3 perltrap - Perl traps for the unwary
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
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>.
14 Accustomed B<awk> users should take special note of the following:
20 The English module, loaded via
24 allows you to refer to special variables (like C<$/>) with names (like
25 C<$RS>), as though they were in B<awk>; see L<perlvar> for details.
29 Semicolons are required after all simple statements in Perl (except
30 at the end of a block). Newline is not a statement delimiter.
34 Curly brackets are required on C<if>s and C<while>s.
38 Variables begin with "$", "@" or "%" in Perl.
42 Arrays index from 0. Likewise string positions in substr() and
47 You have to decide whether your array has numeric or string indices.
51 Hash values do not spring into existence upon mere reference.
55 You have to decide whether you want to use string or numeric
60 Reading an input line does not split it for you. You get to split it
61 to an array yourself. And the split() operator has different
62 arguments than B<awk>'s.
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>.
72 $E<lt>I<digit>E<gt> does not refer to fields--it refers to substrings matched
73 by the last match pattern.
77 The print() statement does not add field and record separators unless
78 you set C<$,> and C<$\>. You can set $OFS and $ORS if you're using
83 You must open your files before you print to them.
87 The range operator is "..", not comma. The comma operator works as in
92 The match operator is "=~", not "~". ("~" is the one's complement
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.)
103 The concatenation operator is ".", not the null string. (Using the
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
106 slightly context sensitive for operators like "/", "?", and "E<gt>".
107 And in fact, "." itself can be the beginning of a number.)
111 The C<next>, C<exit>, and C<continue> keywords work differently.
116 The following variables work differently:
119 ARGC $#ARGV or scalar @ARGV
123 FS (whatever you like)
124 NF $#Fld, or some such
136 You cannot set $RS to a pattern, only a string.
140 When in doubt, run the B<awk> construct through B<a2p> and see what it
147 Cerebral C programmers should take note of the following:
153 Curly brackets are required on C<if>'s and C<while>'s.
157 You must use C<elsif> rather than C<else if>.
161 The C<break> and C<continue> keywords from C become in
162 Perl C<last> and C<next>, respectively.
163 Unlike in C, these do I<NOT> work within a C<do { } while> construct.
167 There's no switch statement. (But it's easy to build one on the fly.)
171 Variables begin with "$", "@" or "%" in Perl.
175 C<printf()> does not implement the "*" format for interpolating
176 field widths, but it's trivial to use interpolation of double-quoted
177 strings to achieve the same effect.
181 Comments begin with "#", not "/*".
185 You can't take the address of anything, although a similar operator
186 in Perl is the backslash, which creates a reference.
190 C<ARGV> must be capitalized. C<$ARGV[0]> is C's C<argv[1]>, and C<argv[0]>
195 System calls such as link(), unlink(), rename(), etc. return nonzero for
200 Signal handlers deal with signal names, not numbers. Use C<kill -l>
201 to find their names on your system.
207 Seasoned B<sed> programmers should take note of the following:
213 Backreferences in substitutions use "$" rather than "\".
217 The pattern matching metacharacters "(", ")", and "|" do not have backslashes
222 The range operator is C<...>, rather than comma.
228 Sharp shell programmers should take note of the following:
234 The backtick operator does variable interpolation without regard to
235 the presence of single quotes in the command.
239 The backtick operator does no translation of the return value, unlike B<csh>.
243 Shells (especially B<csh>) do several levels of substitution on each
244 command line. Perl does substitution in only certain constructs
245 such as double quotes, backticks, angle brackets, and search patterns.
249 Shells interpret scripts a little bit at a time. Perl compiles the
250 entire program before executing it (except for C<BEGIN> blocks, which
251 execute at compile time).
255 The arguments are available via @ARGV, not $1, $2, etc.
259 The environment is not automatically made available as separate scalar
266 Practicing Perl Programmers should take note of the following:
272 Remember that many operations behave differently in a list
273 context than they do in a scalar one. See L<perldata> for details.
