |
|
|
291 |
|
|
|
292 |
|
|
|
293 |
$x = /foo/; |
|
|
294 |
$x =~ /foo/; |
|
|
295 |
|
|
|
296 |
|
|
|
297 |
The do {} construct isn't a real loop that you can |
|
|
298 |
use loop control on. |
|
|
299 |
|
|
|
300 |
|
|
|
301 |
Use my() for local variables whenever you can get |
|
|
302 |
away with it (but see perlform for where you can't). Using |
|
|
303 |
local() actually gives a local value to a global |
|
|
304 |
variable, which leaves you open to unforeseen side-effects |
|
|
305 |
of dynamic scoping. |
|
|
306 |
|
|
|
307 |
|
|
|
308 |
If you localize an exported variable in a module, its |
|
|
309 |
exported value will not change. The local name becomes an |
|
|
310 |
alias to a new value but the external name is still an alias |
|
|
311 |
for the original. |
|
|
312 |
|
|
|
313 |
|
|
|
314 |
__Perl4 to Perl5 Traps__ |
|
|
315 |
|
|
|
316 |
|
|
|
317 |
Practicing Perl4 Programmers should take note of the |
|
|
318 |
following Perl4-to-Perl5 specific traps. |
|
|
319 |
|
|
|
320 |
|
|
|
321 |
They're crudely ordered according to the following |
|
|
322 |
list: |
|
|
323 |
|
|
|
324 |
|
2 |
perry |
325 |
Discontinuance, Deprecation, and !BugFix traps |
1 |
perry |
326 |
|
|
|
327 |
|
|
|
328 |
Anything that's been fixed as a perl4 bug, removed as a |
|
|
329 |
perl4 feature or deprecated as a perl4 feature with the |
|
|
330 |
intent to encourage usage of some other perl5 |
|
|
331 |
feature. |
|
|
332 |
|
|
|
333 |
|
|
|
334 |
Parsing Traps |
|
|
335 |
|
|
|
336 |
|
|
|
337 |
Traps that appear to stem from the new parser. |
|
|
338 |
|
|
|
339 |
|
|
|
340 |
Numerical Traps |
|
|
341 |
|
|
|
342 |
|
|
|
343 |
Traps having to do with numerical or mathematical |
|
|
344 |
operators. |
|
|
345 |
|
|
|
346 |
|
|
|
347 |
General data type traps |
|
|
348 |
|
|
|
349 |
|
|
|
350 |
Traps involving perl standard data types. |
|
|
351 |
|
|
|
352 |
|
|
|
353 |
Context Traps - scalar, list contexts |
|
|
354 |
|
|
|
355 |
|
|
|
356 |
Traps related to context within lists, scalar |
|
|
357 |
statements/declarations. |
|
|
358 |
|
|
|
359 |
|
|
|
360 |
Precedence Traps |
|
|
361 |
|
|
|
362 |
|
|
|
363 |
Traps related to the precedence of parsing, evaluation, and |
|
|
364 |
execution of code. |
|
|
365 |
|
|
|
366 |
|
|
|
367 |
General Regular Expression Traps using s///, |
|
|
368 |
etc. |
|
|
369 |
|
|
|
370 |
|
|
|
371 |
Traps related to the use of pattern matching. |
|
|
372 |
|
|
|
373 |
|
|
|
374 |
Subroutine, Signal, Sorting Traps |
|
|
375 |
|
|
|
376 |
|
|
|
377 |
Traps related to the use of signals and signal handlers, |
|
|
378 |
general subroutines, and sorting, along with sorting |
|
|
379 |
subroutines. |
|
|
380 |
|
|
|
381 |
|
|
|
382 |
OS Traps |
|
|
383 |
|
|
|
384 |
|
|
|
385 |
OS-specific traps. |
|
|
386 |
|
|
|
387 |
|
|
|
388 |
DBM Traps |
|
|
389 |
|
|
|
390 |
|
|
|
391 |
Traps specific to the use of dbmopen(), and |
|
|
392 |
specific dbm implementations. |
|
|
393 |
|
|
|
394 |
|
|
|
395 |
Unclassified Traps |
|
|
396 |
|
|
|
397 |
|
|
|
398 |
Everything else. |
|
|
399 |
|
|
|
400 |
|
|
|
401 |
If you find an example of a conversion trap that is not |
|
|
402 |
listed here, please submit it to |
|
|
403 |
perlbug@perl.org'' |
|
|
404 |
''use |
|
|
405 |
warnings pragma or the __-w__ switch. |
|
|
406 |
|
|
|
407 |
|
2 |
perry |
408 |
__Discontinuance, Deprecation, and !BugFix |
1 |
perry |
409 |
traps__ |
|
|
410 |
|
|
|
411 |
|
|
|
412 |
Anything that has been discontinued, deprecated, or fixed as |
|
|
413 |
a bug from perl4. |
|
|
414 |
|
|
|
415 |
|
|
|
416 |
Discontinuance |
|
|
417 |
|
|
|
418 |
|
|
|
419 |
Symbols starting with ``_'' are no longer forced into |
|
|
420 |
package main, except for $_ itself (and |
|
|
421 |
@_, etc.). |
|
|
422 |
|
|
|
423 |
|
|
|
424 |
package test; |
|
|
425 |
$_legacy = 1; |
|
|
426 |
package main; |
|
|
427 |
print |
|
|
428 |
# perl4 prints: $_legacy is 1 |
|
|
429 |
# perl5 prints: $_legacy is |
|
|
430 |
|
|
|
431 |
|
|
|
432 |
Deprecation |
|
|
433 |
|
|
|
434 |
|
|
|
435 |
Double-colon is now a valid package separator in a variable |
|
|
436 |
name. Thus these behave differently in perl4 vs. perl5, |
|
|
437 |
because the packages don't exist. |
|
|
438 |
|
|
|
439 |
|
|
|
440 |
$a=1;$b=2;$c=3;$var=4; |
|
|
441 |
print |
|
|
442 |
# perl4 prints: 1::2::3 4::abc::xyz |
|
|
443 |
# perl5 prints: 3 |
|
|
444 |
Given that :: is now the preferred package delimiter, it is debatable whether this should be classed as a bug or not. (The older package delimiter, ' ,is used here) |
|
|
445 |
|
|
|
446 |
|
|
|
447 |
$x = 10 ; |
|
|
448 |
print |
|
|
449 |
# perl4 prints: x=10 |
|
|
450 |
# perl5 prints: Can't find string terminator |
|
|
451 |
You can avoid this problem, and remain compatible with perl4, if you always explicitly include the package name: |
|
|
452 |
|
|
|
453 |
|
|
|
