3 perltrap - Perl traps for the unwary
7 The biggest trap of all is forgetting to C<use warnings> or use the B<-w>
8 switch; see L<perllexwarn> and L<perlrun>. The second biggest trap is not
9 making your entire program runnable under C<use strict>. The third biggest
10 trap is not reading the list of changes in this version of Perl; see
15 Accustomed B<awk> users should take special note of the following:
21 A Perl program executes only once, not once for each input line. You can
22 do an implicit loop with C<-n> or C<-p>.
26 The English module, loaded via
30 allows you to refer to special variables (like C<$/>) with names (like
31 $RS), as though they were in B<awk>; see L<perlvar> for details.
35 Semicolons are required after all simple statements in Perl (except
36 at the end of a block). Newline is not a statement delimiter.
40 Curly brackets are required on C<if>s and C<while>s.
44 Variables begin with "$", "@" or "%" in Perl.
48 Arrays index from 0. Likewise string positions in substr() and
53 You have to decide whether your array has numeric or string indices.
57 Hash values do not spring into existence upon mere reference.
61 You have to decide whether you want to use string or numeric
66 Reading an input line does not split it for you. You get to split it
67 to an array yourself. And the split() operator has different
68 arguments than B<awk>'s.
72 The current input line is normally in $_, not $0. It generally does
73 not have the newline stripped. ($0 is the name of the program
74 executed.) See L<perlvar>.
78 $<I<digit>> does not refer to fields--it refers to substrings matched
79 by the last match pattern.
83 The print() statement does not add field and record separators unless
84 you set C<$,> and C<$\>. You can set $OFS and $ORS if you're using
89 You must open your files before you print to them.
93 The range operator is "..", not comma. The comma operator works as in
98 The match operator is "=~", not "~". ("~" is the one's complement
103 The exponentiation operator is "**", not "^". "^" is the XOR
104 operator, as in C. (You know, one could get the feeling that B<awk> is
105 basically incompatible with C.)
109 The concatenation operator is ".", not the null string. (Using the
110 null string would render C</pat/ /pat/> unparsable, because the third slash
111 would be interpreted as a division operator--the tokenizer is in fact
112 slightly context sensitive for operators like "/", "?", and ">".
113 And in fact, "." itself can be the beginning of a number.)
117 The C<next>, C<exit>, and C<continue> keywords work differently.
122 The following variables work differently:
125 ARGC scalar @ARGV (compare with $#ARGV)
129 FS (whatever you like)
130 NF $#Fld, or some such
142 You cannot set $RS to a pattern, only a string.
146 When in doubt, run the B<awk> construct through B<a2p> and see what it
153 Cerebral C and C++ programmers should take note of the following:
159 Curly brackets are required on C<if>'s and C<while>'s.
163 You must use C<elsif> rather than C<else if>.
167 The C<break> and C<continue> keywords from C become in Perl C<last>
168 and C<next>, respectively. Unlike in C, these do I<not> work within a
169 C<do { } while> construct. See L<perlsyn/"Loop Control">.
173 There's no switch statement. (But it's easy to build one on the fly,
174 see L<perlsyn/"Basic BLOCKs and Switch Statements">)
178 Variables begin with "$", "@" or "%" in Perl.
182 Comments begin with "#", not "/*" or "//". Perl may interpret C/C++
183 comments as division operators, unterminated regular expressions or
184 the defined-or operator.
188 You can't take the address of anything, although a similar operator
189 in Perl is the backslash, which creates a reference.
193 C<ARGV> must be capitalized. C<$ARGV[0]> is C's C<argv[1]>, and C<argv[0]>
198 System calls such as link(), unlink(), rename(), etc. return nonzero for
199 success, not 0. (system(), however, returns zero for success.)
203 Signal handlers deal with signal names, not numbers. Use C<kill -l>
204 to find their names on your system.
210 Seasoned B<sed> programmers should take note of the following:
216 A Perl program executes only once, not once for each input line. You can
217 do an implicit loop with C<-n> or C<-p>.
221 Backreferences in substitutions use "$" rather than "\".
225 The pattern matching metacharacters "(", ")", and "|" do not have backslashes
230 The range operator is C<...>, rather than comma.
236 Sharp shell programmers should take note of the following:
242 The backtick operator does variable interpolation without regard to
243 the presence of single quotes in the command.
247 The backtick operator does no translation of the return value, unlike B<csh>.
251 Shells (especially B<csh>) do several levels of substitution on each
252 command line. Perl does substitution in only certain constructs
253 such as double quotes, backticks, angle brackets, and search patterns.
257 Shells interpret scripts a little bit at a time. Perl compiles the
258 entire program before executing it (except for C<BEGIN> blocks, which
259 execute at compile time).