277 Avoid barewords if you can, especially all lowercase ones.
278 You can't tell by just looking at it whether a bareword is
279 a function or a string. By using quotes on strings and
280 parentheses on function calls, you won't ever get them confused.
284 You cannot discern from mere inspection which builtins
285 are unary operators (like chop() and chdir())
286 and which are list operators (like print() and unlink()).
287 (User-defined subroutines can be B<only> list operators, never
288 unary ones.) See L<perlop>.
292 People have a hard time remembering that some functions
293 default to $_, or @ARGV, or whatever, but that others which
294 you might expect to do not.
298 The E<lt>FHE<gt> construct is not the name of the filehandle, it is a readline
299 operation on that handle. The data read is assigned to $_ only if the
300 file read is the sole condition in a while loop:
303 while (defined($_ = <FH>)) { }..
304 <FH>; # data discarded!
308 Remember not to use "C<=>" when you need "C<=~>";
309 these two constructs are quite different:
316 The C<do {}> construct isn't a real loop that you can use
321 Use C<my()> for local variables whenever you can get away with
322 it (but see L<perlform> for where you can't).
323 Using C<local()> actually gives a local value to a global
324 variable, which leaves you open to unforeseen side-effects
329 If you localize an exported variable in a module, its exported value will
330 not change. The local name becomes an alias to a new value but the
331 external name is still an alias for the original.
335 =head2 Perl4 to Perl5 Traps
337 Practicing Perl4 Programmers should take note of the following
338 Perl4-to-Perl5 specific traps.
340 They're crudely ordered according to the following list:
344 =item Discontinuance, Deprecation, and BugFix traps
346 Anything that's been fixed as a perl4 bug, removed as a perl4 feature
347 or deprecated as a perl4 feature with the intent to encourage usage of
348 some other perl5 feature.
352 Traps that appear to stem from the new parser.
354 =item Numerical Traps
356 Traps having to do with numerical or mathematical operators.
358 =item General data type traps
360 Traps involving perl standard data types.
362 =item Context Traps - scalar, list contexts
364 Traps related to context within lists, scalar statements/declarations.
366 =item Precedence Traps
368 Traps related to the precedence of parsing, evaluation, and execution of
371 =item General Regular Expression Traps using s///, etc.
373 Traps related to the use of pattern matching.
375 =item Subroutine, Signal, Sorting Traps
377 Traps related to the use of signals and signal handlers, general subroutines,
378 and sorting, along with sorting subroutines.
386 Traps specific to the use of C<dbmopen()>, and specific dbm implementations.
388 =item Unclassified Traps
394 If you find an example of a conversion trap that is not listed here,
395 please submit it to Bill Middleton <F<wjm@best.com>> for inclusion.
396 Also note that at least some of these can be caught with B<-w>.
398 =head2 Discontinuance, Deprecation, and BugFix traps
400 Anything that has been discontinued, deprecated, or fixed as
405 =item * Discontinuance
407 Symbols starting with "_" are no longer forced into package main, except
408 for C<$_> itself (and C<@_>, etc.).
414 print "\$_legacy is ",$_legacy,"\n";
416 # perl4 prints: $_legacy is 1
417 # perl5 prints: $_legacy is
421 Double-colon is now a valid package separator in a variable name. Thus these
422 behave differently in perl4 vs. perl5, because the packages don't exist.
424 $a=1;$b=2;$c=3;$var=4;
426 print "$var::abc::xyz\n";
428 # perl4 prints: 1::2::3 4::abc::xyz
431 Given that C<::> is now the preferred package delimiter, it is debatable
432 whether this should be classed as a bug or not.
433 (The older package delimiter, ' ,is used here)
439 # perl5 prints: Can't find string terminator "'" anywhere before EOF
441 You can avoid this problem, and remain compatible with perl4, if you
442 always explicitly include the package name:
445 print "x=${main'x}\n" ;
447 Also see precedence traps, for parsing C<$:>.
451 The second and third arguments of C<splice()> are now evaluated in scalar
452 context (as the Camel says) rather than list context.