454 |
$x = 10 ; |
|
|
455 |
print |
|
|
456 |
Also see precedence traps, for parsing $:. |
|
|
457 |
|
|
|
458 |
|
2 |
perry |
459 |
!BugFix |
1 |
perry |
460 |
|
|
|
461 |
|
|
|
462 |
The second and third arguments of splice() are now |
|
|
463 |
evaluated in scalar context (as the Camel says) rather than |
|
|
464 |
list context. |
|
|
465 |
|
|
|
466 |
|
|
|
467 |
sub sub1{return(0,2) } # return a 2-element list |
|
|
468 |
sub sub2{ return(1,2,3)} # return a 3-element list |
|
|
469 |
@a1 = ( |
|
|
470 |
# perl4 prints: a b |
|
|
471 |
# perl5 prints: c d e |
|
|
472 |
|
|
|
473 |
|
|
|
474 |
Discontinuance |
|
|
475 |
|
|
|
476 |
|
|
|
477 |
You can't do a goto into a block that is optimized |
|
|
478 |
away. Darn. |
|
|
479 |
|
|
|
480 |
|
|
|
481 |
goto marker1; |
|
|
482 |
for(1){ |
|
|
483 |
marker1: |
|
|
484 |
print |
|
|
485 |
# perl4 prints: Here I is! |
|
|
486 |
# perl5 errors: Can't |
|
|
487 |
|
|
|
488 |
|
|
|
489 |
Discontinuance |
|
|
490 |
|
|
|
491 |
|
|
|
492 |
It is no longer syntactically legal to use whitespace as the |
|
|
493 |
name of a variable, or as a delimiter for any kind of quote |
|
|
494 |
construct. Double darn. |
|
|
495 |
|
|
|
496 |
|
|
|
497 |
$a = ( |
|
|
498 |
# perl4 prints: a is foo bar, b is baz |
|
|
499 |
# perl5 errors: Bareword found where operator expected |
|
|
500 |
|
|
|
501 |
|
|
|
502 |
Discontinuance |
|
|
503 |
|
|
|
504 |
|
|
|
505 |
The archaic while/if BLOCK BLOCK syntax is no |
|
|
506 |
longer supported. |
|
|
507 |
|
|
|
508 |
|
|
|
509 |
if { 1 } { |
|
|
510 |
print |
|
|
511 |
# perl4 prints: True! |
|
|
512 |
# perl5 errors: syntax error at test.pl line 1, near |
|
|
513 |
|
|
|
514 |
|
2 |
perry |
515 |
!BugFix |
1 |
perry |
516 |
|
|
|
517 |
|
|
|
518 |
The ** operator now binds more tightly than unary |
|
|
519 |
minus. It was documented to work this way before, but |
|
|
520 |
didn't. |
|
|
521 |
|
|
|
522 |
|
|
|
523 |
print -4**2, |
|
|
524 |
# perl4 prints: 16 |
|
|
525 |
# perl5 prints: -16 |
|
|
526 |
|
|
|
527 |
|
|
|
528 |
Discontinuance |
|
|
529 |
|
|
|
530 |
|
|
|
531 |
The meaning of foreach{} has changed slightly when |
|
|
532 |
it is iterating over a list which is not an array. This used |
|
|
533 |
to assign the list to a temporary array, but no longer does |
|
|
534 |
so (for efficiency). This means that you'll now be iterating |
|
|
535 |
over the actual values, not over copies of the values. |
|
|
536 |
Modifications to the loop variable can change the original |
|
|
537 |
values. |
|
|
538 |
|
|
|
539 |
|
|
|
540 |
@list = ('ab','abc','bcd','def'); |
|
|
541 |
foreach $var (grep(/ab/,@list)){ |
|
|
542 |
$var = 1; |
|
|
543 |
} |
|
|
544 |
print (join(':',@list)); |
|
|
545 |
# perl4 prints: ab:abc:bcd:def |
|
|
546 |
# perl5 prints: 1:1:bcd:def |
|
|
547 |
To retain Perl4 semantics you need to assign your list explicitly to a temporary array and then iterate over that. For example, you might need to change |
|
|
548 |
|
|
|
549 |
|
|
|
550 |
foreach $var (grep(/ab/,@list)){ |
|
|
551 |
to |
|
|
552 |
|
|
|
553 |
|
|
|
554 |
foreach $var (@tmp = grep(/ab/,@list)){ |
|
|
555 |
Otherwise changing $var will clobber the values of @list. (This most often happens when you use $_ for the loop variable, and call subroutines in the loop that don't properly localize $_.) |
|
|
556 |
|
|
|
557 |
|
|
|
558 |
Discontinuance |
|
|
559 |
|
|
|
560 |
|
|
|
561 |
split with no arguments now behaves like split |
|
|
562 |
' ' (which doesn't return an initial null field if |
|
|
563 |
$_ starts with whitespace), it used to behave like |
|
|
564 |
split /s+/ (which does). |
|
|
565 |
|
|
|
566 |
|
|
|
567 |
$_ = ' hi mom'; |
|
|
568 |
print join(':', split); |
|
|
569 |
# perl4 prints: :hi:mom |
|
|
570 |
# perl5 prints: hi:mom |
|
|
571 |
|
|
|
572 |
|
2 |
perry |
573 |
!BugFix |
1 |
perry |
574 |
|
|
|
575 |
|
|
|
576 |
Perl 4 would ignore any text which was attached to an |
|
|
577 |
__-e__ switch, always taking the code snippet from the |
|
|
578 |
following arg. Additionally, it would silently accept an |
|
|
579 |
__-e__ switch without a following arg. Both of these |
|
|
580 |
behaviors have been fixed. |
|
|
581 |
|
|
|
582 |
|
|
|
583 |
perl -e'print |
|
|
584 |
# perl4 prints: separate arg |
|
|
585 |
# perl5 prints: attached to -e |
|
|
586 |
perl -e |
|
|
587 |
# perl4 prints: |
|
|
588 |
# perl5 dies: No code specified for -e. |
|
|
589 |
|
|
|
590 |
|
|
|
591 |
Discontinuance |
|
|
592 |
|
|
|
593 |
|
|
|
594 |
In Perl 4 the return value of push was |
|
|
595 |
undocumented, but it was actually the last value being |
|
|
596 |
pushed onto the target list. In Perl 5 the return value of |
|
|
597 |
push is documented, but has changed, it is the |
|
|
598 |
number of elements in the resulting list. |
|
|
599 |
|
|
|
600 |
|
|
|
601 |
@x = ('existing'); |
|
|
602 |