263 The arguments are available via @ARGV, not $1, $2, etc.
267 The environment is not automatically made available as separate scalar
272 The shell's C<test> uses "=", "!=", "<" etc for string comparisons and "-eq",
273 "-ne", "-lt" etc for numeric comparisons. This is the reverse of Perl, which
274 uses C<eq>, C<ne>, C<lt> for string comparisons, and C<==>, C<!=> C<< < >> etc
275 for numeric comparisons.
281 Practicing Perl Programmers should take note of the following:
287 Remember that many operations behave differently in a list
288 context than they do in a scalar one. See L<perldata> for details.
292 Avoid barewords if you can, especially all lowercase ones.
293 You can't tell by just looking at it whether a bareword is
294 a function or a string. By using quotes on strings and
295 parentheses on function calls, you won't ever get them confused.
299 You cannot discern from mere inspection which builtins
300 are unary operators (like chop() and chdir())
301 and which are list operators (like print() and unlink()).
302 (Unless prototyped, user-defined subroutines can B<only> be list
303 operators, never unary ones.) See L<perlop> and L<perlsub>.
307 People have a hard time remembering that some functions
308 default to $_, or @ARGV, or whatever, but that others which
309 you might expect to do not.
313 The <FH> construct is not the name of the filehandle, it is a readline
314 operation on that handle. The data read is assigned to $_ only if the
315 file read is the sole condition in a while loop:
318 while (defined($_ = <FH>)) { }..
319 <FH>; # data discarded!
323 Remember not to use C<=> when you need C<=~>;
324 these two constructs are quite different:
331 The C<do {}> construct isn't a real loop that you can use
336 Use C<my()> for local variables whenever you can get away with
337 it (but see L<perlform> for where you can't).
338 Using C<local()> actually gives a local value to a global
339 variable, which leaves you open to unforeseen side-effects
344 If you localize an exported variable in a module, its exported value will
345 not change. The local name becomes an alias to a new value but the
346 external name is still an alias for the original.
350 =head2 Perl4 to Perl5 Traps
352 Practicing Perl4 Programmers should take note of the following
353 Perl4-to-Perl5 specific traps.
355 They're crudely ordered according to the following list:
359 =item Discontinuance, Deprecation, and BugFix traps
361 Anything that's been fixed as a perl4 bug, removed as a perl4 feature
362 or deprecated as a perl4 feature with the intent to encourage usage of
363 some other perl5 feature.
367 Traps that appear to stem from the new parser.
369 =item Numerical Traps
371 Traps having to do with numerical or mathematical operators.
373 =item General data type traps
375 Traps involving perl standard data types.
377 =item Context Traps - scalar, list contexts
379 Traps related to context within lists, scalar statements/declarations.
381 =item Precedence Traps
383 Traps related to the precedence of parsing, evaluation, and execution of
386 =item General Regular Expression Traps using s///, etc.
388 Traps related to the use of pattern matching.
390 =item Subroutine, Signal, Sorting Traps
392 Traps related to the use of signals and signal handlers, general subroutines,
393 and sorting, along with sorting subroutines.
401 Traps specific to the use of C<dbmopen()>, and specific dbm implementations.
403 =item Unclassified Traps
409 If you find an example of a conversion trap that is not listed here,
410 please submit it to <F<perlbug@perl.org>> for inclusion.
411 Also note that at least some of these can be caught with the
412 C<use warnings> pragma or the B<-w> switch.
414 =head2 Discontinuance, Deprecation, and BugFix traps
416 Anything that has been discontinued, deprecated, or fixed as
421 =item * Symbols starting with "_" no longer forced into main
423 Symbols starting with "_" are no longer forced into package main, except
424 for C<$_> itself (and C<@_>, etc.).
430 print "\$_legacy is ",$_legacy,"\n";
432 # perl4 prints: $_legacy is 1
433 # perl5 prints: $_legacy is
435 =item * Double-colon valid package separator in variable name
437 Double-colon is now a valid package separator in a variable name. Thus these
438 behave differently in perl4 vs. perl5, because the packages don't exist.
440 $a=1;$b=2;$c=3;$var=4;
442 print "$var::abc::xyz\n";
444 # perl4 prints: 1::2::3 4::abc::xyz
447 Given that C<::> is now the preferred package delimiter, it is debatable
448 whether this should be classed as a bug or not.
449 (The older package delimiter, ' ,is used here)
455 # perl5 prints: Can't find string terminator "'" anywhere before EOF
457 You can avoid this problem, and remain compatible with perl4, if you
458 always explicitly include the package name:
461 print "x=${main'x}\n" ;
463 Also see precedence traps, for parsing C<$:>.
465 =item * 2nd and 3rd args to C<splice()> are now in scalar context
467 The second and third arguments of C<splice()> are now evaluated in scalar
468 context (as the Camel says) rather than list context.