454 sub sub1{return(0,2) } # return a 2-element list
455 sub sub2{ return(1,2,3)} # return a 3-element list
456 @a1 = ("a","b","c","d","e");
457 @a2 = splice(@a1,&sub1,&sub2);
458 print join(' ',@a2),"\n";
461 # perl5 prints: c d e
463 =item * Discontinuance
465 You can't do a C<goto> into a block that is optimized away. Darn.
471 print "Here I is!\n";
474 # perl4 prints: Here I is!
475 # perl5 dumps core (SEGV)
477 =item * Discontinuance
479 It is no longer syntactically legal to use whitespace as the name
480 of a variable, or as a delimiter for any kind of quote construct.
485 print "a is $a, b is $b\n";
487 # perl4 prints: a is foo bar, b is baz
488 # perl5 errors: Bareword found where operator expected
490 =item * Discontinuance
492 The archaic while/if BLOCK BLOCK syntax is no longer supported.
501 # perl4 prints: True!
502 # perl5 errors: syntax error at test.pl line 1, near "if {"
506 The C<**> operator now binds more tightly than unary minus.
507 It was documented to work this way before, but didn't.
514 =item * Discontinuance
516 The meaning of C<foreach{}> has changed slightly when it is iterating over a
517 list which is not an array. This used to assign the list to a
518 temporary array, but no longer does so (for efficiency). This means
519 that you'll now be iterating over the actual values, not over copies of
520 the values. Modifications to the loop variable can change the original
523 @list = ('ab','abc','bcd','def');
524 foreach $var (grep(/ab/,@list)){
527 print (join(':',@list));
529 # perl4 prints: ab:abc:bcd:def
530 # perl5 prints: 1:1:bcd:def
532 To retain Perl4 semantics you need to assign your list
533 explicitly to a temporary array and then iterate over that. For
534 example, you might need to change
536 foreach $var (grep(/ab/,@list)){
540 foreach $var (@tmp = grep(/ab/,@list)){
542 Otherwise changing $var will clobber the values of @list. (This most often
543 happens when you use C<$_> for the loop variable, and call subroutines in
544 the loop that don't properly localize C<$_>.)
546 =item * Discontinuance
548 C<split> with no arguments now behaves like C<split ' '> (which doesn't
549 return an initial null field if $_ starts with whitespace), it used to
550 behave like C<split /\s+/> (which does).
553 print join(':', split);
555 # perl4 prints: :hi:mom
556 # perl5 prints: hi:mom
560 Perl 4 would ignore any text which was attached to an B<-e> switch,
561 always taking the code snippet from the following arg. Additionally, it
562 would silently accept an B<-e> switch without a following arg. Both of
563 these behaviors have been fixed.
565 perl -e'print "attached to -e"' 'print "separate arg"'
567 # perl4 prints: separate arg
568 # perl5 prints: attached to -e
573 # perl5 dies: No code specified for -e.
575 =item * Discontinuance
577 In Perl 4 the return value of C<push> was undocumented, but it was
578 actually the last value being pushed onto the target list. In Perl 5
579 the return value of C<push> is documented, but has changed, it is the
580 number of elements in the resulting list.
583 print push(@x, 'first new', 'second new');
585 # perl4 prints: second new
588 =item * Discontinuance
590 In Perl 4 (and versions of Perl 5 before 5.004), C<'\r'> characters in
591 Perl code were silently allowed, although they could cause (mysterious!)
592 failures in certain constructs, particularly here documents. Now,
593 C<'\r'> characters cause an immediate fatal error. (Note: In this
594 example, the notation B<\015> represents the incorrect line
595 ending. Depending upon your text viewer, it will look different.)