print push(@x, 'first new', 'second new'); |
|
|
603 |
# perl4 prints: second new |
|
|
604 |
# perl5 prints: 3 |
|
|
605 |
|
|
|
606 |
|
|
|
607 |
Deprecation |
|
|
608 |
|
|
|
609 |
|
|
|
610 |
Some error messages will be different. |
|
|
611 |
|
|
|
612 |
|
|
|
613 |
Discontinuance |
|
|
614 |
|
|
|
615 |
|
|
|
616 |
In Perl 4, if in list context the delimiters to the first |
|
|
617 |
argument of split() were ??, the result |
|
|
618 |
would be placed in @_ as well as being returned. |
|
|
619 |
Perl 5 has more respect for your subroutine |
|
|
620 |
arguments. |
|
|
621 |
|
|
|
622 |
|
|
|
623 |
Discontinuance |
|
|
624 |
|
|
|
625 |
|
|
|
626 |
Some bugs may have been inadvertently removed. |
|
|
627 |
:-) |
|
|
628 |
|
|
|
629 |
|
|
|
630 |
__Parsing Traps__ |
|
|
631 |
|
|
|
632 |
|
|
|
633 |
Perl4-to-Perl5 traps from having to do with |
|
|
634 |
parsing. |
|
|
635 |
|
|
|
636 |
|
|
|
637 |
Parsing |
|
|
638 |
|
|
|
639 |
|
|
|
640 |
Note the space between . and = |
|
|
641 |
|
|
|
642 |
|
|
|
643 |
$string . = |
|
|
644 |
# perl4 prints: more string |
|
|
645 |
# perl5 prints: syntax error at - line 1, near |
|
|
646 |
|
|
|
647 |
|
|
|
648 |
Parsing |
|
|
649 |
|
|
|
650 |
|
|
|
651 |
Better parsing in perl 5 |
|
|
652 |
|
|
|
653 |
|
|
|
654 |
sub foo {} |
|
|
655 |
# perl4 prints: hello, world |
|
|
656 |
# perl5 prints: syntax error |
|
|
657 |
|
|
|
658 |
|
|
|
659 |
Parsing |
|
|
660 |
|
|
|
661 |
|
|
|
662 |
``if it looks like a function, it is a function'' |
|
|
663 |
rule. |
|
|
664 |
|
|
|
665 |
|
|
|
666 |
print |
|
|
667 |
($foo == 1) ? |
|
|
668 |
# perl4 prints: is zero |
|
|
669 |
# perl5 warns: |
|
|
670 |
|
|
|
671 |
|
|
|
672 |
Parsing |
|
|
673 |
|
|
|
674 |
|
|
|
675 |
String interpolation of the $#array construct |
|
|
676 |
differs when braces are to used around the |
|
|
677 |
name. |
|
|
678 |
|
|
|
679 |
|
|
|
680 |
@a = (1..3); |
|
|
681 |
print |
|
|
682 |
# perl4 prints: 2 |
|
|
683 |
# perl5 fails with syntax error |
|
|
684 |
@ = (1..3); |
|
|
685 |
print |
|
|
686 |
# perl4 prints: {a} |
|
|
687 |
# perl5 prints: 2 |
|
|
688 |
|
|
|
689 |
|
|
|
690 |
__Numerical Traps__ |
|
|
691 |
|
|
|
692 |
|
|
|
693 |
Perl4-to-Perl5 traps having to do with numerical operators, |
|
|
694 |
operands, or output from same. |
|
|
695 |
|
|
|
696 |
|
|
|
697 |
Numerical |
|
|
698 |
|
|
|
699 |
|
|
|
700 |
Formatted output and significant digits |
|
|
701 |
|
|
|
702 |
|
|
|
703 |
print 7.373504 - 0, |
|
|
704 |
# Perl4 prints: |
|
|
705 |
7.375039999999996141 |
|
|
706 |
7.37503999999999614 |
|
|
707 |
# Perl5 prints: |
|
|
708 |
7.373504 |
|
|
709 |
7.37503999999999614 |
|
|
710 |
|
|
|
711 |
|
|
|
712 |
Numerical |
|
|
713 |
|
|
|
714 |
|
|
|
715 |
This specific item has been deleted. It demonstrated how the |
|
|
716 |
auto-increment operator would not catch when a number went |
|
|
717 |
over the signed int limit. Fixed in version 5.003_04. But |
|
|
718 |
always be wary when using large integers. If in |
|
|
719 |
doubt: |
|
|
720 |
|
|
|
721 |
|
2 |
perry |
722 |
use Math::!BigInt; |
1 |
perry |
723 |
|
|
|
724 |
|
|
|
725 |
Numerical |
|
|
726 |
|
|
|
727 |
|
|
|
728 |
Assignment of return values from numeric equality tests does |
|
|
729 |
not work in perl5 when the test evaluates to false (0). |
|
|
730 |
Logical tests now return an null, instead of 0 |
|
|
731 |
|
|
|
732 |
|
|
|
733 |
$p = ($test == 1); |
|
|
734 |
print $p, |
|
|
735 |
# perl4 prints: 0 |
|
|
736 |
# perl5 prints: |
|
|
737 |
Also see ``General Regular Expression Traps using s///, etc.'' for another example of this new feature... |
|
|
738 |
|
|
|
739 |
|
|
|
740 |
Bitwise string ops |
|
|
741 |
|
|
|
742 |
|
|
|
743 |
When bitwise operators which can operate upon either numbers |
|
|
744 |
or strings () are given only strings as |
|
|
745 |
arguments, perl4 would treat the operands as bitstrings so |
|
|
746 |
long as the program contained a call to the vec() |
|
|
747 |
function. perl5 treats the string operands as bitstrings. |
|
|
748 |
(See ``Bitwise String Operators'' in perlop for more |
|
|
749 |
details.) |
|
|
750 |
|
|
|
751 |
|
|
|
752 |
$fred = |
|
|
753 |
# Perl4 prints: |
|
|
754 |
8 |
|
|
755 |
# Perl5 prints: |
|
|
756 |
10 |
|
|
757 |
# If vec() is used anywhere in the program, both print: |
|
|
758 |
10 |
|
|
759 |
|
|
|
760 |
|
|
|
761 |
__General data type traps__ |
|
|
762 |
|
|
|
763 |
|
|
|
764 |
Perl4-to-Perl5 traps involving most data-types, and their |
|
|
765 |
usage within certain expressions and/or |
|
|
766 |
context. |
|
|
767 |
|
|
|
768 |
|
|
|
769 |
(Arrays) |
|
|
770 |
|
|
|
771 |
|
|
|
772 |
Negative array subscripts now count from the end of the |
|
|
773 |
array. |
|
|
774 |
|
|
|
775 |
|
|
|
776 |
@a = (1, 2, 3, 4, 5); |
|
|
777 |
print |
|
|
778 |
# perl4 prints: The third element of the array is 4 also expressed as |
|