470 sub sub1{return(0,2) } # return a 2-element list
471 sub sub2{ return(1,2,3)} # return a 3-element list
472 @a1 = ("a","b","c","d","e");
473 @a2 = splice(@a1,&sub1,&sub2);
474 print join(' ',@a2),"\n";
477 # perl5 prints: c d e
479 =item * Can't do C<goto> into a block that is optimized away
481 You can't do a C<goto> into a block that is optimized away. Darn.
487 print "Here I is!\n";
490 # perl4 prints: Here I is!
491 # perl5 errors: Can't "goto" into the middle of a foreach loop
493 =item * Can't use whitespace as variable name or quote delimiter
495 It is no longer syntactically legal to use whitespace as the name
496 of a variable, or as a delimiter for any kind of quote construct.
501 print "a is $a, b is $b\n";
503 # perl4 prints: a is foo bar, b is baz
504 # perl5 errors: Bareword found where operator expected
506 =item * C<while/if BLOCK BLOCK> gone
508 The archaic while/if BLOCK BLOCK syntax is no longer supported.
517 # perl4 prints: True!
518 # perl5 errors: syntax error at test.pl line 1, near "if {"
520 =item * C<**> binds tighter than unary minus
522 The C<**> operator now binds more tightly than unary minus.
523 It was documented to work this way before, but didn't.
530 =item * C<foreach> changed when iterating over a list
532 The meaning of C<foreach{}> has changed slightly when it is iterating over a
533 list which is not an array. This used to assign the list to a
534 temporary array, but no longer does so (for efficiency). This means
535 that you'll now be iterating over the actual values, not over copies of
536 the values. Modifications to the loop variable can change the original
539 @list = ('ab','abc','bcd','def');
540 foreach $var (grep(/ab/,@list)){
543 print (join(':',@list));
545 # perl4 prints: ab:abc:bcd:def
546 # perl5 prints: 1:1:bcd:def
548 To retain Perl4 semantics you need to assign your list
549 explicitly to a temporary array and then iterate over that. For
550 example, you might need to change
552 foreach $var (grep(/ab/,@list)){
556 foreach $var (@tmp = grep(/ab/,@list)){
558 Otherwise changing $var will clobber the values of @list. (This most often
559 happens when you use C<$_> for the loop variable, and call subroutines in
560 the loop that don't properly localize C<$_>.)
562 =item * C<split> with no args behavior changed
564 C<split> with no arguments now behaves like C<split ' '> (which doesn't
565 return an initial null field if $_ starts with whitespace), it used to
566 behave like C<split /\s+/> (which does).
569 print join(':', split);
571 # perl4 prints: :hi:mom
572 # perl5 prints: hi:mom
574 =item * B<-e> behavior fixed
576 Perl 4 would ignore any text which was attached to an B<-e> switch,
577 always taking the code snippet from the following arg. Additionally, it
578 would silently accept an B<-e> switch without a following arg. Both of
579 these behaviors have been fixed.
581 perl -e'print "attached to -e"' 'print "separate arg"'
583 # perl4 prints: separate arg
584 # perl5 prints: attached to -e
589 # perl5 dies: No code specified for -e.
591 =item * C<push> returns number of elements in resulting list
593 In Perl 4 the return value of C<push> was undocumented, but it was
594 actually the last value being pushed onto the target list. In Perl 5
595 the return value of C<push> is documented, but has changed, it is the
596 number of elements in the resulting list.
599 print push(@x, 'first new', 'second new');
601 # perl4 prints: second new
604 =item * Some error messages differ
606 Some error messages will be different.
608 =item * C<split()> honors subroutine args
610 In Perl 4, if in list context the delimiters to the first argument of
611 C<split()> were C<??>, the result would be placed in C<@_> as well as
612 being returned. Perl 5 has more respect for your subroutine arguments.
616 Some bugs may have been inadvertently removed. :-)
622 Perl4-to-Perl5 traps from having to do with parsing.
626 =item * Space between . and = triggers syntax error
628 Note the space between . and =
630 $string . = "more string";
633 # perl4 prints: more string
634 # perl5 prints: syntax error at - line 1, near ". ="
636 =item * Better parsing in perl 5
638 Better parsing in perl 5
642 print("hello, world\n");
644 # perl4 prints: hello, world
645 # perl5 prints: syntax error
647 =item * Function parsing
649 "if it looks like a function, it is a function" rule.