600 # perl4 prints: foobar
601 # perl5.003 prints: foobar
602 # perl5.004 dies: Illegal character \015 (carriage return)
604 See L<perldiag> for full details.
608 Some error messages will be different.
610 =item * Discontinuance
612 Some bugs may have been inadvertently removed. :-)
618 Perl4-to-Perl5 traps from having to do with parsing.
624 Note the space between . and =
626 $string . = "more string";
629 # perl4 prints: more string
630 # perl5 prints: syntax error at - line 1, near ". ="
634 Better parsing in perl 5
638 print("hello, world\n");
640 # perl4 prints: hello, world
641 # perl5 prints: syntax error
645 "if it looks like a function, it is a function" rule.
648 ($foo == 1) ? "is one\n" : "is zero\n";
650 # perl4 prints: is zero
651 # perl5 warns: "Useless use of a constant in void context" if using -w
655 =head2 Numerical Traps
657 Perl4-to-Perl5 traps having to do with numerical operators,
658 operands, or output from same.
664 Formatted output and significant digits
666 print 7.373504 - 0, "\n";
667 printf "%20.18f\n", 7.373504 - 0;
679 This specific item has been deleted. It demonstrated how the auto-increment
680 operator would not catch when a number went over the signed int limit. Fixed
681 in version 5.003_04. But always be wary when using large integers.
688 Assignment of return values from numeric equality tests
689 does not work in perl5 when the test evaluates to false (0).
690 Logical tests now return an null, instead of 0
698 Also see L<"General Regular Expression Traps using s///, etc.">
699 for another example of this new feature...
703 =head2 General data type traps
705 Perl4-to-Perl5 traps involving most data-types, and their usage
706 within certain expressions and/or context.
712 Negative array subscripts now count from the end of the array.
714 @a = (1, 2, 3, 4, 5);
715 print "The third element of the array is $a[3] also expressed as $a[-2] \n";
717 # perl4 prints: The third element of the array is 4 also expressed as
718 # perl5 prints: The third element of the array is 4 also expressed as 4
722 Setting C<$#array> lower now discards array elements, and makes them
723 impossible to recover.
726 print "Before: ",join('',@a);
728 print ", After: ",join('',@a);
730 print ", Recovered: ",join('',@a),"\n";
732 # perl4 prints: Before: abcde, After: ab, Recovered: abcd
733 # perl5 prints: Before: abcde, After: ab, Recovered: ab
737 Hashes get defined before use
740 die "scalar \$s defined" if defined($s);
741 die "array \@a defined" if defined(@a);
742 die "hash \%h defined" if defined(%h);
745 # perl5 dies: hash %h defined
749 glob assignment from variable to variable will fail if the assigned
750 variable is localized subsequent to the assignment
752 @a = ("This is Perl 4");
757 # perl4 prints: This is Perl 4
762 Assigning C<undef> to a glob has no effect in Perl 5. In Perl 4
763 it undefines the associated scalar (but may have other side effects
766 =item * (Scalar String)
768 Changes in unary negation (of strings)
769 This change effects both the return value and what it
770 does to auto(magic)increment.
777 # perl4 prints: aab : -0 : 1
778 # perl5 prints: aab : -aab : aac
782 perl 4 lets you modify constants:
786 for ($x = 0; $x < 3; $x++) {
790 print "before: $_[0]";
792 print " after: $_[0]\n";
803 # Modification of a read-only value attempted at foo.pl line 12.
808 The behavior is slightly different for:
810 print "$x", defined $x
813 # perl 5: <no output, $x is not called into existence>
815 =item * (Variable Suicide)
817 Variable suicide behavior is more consistent under Perl 5.
818 Perl5 exhibits the same behavior for hashes and scalars,
819 that perl4 exhibits for only scalars.
821 $aGlobal{ "aKey" } = "global value";
822 print "MAIN:", $aGlobal{"aKey"}, "\n";
827 local( *theArgument ) = @_;
828 local( %aNewLocal ); # perl 4 != 5.001l,m
829 $aNewLocal{"aKey"} = "this should never appear";
830 print "SUB: ", $theArgument{"aKey"}, "\n";
831 $aNewLocal{"aKey"} = "level $GlobalLevel"; # what should print
833 if( $GlobalLevel<4 ) {
848 # SUB: this should never appear
849 # SUB: this should never appear
850 # SUB: this should never appear
854 =head2 Context Traps - scalar, list contexts
858 =item * (list context)
860 The elements of argument lists for formats are now evaluated in list
861 context. This means you can interpolate list values now.