|
779 |
# perl5 prints: The third element of the array is 4 also expressed as 4 |
|
|
780 |
|
|
|
781 |
|
|
|
782 |
(Arrays) |
|
|
783 |
|
|
|
784 |
|
|
|
785 |
Setting $#array lower now discards array elements, |
|
|
786 |
and makes them impossible to recover. |
|
|
787 |
|
|
|
788 |
|
|
|
789 |
@a = (a,b,c,d,e); |
|
|
790 |
print |
|
|
791 |
# perl4 prints: Before: abcde, After: ab, Recovered: abcd |
|
|
792 |
# perl5 prints: Before: abcde, After: ab, Recovered: ab |
|
|
793 |
|
|
|
794 |
|
|
|
795 |
(Hashes) |
|
|
796 |
|
|
|
797 |
|
|
|
798 |
Hashes get defined before use |
|
|
799 |
|
|
|
800 |
|
|
|
801 |
local($s,@a,%h); |
|
|
802 |
die |
|
|
803 |
# perl4 prints: |
|
|
804 |
# perl5 dies: hash %h defined |
|
|
805 |
Perl will now generate a warning when it sees defined(@a) and defined(%h). |
|
|
806 |
|
|
|
807 |
|
|
|
808 |
(Globs) |
|
|
809 |
|
|
|
810 |
|
|
|
811 |
glob assignment from variable to variable will fail if the |
|
|
812 |
assigned variable is localized subsequent to the |
|
|
813 |
assignment |
|
|
814 |
|
|
|
815 |
|
|
|
816 |
@a = ( |
|
|
817 |
# perl4 prints: This is Perl 4 |
|
|
818 |
# perl5 prints: |
|
|
819 |
|
|
|
820 |
|
|
|
821 |
(Globs) |
|
|
822 |
|
|
|
823 |
|
|
|
824 |
Assigning undef to a glob has no effect in Perl 5. |
|
|
825 |
In Perl 4 it undefines the associated scalar (but may have |
|
|
826 |
other side effects including SEGVs). Perl 5 will also warn |
|
|
827 |
if undef is assigned to a typeglob. (Note that |
|
|
828 |
assigning undef to a typeglob is different than |
|
|
829 |
calling the undef function on a typeglob (undef |
|
|
830 |
*foo), which has quite a few effects. |
|
|
831 |
|
|
|
832 |
|
|
|
833 |
$foo = |
|
|
834 |
# perl4 prints: |
|
|
835 |
# perl4 warns: |
|
|
836 |
|
|
|
837 |
|
|
|
838 |
(Scalar String) |
|
|
839 |
|
|
|
840 |
|
|
|
841 |
Changes in unary negation (of strings) This change effects |
|
|
842 |
both the return value and what it does to |
|
|
843 |
auto(magic)increment. |
|
|
844 |
|
|
|
845 |
|
|
|
846 |
$x = |
|
|
847 |
# perl4 prints: aab : -0 : 1 |
|
|
848 |
# perl5 prints: aab : -aab : aac |
|
|
849 |
|
|
|
850 |
|
|
|
851 |
(Constants) |
|
|
852 |
|
|
|
853 |
|
|
|
854 |
perl 4 lets you modify constants: |
|
|
855 |
|
|
|
856 |
|
|
|
857 |
$foo = |
|
|
858 |
# perl4: |
|
|
859 |
# before: x after: m |
|
|
860 |
# before: a after: m |
|
|
861 |
# before: m after: m |
|
|
862 |
# before: m after: m |
|
|
863 |
# Perl5: |
|
|
864 |
# before: x after: m |
|
|
865 |
# Modification of a read-only value attempted at foo.pl line 12. |
|
|
866 |
# before: a |
|
|
867 |
|
|
|
868 |
|
|
|
869 |
(Scalars) |
|
|
870 |
|
|
|
871 |
|
|
|
872 |
The behavior is slightly different for: |
|
|
873 |
|
|
|
874 |
|
|
|
875 |
print |
|
|
876 |
# perl 4: 1 |
|
|
877 |
# perl 5: |
|
|
878 |
|
|
|
879 |
|
|
|
880 |
(Variable Suicide) |
|
|
881 |
|
|
|
882 |
|
|
|
883 |
Variable suicide behavior is more consistent under Perl 5. |
|
|
884 |
Perl5 exhibits the same behavior for hashes and scalars, |
|
|
885 |
that perl4 exhibits for only scalars. |
|
|
886 |
|
|
|
887 |
|
|
|
888 |
$aGlobal{ |
|
|
889 |
sub test { |
|
|
890 |
local( *theArgument ) = @_; |
|
|
891 |
local( %aNewLocal ); # perl 4 != 5.001l,m |
|
|
892 |
$aNewLocal{ |
|
|
893 |
# Perl4: |
|
|
894 |
# MAIN:global value |
|
|
895 |
# SUB: global value |
|
|
896 |
# SUB: level 0 |
|
|
897 |
# SUB: level 1 |
|
|
898 |
# SUB: level 2 |
|
|
899 |
# Perl5: |
|
|
900 |
# MAIN:global value |
|
|
901 |
# SUB: global value |
|
|
902 |
# SUB: this should never appear |
|
|
903 |
# SUB: this should never appear |
|
|
904 |
# SUB: this should never appear |
|
|
905 |
|
|
|
906 |
|
|
|
907 |
__Context Traps - scalar, list contexts__ |
|
|
908 |
|
|
|
909 |
|
|
|
910 |
(list context) |
|
|
911 |
|
|
|
912 |
|
|
|
913 |
The elements of argument lists for formats are now evaluated |
|
|
914 |
in list context. This means you can interpolate list values |
|
|
915 |
now. |
|
|
916 |
|
|
|
917 |
|
|
|
918 |
@fmt = ( |
|
|
919 |
# perl4 errors: Please use commas to separate fields in file |
|
|
920 |
# perl5 prints: foo bar baz |
|
|
921 |
|
|
|
922 |
|
|
|
923 |
(scalar context) |
|
|
924 |
|
|
|
925 |
|
|
|
926 |
The caller() function now returns a false value in |
|
|
927 |
a scalar context if there is no caller. This lets library |
|
|
928 |
files determine if they're being required. |
|
|
929 |
|
|
|
930 |
|
|
|
931 |
caller() ? (print |
|
|
932 |
# perl4 errors: There is no caller |
|
|
933 |
# perl5 prints: Got a 0 |
|
|
934 |
|
|
|
935 |
|
|
|
936 |
(scalar context) |
|
|
937 |
|
|
|
938 |
|
|
|
939 |
The comma operator in a scalar context is now guaranteed to |
|
|
940 |
give a scalar context to its arguments. |
|
|
941 |
|
|
|
942 |
|
|
|
943 |
@y= ('a','b','c'); |
|
|
944 |
$x = (1, 2, @y); |
|
|
945 |
print |
|
|
946 |