652 ($foo == 1) ? "is one\n" : "is zero\n";
654 # perl4 prints: is zero
655 # perl5 warns: "Useless use of a constant in void context" if using -w
657 =item * String interpolation of C<$#array> differs
659 String interpolation of the C<$#array> construct differs when braces
660 are to used around the name.
666 # perl5 fails with syntax error
674 =item * Perl guesses on C<map>, C<grep> followed by C<{> whether C<{> starts BLOCK or hash ref
676 When perl sees C<map {> (or C<grep {>), it has to guess whether the C<{>
677 starts a BLOCK or a hash reference. If it guesses wrong, it will report
678 a syntax error near the C<}> and the missing (or unexpected) comma.
680 Use unary C<+> before C<{> on a hash reference, and unary C<+> applied
681 to the first thing in a BLOCK (after C<{>), for perl to guess right all
682 the time. (See L<perlfunc/map>.)
686 =head2 Numerical Traps
688 Perl4-to-Perl5 traps having to do with numerical operators,
689 operands, or output from same.
693 =item * Formatted output and significant digits
695 Formatted output and significant digits. In general, Perl 5
696 tries to be more precise. For example, on a Solaris Sparc:
698 print 7.373504 - 0, "\n";
699 printf "%20.18f\n", 7.373504 - 0;
702 7.3750399999999996141
709 Notice how the first result looks better in Perl 5.
711 Your results may vary, since your floating point formatting routines
712 and even floating point format may be slightly different.
714 =item * Auto-increment operator over signed int limit deleted
716 This specific item has been deleted. It demonstrated how the auto-increment
717 operator would not catch when a number went over the signed int limit. Fixed
718 in version 5.003_04. But always be wary when using large integers.
723 =item * Assignment of return values from numeric equality tests doesn't works
725 Assignment of return values from numeric equality tests
726 does not work in perl5 when the test evaluates to false (0).
727 Logical tests now return a null, instead of 0
735 Also see L<"General Regular Expression Traps using s///, etc.">
736 for another example of this new feature...
738 =item * Bitwise string ops
740 When bitwise operators which can operate upon either numbers or
741 strings (C<& | ^ ~>) are given only strings as arguments, perl4 would
742 treat the operands as bitstrings so long as the program contained a call
743 to the C<vec()> function. perl5 treats the string operands as bitstrings.
744 (See L<perlop/Bitwise String Operators> for more details.)
748 $betty = $fred & $barney;
750 # Uncomment the next line to change perl4's behavior
751 # ($dummy) = vec("dummy", 0, 0);
759 # If vec() is used anywhere in the program, both print:
764 =head2 General data type traps
766 Perl4-to-Perl5 traps involving most data-types, and their usage
767 within certain expressions and/or context.
771 =item * Negative array subscripts now count from the end of array
773 Negative array subscripts now count from the end of the array.
775 @a = (1, 2, 3, 4, 5);
776 print "The third element of the array is $a[3] also expressed as $a[-2] \n";
778 # perl4 prints: The third element of the array is 4 also expressed as
779 # perl5 prints: The third element of the array is 4 also expressed as 4
781 =item * Setting C<$#array> lower now discards array elements
783 Setting C<$#array> lower now discards array elements, and makes them
784 impossible to recover.
787 print "Before: ",join('',@a);
789 print ", After: ",join('',@a);
791 print ", Recovered: ",join('',@a),"\n";
793 # perl4 prints: Before: abcde, After: ab, Recovered: abcd
794 # perl5 prints: Before: abcde, After: ab, Recovered: ab
796 =item * Hashes get defined before use
798 Hashes get defined before use
801 die "scalar \$s defined" if defined($s);
802 die "array \@a defined" if defined(@a);
803 die "hash \%h defined" if defined(%h);
806 # perl5 dies: hash %h defined
808 Perl will now generate a warning when it sees defined(@a) and
811 =item * Glob assignment from localized variable to variable
813 glob assignment from variable to variable will fail if the assigned
814 variable is localized subsequent to the assignment
816 @a = ("This is Perl 4");
821 # perl4 prints: This is Perl 4
824 =item * Assigning C<undef> to glob
826 Assigning C<undef> to a glob has no effect in Perl 5. In Perl 4
827 it undefines the associated scalar (but may have other side effects
828 including SEGVs). Perl 5 will also warn if C<undef> is assigned to a
829 typeglob. (Note that assigning C<undef> to a typeglob is different
830 than calling the C<undef> function on a typeglob (C<undef *foo>), which
831 has quite a few effects.
838 # perl4 warns: "Use of uninitialized variable" if using -w
840 # perl5 warns: "Undefined value assigned to typeglob" if using -w
842 =item * Changes in unary negation (of strings)
844 Changes in unary negation (of strings)
845 This change effects both the return value and what it
846 does to auto(magic)increment.
853 # perl4 prints: aab : -0 : 1
854 # perl5 prints: aab : -aab : aac
856 =item * Modifying of constants prohibited
858 perl 4 lets you modify constants:
862 for ($x = 0; $x < 3; $x++) {
866 print "before: $_[0]";
868 print " after: $_[0]\n";
879 # Modification of a read-only value attempted at foo.pl line 12.
882 =item * C<defined $var> behavior changed
884 The behavior is slightly different for:
886 print "$x", defined $x
889 # perl 5: <no output, $x is not called into existence>
891 =item * Variable Suicide
893 Variable suicide behavior is more consistent under Perl 5.
894 Perl5 exhibits the same behavior for hashes and scalars,
895 that perl4 exhibits for only scalars.
897 $aGlobal{ "aKey" } = "global value";
898 print "MAIN:", $aGlobal{"aKey"}, "\n";
903 local( *theArgument ) = @_;
904 local( %aNewLocal ); # perl 4 != 5.001l,m
905 $aNewLocal{"aKey"} = "this should never appear";
906 print "SUB: ", $theArgument{"aKey"}, "\n";
907 $aNewLocal{"aKey"} = "level $GlobalLevel"; # what should print
909 if( $GlobalLevel<4 ) {
924 # SUB: this should never appear
925 # SUB: this should never appear
926 # SUB: this should never appear
930 =head2 Context Traps - scalar, list contexts
934 =item * Elements of argument lists for formats evaluated in list context
936 The elements of argument lists for formats are now evaluated in list
937 context. This means you can interpolate list values now.
939 @fmt = ("foo","bar","baz");
946 # perl4 errors: Please use commas to separate fields in file
947 # perl5 prints: foo bar baz
949 =item * C<caller()> returns false value in scalar context if no caller present
951 The C<caller()> function now returns a false value in a scalar context
952 if there is no caller. This lets library files determine if they're
955 caller() ? (print "You rang?\n") : (print "Got a 0\n");
957 # perl4 errors: There is no caller
958 # perl5 prints: Got a 0
960 =item * Comma operator in scalar context gives scalar context to args
962 The comma operator in a scalar context is now guaranteed to give a
963 scalar context to its arguments.
969 # Perl4 prints: x = c # Thinks list context interpolates list
970 # Perl5 prints: x = 3 # Knows scalar uses length of list
972 =item * C<sprintf()> prototyped as C<($;@)>
974 C<sprintf()> is prototyped as ($;@), so its first argument is given scalar
975 context. Thus, if passed an array, it will probably not do what you want,
978 @z = ('%s%s', 'foo', 'bar');
982 # perl4 prints: foobar
985 C<printf()> works the same as it did in Perl 4, though:
987 @z = ('%s%s', 'foo', 'bar');
990 # perl4 prints: foobar
991 # perl5 prints: foobar
995 =head2 Precedence Traps
997 Perl4-to-Perl5 traps involving precedence order.
999 Perl 4 has almost the same precedence rules as Perl 5 for the operators
1000 that they both have. Perl 4 however, seems to have had some
1001 inconsistencies that made the behavior differ from what was documented.
1005 =item * LHS vs. RHS of any assignment operator
1007 LHS vs. RHS of any assignment operator. LHS is evaluated first
1008 in perl4, second in perl5; this can affect the relationship
1009 between side-effects in sub-expressions.
1011 @arr = ( 'left', 'right' );
1012 $a{shift @arr} = shift @arr;
1013 print join( ' ', keys %a );
1015 # perl4 prints: left
1016 # perl5 prints: right
1018 =item * Semantic errors introduced due to precedence
1020 These are now semantic errors because of precedence:
1022 @list = (1,2,3,4,5);
1023 %map = ("a",1,"b",2,"c",3,"d",4);
1024 $n = shift @list + 2; # first item in list plus 2
1026 $m = keys %map + 2; # number of items in hash plus 2
1029 # perl4 prints: n is 3, m is 6
1030 # perl5 errors and fails to compile
1032 =item * Precedence of assignment operators same as the precedence of assignment
1034 The precedence of assignment operators is now the same as the precedence
1035 of assignment. Perl 4 mistakenly gave them the precedence of the associated
1036 operator. So you now must parenthesize them in expressions like
1038 /foo/ ? ($a += 2) : ($a -= 2);
1042 /foo/ ? $a += 2 : $a -= 2
1044 would be erroneously parsed as
1046 (/foo/ ? $a += 2 : $a) -= 2;
1050 $a += /foo/ ? 1 : 2;
1052 now works as a C programmer would expect.
1054 =item * C<open> requires parentheses around filehandle
1058 is now incorrect. You need parentheses around the filehandle.