# Perl4 prints: x = c # Thinks list context interpolates list |
|
|
947 |
# Perl5 prints: x = 3 # Knows scalar uses length of list |
|
|
948 |
|
|
|
949 |
|
|
|
950 |
(list, builtin) |
|
|
951 |
|
|
|
952 |
|
|
|
953 |
sprintf() is prototyped as ($;@), so its first |
|
|
954 |
argument is given scalar context. Thus, if passed an array, |
|
|
955 |
it will probably not do what you want, unlike Perl |
|
|
956 |
4: |
|
|
957 |
|
|
|
958 |
|
|
|
959 |
@z = ('%s%s', 'foo', 'bar'); |
|
|
960 |
$x = sprintf(@z); |
|
|
961 |
print $x; |
|
|
962 |
# perl4 prints: foobar |
|
|
963 |
# perl5 prints: 3 |
|
|
964 |
printf() works the same as it did in Perl 4, though: |
|
|
965 |
|
|
|
966 |
|
|
|
967 |
@z = ('%s%s', 'foo', 'bar'); |
|
|
968 |
printf STDOUT (@z); |
|
|
969 |
# perl4 prints: foobar |
|
|
970 |
# perl5 prints: foobar |
|
|
971 |
|
|
|
972 |
|
|
|
973 |
__Precedence Traps__ |
|
|
974 |
|
|
|
975 |
|
|
|
976 |
Perl4-to-Perl5 traps involving precedence |
|
|
977 |
order. |
|
|
978 |
|
|
|
979 |
|
|
|
980 |
Perl 4 has almost the same precedence rules as Perl 5 for |
|
|
981 |
the operators that they both have. Perl 4 however, seems to |
|
|
982 |
have had some inconsistencies that made the behavior differ |
|
|
983 |
from what was documented. |
|
|
984 |
|
|
|
985 |
|
|
|
986 |
Precedence |
|
|
987 |
|
|
|
988 |
|
|
|
989 |
LHS vs. RHS of any assignment |
|
|
990 |
operator. LHS is evaluated first in perl4, |
|
|
991 |
second in perl5; this can affect the relationship between |
|
|
992 |
side-effects in sub-expressions. |
|
|
993 |
|
|
|
994 |
|
|
|
995 |
@arr = ( 'left', 'right' ); |
|
|
996 |
$a{shift @arr} = shift @arr; |
|
|
997 |
print join( ' ', keys %a ); |
|
|
998 |
# perl4 prints: left |
|
|
999 |
# perl5 prints: right |
|
|
1000 |
|
|
|
1001 |
|
|
|
1002 |
Precedence |
|
|
1003 |
|
|
|
1004 |
|
|
|
1005 |
These are now semantic errors because of |
|
|
1006 |
precedence: |
|
|
1007 |
|
|
|
1008 |
|
|
|
1009 |
@list = (1,2,3,4,5); |
|
|
1010 |
%map = ( |
|
|
1011 |
# perl4 prints: n is 3, m is 6 |
|
|
1012 |
# perl5 errors and fails to compile |
|
|
1013 |
|
|
|
1014 |
|
|
|
1015 |
Precedence |
|
|
1016 |
|
|
|
1017 |
|
|
|
1018 |
The precedence of assignment operators is now the same as |
|
|
1019 |
the precedence of assignment. Perl 4 mistakenly gave them |
|
|
1020 |
the precedence of the associated operator. So you now must |
|
|
1021 |
parenthesize them in expressions like |
|
|
1022 |
|
|
|
1023 |
|
|
|
1024 |
/foo/ ? ($a += 2) : ($a -= 2); |
|
|
1025 |
Otherwise |
|
|
1026 |
|
|
|
1027 |
|
|
|
1028 |
/foo/ ? $a += 2 : $a -= 2 |
|
|
1029 |
would be erroneously parsed as |
|
|
1030 |
|
|
|
1031 |
|
|
|
1032 |
(/foo/ ? $a += 2 : $a) -= 2; |
|
|
1033 |
On the other hand, |
|
|
1034 |
|
|
|
1035 |
|
|
|
1036 |
$a += /foo/ ? 1 : 2; |
|
|
1037 |
now works as a C programmer would expect. |
|
|
1038 |
|
|
|
1039 |
|
|
|
1040 |
Precedence |
|
|
1041 |
|
|
|
1042 |
|
|
|
1043 |
open FOO die; |
|
|
1044 |
is now incorrect. You need parentheses around the filehandle. Otherwise, perl5 leaves the statement as its default precedence: |
|
|
1045 |
|
|
|
1046 |
|
|
|
1047 |
open(FOO die); |
|
|
1048 |
# perl4 opens or dies |
|
|
1049 |
# perl5 opens FOO, dying only if 'FOO' is false, i.e. never |
|
|
1050 |
|
|
|
1051 |
|
|
|
1052 |
Precedence |
|
|
1053 |
|
|
|
1054 |
|
|
|
1055 |
perl4 gives the special variable, $: precedence, |
|
|
1056 |
where perl5 treats $:: as main |
|
|
1057 |
package |
|
|
1058 |
|
|
|
1059 |
|
|
|
1060 |
$a = |
|
|
1061 |
# perl 4 prints: -:a |
|
|
1062 |
# perl 5 prints: x |
|
|
1063 |
|
|
|
1064 |
|
|
|
1065 |
Precedence |
|
|
1066 |
|
|
|
1067 |
|
|
|
1068 |
perl4 had buggy precedence for the file test operators |
|
|
1069 |
vis-a-vis the assignment operators. Thus, although the |
|
|
1070 |
precedence table for perl4 leads one to believe -e $foo |
|
|
1071 |
.= should parse as ((-e $foo) .= |
|
|
1072 |
, it actually parses as (-e ($foo .= |
|
|
1073 |
. In perl5, the precedence is as |
|
|
1074 |
documented. |
|
|
1075 |
|
|
|
1076 |
|
|
|
1077 |
-e $foo .= |
|
|
1078 |
# perl4 prints: no output |
|
|
1079 |
# perl5 prints: Can't modify -e in concatenation |
|
|
1080 |
|
|
|
1081 |
|
|
|
1082 |
Precedence |
|
|
1083 |
|
|
|
1084 |
|
|
|
1085 |
In perl4, ''keys()'', ''each()'' and ''values()'' |
|
|
1086 |
were special high-precedence operators that operated on a |
|
|
1087 |
single hash, but in perl5, they are regular named unary |
|
|
1088 |
operators. As documented, named unary operators have lower |
|
|
1089 |
precedence than the arithmetic and concatenation operators |
|
|
1090 |
+ - ., but the perl4 variants of these operators |
|
|
1091 |
actually bind tighter than + - .. Thus, |
|
|
1092 |
for: |
|
|
1093 |
|
|
|
1094 |
|
|
|
1095 |
%foo = 1..10; |
|
|
1096 |
print keys %foo - 1 |
|
|
1097 |
# perl4 prints: 4 |
|
|