1059 Otherwise, perl5 leaves the statement as its default precedence:
1063 # perl4 opens or dies
1064 # perl5 opens FOO, dying only if 'FOO' is false, i.e. never
1066 =item * C<$:> precedence over C<$::> gone
1068 perl4 gives the special variable, C<$:> precedence, where perl5
1069 treats C<$::> as main C<package>
1071 $a = "x"; print "$::a";
1073 # perl 4 prints: -:a
1076 =item * Precedence of file test operators documented
1078 perl4 had buggy precedence for the file test operators vis-a-vis
1079 the assignment operators. Thus, although the precedence table
1080 for perl4 leads one to believe C<-e $foo .= "q"> should parse as
1081 C<((-e $foo) .= "q")>, it actually parses as C<(-e ($foo .= "q"))>.
1082 In perl5, the precedence is as documented.
1086 # perl4 prints: no output
1087 # perl5 prints: Can't modify -e in concatenation
1089 =item * C<keys>, C<each>, C<values> are regular named unary operators
1091 In perl4, keys(), each() and values() were special high-precedence operators
1092 that operated on a single hash, but in perl5, they are regular named unary
1093 operators. As documented, named unary operators have lower precedence
1094 than the arithmetic and concatenation operators C<+ - .>, but the perl4
1095 variants of these operators actually bind tighter than C<+ - .>.
1102 # perl5 prints: Type of arg 1 to keys must be hash (not subtraction)
1104 The perl4 behavior was probably more useful, if less consistent.
1108 =head2 General Regular Expression Traps using s///, etc.
1110 All types of RE traps.
1114 =item * C<s'$lhs'$rhs'> interpolates on either side
1116 C<s'$lhs'$rhs'> now does no interpolation on either side. It used to
1117 interpolate $lhs but not $rhs. (And still does not match a literal
1121 $string = '1 2 $a $b';
1122 $string =~ s'$a'$b';
1125 # perl4 prints: $b 2 $a $b
1126 # perl5 prints: 1 2 $a $b
1128 =item * C<m//g> attaches its state to the searched string
1130 C<m//g> now attaches its state to the searched string rather than the
1131 regular expression. (Once the scope of a block is left for the sub, the
1132 state of the searched string is lost)
1138 sub doit{local($_) = shift; print "Got $_ "}
1140 # perl4 prints: Got blah Got blah Got blah Got blah
1141 # perl5 prints: infinite loop blah...
1143 =item * C<m//o> used within an anonymous sub
1145 Currently, if you use the C<m//o> qualifier on a regular expression
1146 within an anonymous sub, I<all> closures generated from that anonymous
1147 sub will use the regular expression as it was compiled when it was used
1148 the very first time in any such closure. For instance, if you say
1151 my($left,$right) = @_;
1152 return sub { $_[0] =~ /$left stuff $right/o; };
1154 $good = build_match('foo','bar');
1155 $bad = build_match('baz','blarch');
1156 print $good->('foo stuff bar') ? "ok\n" : "not ok\n";
1157 print $bad->('baz stuff blarch') ? "ok\n" : "not ok\n";
1158 print $bad->('foo stuff bar') ? "not ok\n" : "ok\n";
1160 For most builds of Perl5, this will print:
1165 build_match() will always return a sub which matches the contents of
1166 $left and $right as they were the I<first> time that build_match()
1167 was called, not as they are in the current call.
1169 =item * C<$+> isn't set to whole match
1171 If no parentheses are used in a match, Perl4 sets C<$+> to
1172 the whole match, just like C<$&>. Perl5 does not.
1177 # perl4 prints: bcde
1180 =item * Substitution now returns null string if it fails
1182 substitution now returns the null string if it fails
1185 $value = ($string =~ s/foo//);
1191 Also see L<Numerical Traps> for another example of this new feature.
1193 =item * C<s`lhs`rhs`> is now a normal substitution
1195 C<s`lhs`rhs`> (using backticks) is now a normal substitution, with no
1199 $string =~ s`^`hostname`;
1200 print $string, "\n";
1202 # perl4 prints: <the local hostname>
1203 # perl5 prints: hostname
1205 =item * Stricter parsing of variables in regular expressions
1207 Stricter parsing of variables used in regular expressions
1209 s/^([^$grpc]*$grpc[$opt$plus$rep]?)//o;
1211 # perl4: compiles w/o error
1212 # perl5: with Scalar found where operator expected ..., near "$opt$plus"
1214 an added component of this example, apparently from the same script, is
1215 the actual value of the s'd string after the substitution.