1098 |
# perl5 prints: Type of arg 1 to keys must be hash (not subtraction) |
|
|
1099 |
The perl4 behavior was probably more useful, if less consistent. |
|
|
1100 |
|
|
|
1101 |
|
|
|
1102 |
__General Regular Expression Traps using s///, |
|
|
1103 |
etc.__ |
|
|
1104 |
|
|
|
1105 |
|
|
|
1106 |
All types of RE traps. |
|
|
1107 |
|
|
|
1108 |
|
|
|
1109 |
Regular Expression |
|
|
1110 |
|
|
|
1111 |
|
|
|
1112 |
s'$lhs'$rhs' now does no interpolation on either |
|
|
1113 |
side. It used to interpolate $lhs but not |
|
|
1114 |
$rhs. (And still does not match a literal '$' in |
|
|
1115 |
string) |
|
|
1116 |
|
|
|
1117 |
|
|
|
1118 |
$a=1;$b=2; |
|
|
1119 |
$string = '1 2 $a $b'; |
|
|
1120 |
$string =~ s'$a'$b'; |
|
|
1121 |
print $string, |
|
|
1122 |
# perl4 prints: $b 2 $a $b |
|
|
1123 |
# perl5 prints: 1 2 $a $b |
|
|
1124 |
|
|
|
1125 |
|
|
|
1126 |
Regular Expression |
|
|
1127 |
|
|
|
1128 |
|
|
|
1129 |
m//g now attaches its state to the searched string |
|
|
1130 |
rather than the regular expression. (Once the scope of a |
|
|
1131 |
block is left for the sub, the state of the searched string |
|
|
1132 |
is lost) |
|
|
1133 |
|
|
|
1134 |
|
|
|
1135 |
$_ = |
|
|
1136 |
# perl4 prints: Got blah Got blah Got blah Got blah |
|
|
1137 |
# perl5 prints: infinite loop blah... |
|
|
1138 |
|
|
|
1139 |
|
|
|
1140 |
Regular Expression |
|
|
1141 |
|
|
|
1142 |
|
|
|
1143 |
Currently, if you use the m//o qualifier on a |
|
|
1144 |
regular expression within an anonymous sub, ''all'' |
|
|
1145 |
closures generated from that anonymous sub will use the |
|
|
1146 |
regular expression as it was compiled when it was used the |
|
|
1147 |
very first time in any such closure. For instance, if you |
|
|
1148 |
say |
|
|
1149 |
|
|
|
1150 |
|
|
|
1151 |
sub build_match { |
|
|
1152 |
my($left,$right) = @_; |
|
|
1153 |
return sub { $_[[0] =~ /$left stuff $right/o; }; |
|
|
1154 |
} |
|
|
1155 |
$good = build_match('foo','bar'); |
|
|
1156 |
$bad = build_match('baz','blarch'); |
|
|
1157 |
print $good- |
|
|
1158 |
For most builds of Perl5, this will print: ok not ok not ok |
|
|
1159 |
|
|
|
1160 |
|
|
|
1161 |
''build_match()'' will always return a sub which matches |
|
|
1162 |
the contents of $left and $right as they |
|
|
1163 |
were the ''first'' time that ''build_match()'' was |
|
|
1164 |
called, not as they are in the current call. |
|
|
1165 |
|
|
|
1166 |
|
|
|
1167 |
Regular Expression |
|
|
1168 |
|
|
|
1169 |
|
|
|
1170 |
If no parentheses are used in a match, Perl4 sets |
|
|
1171 |
$+ to the whole match, just like $. |
|
|
1172 |
Perl5 does not. |
|
|
1173 |
|
|
|
1174 |
|
|
|
1175 |
|
|
|
1176 |
# perl4 prints: bcde |
|
|
1177 |
# perl5 prints: |
|
|
1178 |
|
|
|
1179 |
|
|
|
1180 |
Regular Expression |
|
|
1181 |
|
|
|
1182 |
|
|
|
1183 |
substitution now returns the null string if it |
|
|
1184 |
fails |
|
|
1185 |
|
|
|
1186 |
|
|
|
1187 |
$string = |
|
|
1188 |
# perl4 prints: 0 |
|
|
1189 |
# perl5 prints: |
|
|
1190 |
Also see ``Numerical Traps'' for another example of this new feature. |
|
|
1191 |
|
|
|
1192 |
|
|
|
1193 |
Regular Expression |
|
|
1194 |
|
|
|
1195 |
|
|
|
1196 |
s`lhs`rhs` (using backticks) is now a normal |
|
|
1197 |
substitution, with no backtick expansion |
|
|
1198 |
|
|
|
1199 |
|
|
|
1200 |
$string = |
|
|
1201 |
# perl4 prints: |
|
|
1202 |
|
|
|
1203 |
|
|
|
1204 |
Regular Expression |
|
|
1205 |
|
|
|
1206 |
|
|
|
1207 |
Stricter parsing of variables used in regular |
|
|
1208 |
expressions |
|
|
1209 |
|
|
|
1210 |
|
|
|
1211 |
s/^([[^$grpc]*$grpc[[$opt$plus$rep]?)//o; |
|
|
1212 |
# perl4: compiles w/o error |
|
|
1213 |
# perl5: with Scalar found where operator expected ..., near |
|
|
1214 |
an added component of this example, apparently from the same script, is the actual value of the s'd string after the substitution. [[$opt] is a character class in perl4 and an array subscript in perl5 |
|
|
1215 |
|
|
|
1216 |
|
|
|
1217 |
$grpc = 'a'; |
|
|
1218 |
$opt = 'r'; |
|
|
1219 |
$_ = 'bar'; |
|
|
1220 |
s/^([[^$grpc]*$grpc[[$opt]?)/foo/; |
|
|
1221 |
print ; |
|
|
1222 |
# perl4 prints: foo |
|
|
1223 |
# perl5 prints: foobar |
|
|
1224 |
|
|
|
1225 |
|
|
|
1226 |
Regular Expression |
|
|
1227 |
|
|
|
1228 |
|
|
|
1229 |
Under perl5, m?x? matches only once, like |
|
|
1230 |
?x?. Under perl4, it matched repeatedly, like |
|
|
1231 |
/x/ or m!x!. |
|
|
1232 |
|
|
|
1233 |
|
|
|
1234 |
$test = |
|
|
1235 |
# perl4 prints: perl4 |
|
|
1236 |
# perl5 prints: perl5 |
|
|
1237 |
|
|
|
1238 |
|
|
|
1239 |
__Subroutine, Signal, Sorting Traps__ |
|
|
1240 |
|
|
|
1241 |
|
|
|
1242 |
The general group of Perl4-to-Perl5 traps having to do with |
|
|
1243 |
Signals, Sorting, and their related subroutines, as well as |
|
|
1244 |
general subroutine traps. Includes some OS-Specific |
|
|
1245 |
traps. |
|
|