1216 C<[$opt]> is a character class in perl4 and an array subscript in perl5
1221 s/^([^$grpc]*$grpc[$opt]?)/foo/;
1225 # perl5 prints: foobar
1227 =item * C<m?x?> matches only once
1229 Under perl5, C<m?x?> matches only once, like C<?x?>. Under perl4, it matched
1230 repeatedly, like C</x/> or C<m!x!>.
1233 sub match { $test =~ m?once?; }
1236 # m?x? matches more then once
1239 # m?x? matches only once
1243 # perl4 prints: perl4
1244 # perl5 prints: perl5
1246 =item * Failed matches don't reset the match variables
1248 Unlike in Ruby, failed matches in Perl do not reset the match variables
1249 ($1, $2, ..., C<$`>, ...).
1253 =head2 Subroutine, Signal, Sorting Traps
1255 The general group of Perl4-to-Perl5 traps having to do with
1256 Signals, Sorting, and their related subroutines, as well as
1257 general subroutine traps. Includes some OS-Specific traps.
1261 =item * Barewords that used to look like strings look like subroutine calls if a subroutine by that name is defined
1263 Barewords that used to look like strings to Perl will now look like subroutine
1264 calls if a subroutine by that name is defined before the compiler sees them.
1266 sub SeeYa { warn"Hasta la vista, baby!" }
1267 $SIG{'TERM'} = SeeYa;
1268 print "SIGTERM is now $SIG{'TERM'}\n";
1270 # perl4 prints: SIGTERM is now main'SeeYa
1271 # perl5 prints: SIGTERM is now main::1 (and warns "Hasta la vista, baby!")
1273 Use B<-w> to catch this one
1275 =item * Reverse is no longer allowed as the name of a sort subroutine
1277 reverse is no longer allowed as the name of a sort subroutine.
1279 sub reverse{ print "yup "; $a <=> $b }
1280 print sort reverse (2,1,3);
1282 # perl4 prints: yup yup 123
1284 # perl5 warns (if using -w): Ambiguous call resolved as CORE::reverse()
1286 =item * C<warn()> won't let you specify a filehandle.
1288 Although it _always_ printed to STDERR, warn() would let you specify a
1289 filehandle in perl4. With perl5 it does not.
1293 # perl4 prints: Foo!
1294 # perl5 prints: String found where operator expected
1302 =item * SysV resets signal handler correctly
1304 Under HPUX, and some other SysV OSes, one had to reset any signal handler,
1305 within the signal handler function, each time a signal was handled with
1306 perl4. With perl5, the reset is now done correctly. Any code relying
1307 on the handler _not_ being reset will have to be reworked.
1309 Since version 5.002, Perl uses sigaction() under SysV.
1314 $SIG{'INT'} = 'gotit';
1323 while (1) {sleep(10);}
1326 # perl4 (HPUX) prints: Got INT...
1327 # perl5 (HPUX) prints: Got INT... Got INT...
1329 =item * SysV C<seek()> appends correctly
1331 Under SysV OSes, C<seek()> on a file opened to append C<<< >> >>> now does
1332 the right thing w.r.t. the fopen() manpage. e.g., - When a file is opened
1333 for append, it is impossible to overwrite information already in
1336 open(TEST,">>seek.test");
1337 $start = tell TEST ;
1342 seek(TEST,$start,0);
1343 print TEST "18 characters here";
1345 # perl4 (solaris) seek.test has: 18 characters here
1346 # perl5 (solaris) seek.test has: 1 2 3 4 5 6 7 8 9 18 characters here
1352 =head2 Interpolation Traps
1354 Perl4-to-Perl5 traps having to do with how things get interpolated
1355 within certain expressions, statements, contexts, or whatever.
1359 =item * C<@> always interpolates an array in double-quotish strings
1361 @ now always interpolates an array in double-quotish strings.
1363 print "To: someone@somewhere.com\n";
1365 # perl4 prints: To:someone@somewhere.com
1366 # perl < 5.6.1, error : In string, @somewhere now must be written as \@somewhere
1367 # perl >= 5.6.1, warning : Possible unintended interpolation of @somewhere in string
1369 =item * Double-quoted strings may no longer end with an unescaped $
1371 Double-quoted strings may no longer end with an unescaped $.
1374 print "foo is $foo\n";
1376 # perl4 prints: foo is foo$
1377 # perl5 errors: Final $ should be \$ or $name
1379 Note: perl5 DOES NOT error on the terminating @ in $bar
1381 =item * Arbitrary expressions are evaluated inside braces that occur within double quotes
1383 Perl now sometimes evaluates arbitrary expressions inside braces that occur
1384 within double quotes (usually when the opening brace is preceded by C<$>
1390 sub foo { return "bar" };
1391 print "|@{w.w.w}|${main'foo}|";
1393 # perl4 prints: |@{w.w.w}|foo|
1394 # perl5 prints: |buz|bar|
1396 Note that you can C<use strict;> to ward off such trappiness under perl5.
1398 =item * C<$$x> now tries to dereference $x
1400 The construct "this is $$x" used to interpolate the pid at that point, but
1401 now tries to dereference $x. C<$$> by itself still works fine, however.
1405 print "this is $$x\n";
1407 # perl4 prints: this is XXXx (XXX is the current pid)
1408 # perl5 prints: this is a reference
1410 =item * Creation of hashes on the fly with C<eval "EXPR"> requires protection of C<$>'s or both curlies
1412 Creation of hashes on the fly with C<eval "EXPR"> now requires either both
1413 C<$>'s to be protected in the specification of the hash name, or both curlies
1414 to be protected. If both curlies are protected, the result will be compatible
1415 with perl4 and perl5. This is a very common practice, and should be changed
1416 to use the block form of C<eval{}> if possible.
1418 $hashname = "foobar";
1421 eval "\$$hashname{'$key'} = q|$value|";
1422 (defined($foobar{'baz'})) ? (print "Yup") : (print "Nope");
1425 # perl5 prints: Nope
1429 eval "\$$hashname{'$key'} = q|$value|";
1433 eval "\$\$hashname{'$key'} = q|$value|";
1435 causes the following result:
1437 # perl4 prints: Nope
1442 eval "\$$hashname\{'$key'\} = q|$value|";
1444 causes the following result:
1448 # and is compatible for both versions
1451 =item * Bugs in earlier perl versions
1453 perl4 programs which unconsciously rely on the bugs in earlier perl versions.
1455 perl -e '$bar=q/not/; print "This is $foo{$bar} perl5"'
1457 # perl4 prints: This is not perl5
1458 # perl5 prints: This is perl5
1460 =item * Array and hash brackets during interpolation
1462 You also have to be careful about array and hash brackets during
1468 perl 5 prints: syntax error
1473 perl 5 prints: syntax error
1475 Perl 5 is expecting to find an index or key name following the respective
1476 brackets, as well as an ending bracket of the appropriate type. In order
1477 to mimic the behavior of Perl 4, you must escape the bracket like so.
1482 =item * Interpolation of C<\$$foo{bar}>
1484 Similarly, watch out for: C<\$$foo{bar}>
1487 print "\$$foo{bar}\n";
1489 # perl4 prints: $baz{bar}
1492 Perl 5 is looking for C<$foo{bar}> which doesn't exist, but perl 4 is
1493 happy just to expand $foo to "baz" by itself. Watch out for this
1494 especially in C<eval>'s.
1496 =item * C<qq()> string passed to C<eval> will not find string terminator
1498 C<qq()> string passed to C<eval>
1501 foreach \$y (keys %\$x\) {
1506 # perl4 runs this ok
1507 # perl5 prints: Can't find string terminator ")"
1517 =item * Perl5 must have been linked with the same dbm/ndbm as the default for C<dbmopen()> to function properly without C<tie>'ing
1519 Existing dbm databases created under perl4 (or any other dbm/ndbm tool)
1520 may cause the same script, run under perl5, to fail. The build of perl5
1521 must have been linked with the same dbm/ndbm as the default for C<dbmopen()>
1522 to function properly without C<tie>'ing to an extension dbm implementation.
1524 dbmopen (%dbm, "file", undef);
1528 # perl5 prints: ok (IFF linked with -ldbm or -lndbm)
1531 =item * DBM exceeding the limit on the key/value size will cause perl5 to exit immediately
1533 Existing dbm databases created under perl4 (or any other dbm/ndbm tool)
1534 may cause the same script, run under perl5, to fail. The error generated
1535 when exceeding the limit on the key/value size will cause perl5 to exit
1538 dbmopen(DB, "testdb",0600) || die "couldn't open db! $!";
1539 $DB{'trap'} = "x" x 1024; # value too large for most dbm/ndbm
1543 dbm store returned -1, errno 28, key "trap" at - line 3.
1547 dbm store returned -1, errno 28, key "trap" at - line 3.
1551 =head2 Unclassified Traps
1557 =item * C<require>/C<do> trap using returned value
1559 If the file doit.pl has:
1567 And the do.pl file has the following single line:
1571 Running doit.pl gives the following:
1573 # perl 4 prints: 3 (aborts the subroutine early)
1576 Same behavior if you replace C<do> with C<require>.
1578 =item * C<split> on empty string with LIMIT specified
1581 @list = split(/foo/, $string, 2)
1583 Perl4 returns a one element list containing the empty string but Perl5
1584 returns an empty list.
1588 As always, if any of these are ever officially declared as bugs,
1589 they'll be fixed and removed.