1246 |
|
|
|
1247 |
|
|
|
1248 |
(Signals) |
|
|
1249 |
|
|
|
1250 |
|
|
|
1251 |
Barewords that used to look like strings to Perl will now |
|
|
1252 |
look like subroutine calls if a subroutine by that name is |
|
|
1253 |
defined before the compiler sees them. |
|
|
1254 |
|
|
|
1255 |
|
2 |
perry |
1256 |
sub !SeeYa { warn |
|
|
1257 |
# perl4 prints: SIGTERM is now main'!SeeYa |
1 |
perry |
1258 |
# perl5 prints: SIGTERM is now main::1 (and warns |
|
|
1259 |
Use __-w__ to catch this one |
|
|
1260 |
|
|
|
1261 |
|
|
|
1262 |
(Sort Subroutine) |
|
|
1263 |
|
|
|
1264 |
|
|
|
1265 |
reverse is no longer allowed as the name of a sort |
|
|
1266 |
subroutine. |
|
|
1267 |
|
|
|
1268 |
|
|
|
1269 |
sub reverse{ print |
|
|
1270 |
# perl4 prints: yup yup 123 |
|
|
1271 |
# perl5 prints: 123 |
|
|
1272 |
# perl5 warns (if using -w): Ambiguous call resolved as CORE::reverse() |
|
|
1273 |
|
|
|
1274 |
|
|
|
1275 |
''warn()'' won't let you specify a |
|
|
1276 |
filehandle. |
|
|
1277 |
|
|
|
1278 |
|
|
|
1279 |
Although it _always_ printed to STDERR , |
|
|
1280 |
''warn()'' would let you specify a filehandle in perl4. |
|
|
1281 |
With perl5 it does not. |
|
|
1282 |
|
|
|
1283 |
|
|
|
1284 |
warn STDERR |
|
|
1285 |
# perl4 prints: Foo! |
|
|
1286 |
# perl5 prints: String found where operator expected |
|
|
1287 |
|
|
|
1288 |
|
|
|
1289 |
__OS Traps__ |
|
|
1290 |
|
|
|
1291 |
|
|
|
1292 |
(SysV) |
|
|
1293 |
|
|
|
1294 |
|
|
|
1295 |
Under HPUX , and some other SysV OSes, one |
|
|
1296 |
had to reset any signal handler, within the signal handler |
|
|
1297 |
function, each time a signal was handled with perl4. With |
|
|
1298 |
perl5, the reset is now done correctly. Any code relying on |
|
|
1299 |
the handler _not_ being reset will have to be |
|
|
1300 |
reworked. |
|
|
1301 |
|
|
|
1302 |
|
|
|
1303 |
Since version 5.002, Perl uses ''sigaction()'' under |
|
|
1304 |
SysV. |
|
|
1305 |
|
|
|
1306 |
|
|
|
1307 |
sub gotit { |
|
|
1308 |
print |
|
|
1309 |
$ = 1; |
|
|
1310 |
$pid = fork; |
|
|
1311 |
if ($pid) { |
|
|
1312 |
kill('INT', $pid); |
|
|
1313 |
sleep(1); |
|
|
1314 |
kill('INT', $pid); |
|
|
1315 |
} else { |
|
|
1316 |
while (1) {sleep(10);} |
|
|
1317 |
} |
|
|
1318 |
# perl4 (HPUX) prints: Got INT... |
|
|
1319 |
# perl5 (HPUX) prints: Got INT... Got INT... |
|
|
1320 |
|
|
|
1321 |
|
|
|
1322 |
(SysV) |
|
|
1323 |
|
|
|
1324 |
|
|
|
1325 |
Under SysV OSes, seek() on a file opened to append |
|
|
1326 |
now does the right thing w.r.t. the |
|
|
1327 |
''fopen()'' manpage. e.g., - When a file is opened for |
|
|
1328 |
append, it is impossible to overwrite information already in |
|
|
1329 |
the file. |
|
|
1330 |
|
|
|
1331 |
|
|
|
1332 |
open(TEST, |
|
|
1333 |
# perl4 (solaris) seek.test has: 18 characters here |
|
|
1334 |
# perl5 (solaris) seek.test has: 1 2 3 4 5 6 7 8 9 18 characters here |
|
|
1335 |
|
|
|
1336 |
|
|
|
1337 |
__Interpolation Traps__ |
|
|
1338 |
|
|
|
1339 |
|
|
|
1340 |
Perl4-to-Perl5 traps having to do with how things get |
|
|
1341 |
interpolated within certain expressions, statements, |
|
|
1342 |
contexts, or whatever. |
|
|
1343 |
|
|
|
1344 |
|
|
|
1345 |
Interpolation |
|
|
1346 |
|
|
|
1347 |
|
|
|
1348 |
@ now always interpolates an array in double-quotish |
|
|
1349 |
strings. |
|
|
1350 |
|
|
|
1351 |
|
|
|
1352 |
print |
|
|
1353 |
# perl4 prints: To:someone@somewhere.com |
|
|
1354 |
# perl |
|
|
1355 |
|
|
|
1356 |
|
|
|
1357 |
Interpolation |
|
|
1358 |
|
|
|
1359 |
|
|
|
1360 |
Double-quoted strings may no longer end with an unescaped $ |
|
|
1361 |
or @. |
|
|
1362 |
|
|
|
1363 |
|
|
|
1364 |
$foo = |
|
|
1365 |
# perl4 prints: foo is foo$, bar is bar@ |
|
|
1366 |
# perl5 errors: Final $ should be $ or $name |
|
|
1367 |
Note: perl5 DOES NOT error on the terminating @ in $bar |
|
|
1368 |
|
|
|
1369 |
|
|
|
1370 |
Interpolation |
|
|
1371 |
|
|
|
1372 |
|
|
|
1373 |
Perl now sometimes evaluates arbitrary expressions inside |
|
|
1374 |
braces that occur within double quotes (usually when the |
|
|
1375 |
opening brace is preceded by $ or |
|
|
1376 |
@). |
|
|
1377 |
|
|
|
1378 |
|
|
|
1379 |
@www = |
|
|
1380 |
# perl4 prints: @{w.w.w}foo |
|
|
1381 |
# perl5 prints: buzbar |
|
|
1382 |
Note that you can use strict; to ward off such trappiness under perl5. |
|
|
1383 |
|
|
|
1384 |
|
|
|
1385 |
Interpolation |
|
|
1386 |
|
|
|
1387 |
|
|
|
1388 |
The construct ``this is $$x'' used to interpolate the pid at |
|
|
1389 |
that point, but now tries to dereference $x. |
|
|
1390 |
$$ by itself still works fine, |
|
|
1391 |
however. |
|
|
1392 |
|
|
|
1393 |
|
|
|
1394 |
$s = |
|
|
1395 |
# perl4 prints: this is XXXx (XXX is the current pid) |
|
|
1396 |
# perl5 prints: this is a reference |
|
|
1397 |
|
|
|
1398 |
|
|
|
1399 |
Interpolation |
|
|
1400 |
|
|
|
1401 |
|
|
|
1402 |
Creation of hashes on the fly with eval |
|
|
1403 |
now requires either both $'s |
|
|
1404 |
to be protected in the specification of the hash name, or |
|
|
1405 |
both curlies to be protected. If both curlies are protected, |
|
|
1406 |
the result will be compatible with perl4 and perl5. This is |
|
|
1407 |
a very common practice, and should be changed to use the |
|
|
1408 |
block form of eval{} if possible. |
|
|
1409 |
|
|
|
1410 |
|
|
|
1411 |
$hashname = |
|
|
1412 |
# perl4 prints: Yup |
|
|
1413 |
# perl5 prints: Nope |
|
|
1414 |
Changing |
|
|
1415 |
|
|
|
1416 |
|
|
|
1417 |
eval |
|
|
1418 |
to |
|
|
1419 |
|
|
|
1420 |
|
|
|
1421 |
eval |
|
|
1422 |
causes the following result: |
|
|
1423 |
|
|
|
1424 |
|
|
|
1425 |
# perl4 prints: Nope |
|
|
1426 |
# perl5 prints: Yup |
|
|
1427 |
or, changing to |
|
|
1428 |
|
|
|
1429 |
|
|
|
1430 |
eval |
|
|
1431 |
causes the following result: |
|
|
1432 |
|
|
|
1433 |
|
|
|
1434 |
# perl4 prints: Yup |
|
|
1435 |
# perl5 prints: Yup |
|
|
1436 |
# and is compatible for both versions |
|
|
1437 |
|
|
|
1438 |
|
|
|
1439 |
Interpolation |
|
|
1440 |
|
|
|
1441 |
|
|
|
1442 |
perl4 programs which unconsciously rely on the bugs in |
|
|
1443 |
earlier perl versions. |
|
|
1444 |
|
|
|
1445 |
|
|
|
1446 |
perl -e '$bar=q/not/; print |
|
|
1447 |
# perl4 prints: This is not perl5 |
|
|
1448 |
# perl5 prints: This is perl5 |
|
|
1449 |
|
|
|
1450 |
|
|
|
1451 |
Interpolation |
|
|
1452 |
|
|
|
1453 |
|
|
|
1454 |
You also have to be careful about array |
|
|
1455 |
references. |
|
|
1456 |
|
|
|
1457 |
|
|
|
1458 |
print |
|
|
1459 |
perl 4 prints: { |
|
|
1460 |
perl 5 prints: syntax error |
|
|
1461 |
|
|
|
1462 |
|
|
|
1463 |
Interpolation |
|
|
1464 |
|
|
|
1465 |
|
|
|
1466 |
Similarly, watch out for: |
|
|
1467 |
|
|
|
1468 |
|
|
|
1469 |
$foo = |
|
|
1470 |
# perl4 prints: $baz{bar} |
|
|
1471 |
# perl5 prints: $ |
|
|
1472 |
Perl 5 is looking for $foo{bar} which doesn't exist, but perl 4 is happy just to expand $foo to ``baz'' by itself. Watch out for this especially in eval's. |
|
|
1473 |
|
|
|
1474 |
|
|
|
1475 |
Interpolation |
|
|
1476 |
|
|
|
1477 |
|
|
|
1478 |
qq() string passed to eval |
|
|
1479 |
|
|
|
1480 |
|
|
|
1481 |
eval qq( |
|
|
1482 |
foreach $y (keys %$x) { |
|
|
1483 |
$count++; |
|
|
1484 |
} |
|
|
1485 |
); |
|
|
1486 |
# perl4 runs this ok |
|
|
1487 |
# perl5 prints: Can't find string terminator |
|
|
1488 |
|
|
|
1489 |
|
|
|
1490 |
__DBM Traps__ |
|
|
1491 |
|
|
|
1492 |
|
|
|
1493 |
General DBM traps. |
|
|
1494 |
|
|
|
1495 |
|
|
|
1496 |
DBM |
|
|
1497 |
|
|
|
1498 |
|
|
|
1499 |
Existing dbm databases created under perl4 (or any other |
|
|
1500 |
dbm/ndbm tool) may cause the same script, run under perl5, |
|
|
1501 |
to fail. The build of perl5 must have been linked with the |
|
|
1502 |
same dbm/ndbm as the default for dbmopen() to |
|
|
1503 |
function properly without tie'ing to an extension |
|
|
1504 |
dbm implementation. |
|
|
1505 |
|
|
|
1506 |
|
|
|
1507 |
dbmopen (%dbm, |
|
|
1508 |
# perl4 prints: ok |
|
|
1509 |
# perl5 prints: ok (IFF linked with -ldbm or -lndbm) |
|
|
1510 |
|
|
|
1511 |
|
|
|
1512 |
DBM |
|
|
1513 |
|
|
|
1514 |
|
|
|
1515 |
Existing dbm databases created under perl4 (or any other |
|
|
1516 |
dbm/ndbm tool) may cause the same script, run under perl5, |
|
|
1517 |
to fail. The error generated when exceeding the limit on the |
|
|
1518 |
key/value size will cause perl5 to exit |
|
|
1519 |
immediately. |
|
|
1520 |
|
|
|
1521 |
|
|
|
1522 |
dbmopen(DB, |
|
|
1523 |
# perl4 prints: |
|
|
1524 |
dbm store returned -1, errno 28, key |
|
|
1525 |
# perl5 prints: |
|
|
1526 |
dbm store returned -1, errno 28, key |
|
|
1527 |
|
|
|
1528 |
|
|
|
1529 |
__Unclassified Traps__ |
|
|
1530 |
|
|
|
1531 |
|
|
|
1532 |
Everything else. |
|
|
1533 |
|
|
|
1534 |
|
|
|
1535 |
require/do trap using returned |
|
|
1536 |
value |
|
|
1537 |
|
|
|
1538 |
|
|
|
1539 |
If the file doit.pl has: |
|
|
1540 |
|
|
|
1541 |
|
|
|
1542 |
sub foo { |
|
|
1543 |
$rc = do |
|
|
1544 |
And the do.pl file has the following single line: |
|
|
1545 |
|
|
|
1546 |
|
|
|
1547 |
return 3; |
|
|
1548 |
Running doit.pl gives the following: |
|
|
1549 |
|
|
|
1550 |
|
|
|
1551 |
# perl 4 prints: 3 (aborts the subroutine early) |
|
|
1552 |
# perl 5 prints: 8 |
|
|
1553 |
Same behavior if you replace do with require. |
|
|
1554 |
|
|
|
1555 |
|
|
|
1556 |
split on empty string with LIMIT |
|
|
1557 |
specified |
|
|
1558 |
|
|
|
1559 |
|
|
|
1560 |
$string = ''; |
|
|
1561 |
@list = split(/foo/, $string, 2) |
|
|
1562 |
Perl4 returns a one element list containing the empty string but Perl5 returns an empty list. |
|
|
1563 |
|
|
|
1564 |
|
|
|
1565 |
As always, if any of these are ever officially declared as |
|
|
1566 |
bugs, they'll be fixed and removed. |
|
|
1567 |
---- |