863 @fmt = ("foo","bar","baz");
870 # perl4 errors: Please use commas to separate fields in file
871 # perl5 prints: foo bar baz
873 =item * (scalar context)
875 The C<caller()> function now returns a false value in a scalar context
876 if there is no caller. This lets library files determine if they're
879 caller() ? (print "You rang?\n") : (print "Got a 0\n");
881 # perl4 errors: There is no caller
882 # perl5 prints: Got a 0
884 =item * (scalar context)
886 The comma operator in a scalar context is now guaranteed to give a
887 scalar context to its arguments.
893 # Perl4 prints: x = c # Thinks list context interpolates list
894 # Perl5 prints: x = 3 # Knows scalar uses length of list
896 =item * (list, builtin)
898 C<sprintf()> funkiness (array argument converted to scalar array count)
899 This test could be added to t/op/sprintf.t
901 @z = ('%s%s', 'foo', 'bar');
903 if ($x eq 'foobar') {print "ok 2\n";} else {print "not ok 2 '$x'\n";}
906 # perl5 prints: not ok 2
908 C<printf()> works fine, though:
913 # perl4 prints: foobar
914 # perl5 prints: foobar
920 =head2 Precedence Traps
922 Perl4-to-Perl5 traps involving precedence order.
924 Perl 4 has almost the same precedence rules as Perl 5 for the operators
925 that they both have. Perl 4 however, seems to have had some
926 inconsistencies that made the behavior differ from what was documented.
932 LHS vs. RHS of any assignment operator. LHS is evaluated first
933 in perl4, second in perl5; this can affect the relationship
934 between side-effects in sub-expressions.
936 @arr = ( 'left', 'right' );
937 $a{shift @arr} = shift @arr;
938 print join( ' ', keys %a );
941 # perl5 prints: right
945 These are now semantic errors because of precedence:
948 %map = ("a",1,"b",2,"c",3,"d",4);
949 $n = shift @list + 2; # first item in list plus 2
951 $m = keys %map + 2; # number of items in hash plus 2
954 # perl4 prints: n is 3, m is 6
955 # perl5 errors and fails to compile
959 The precedence of assignment operators is now the same as the precedence
960 of assignment. Perl 4 mistakenly gave them the precedence of the associated
961 operator. So you now must parenthesize them in expressions like
963 /foo/ ? ($a += 2) : ($a -= 2);
967 /foo/ ? $a += 2 : $a -= 2
969 would be erroneously parsed as
971 (/foo/ ? $a += 2 : $a) -= 2;
977 now works as a C programmer would expect.
983 is now incorrect. You need parentheses around the filehandle.
984 Otherwise, perl5 leaves the statement as its default precedence:
988 # perl4 opens or dies
989 # perl5 errors: Precedence problem: open FOO should be open(FOO)
993 perl4 gives the special variable, C<$:> precedence, where perl5
994 treats C<$::> as main C<package>
996 $a = "x"; print "$::a";
1003 perl4 had buggy precedence for the file test operators vis-a-vis
1004 the assignment operators. Thus, although the precedence table
1005 for perl4 leads one to believe C<-e $foo .= "q"> should parse as
1006 C<((-e $foo) .= "q")>, it actually parses as C<(-e ($foo .= "q"))>.
1007 In perl5, the precedence is as documented.
1011 # perl4 prints: no output
1012 # perl5 prints: Can't modify -e in concatenation
1016 In perl4, keys(), each() and values() were special high-precedence operators
1017 that operated on a single hash, but in perl5, they are regular named unary
1018 operators. As documented, named unary operators have lower precedence
1019 than the arithmetic and concatenation operators C<+ - .>, but the perl4
1020 variants of these operators actually bind tighter than C<+ - .>.
1027 # perl5 prints: Type of arg 1 to keys must be hash (not subtraction)
1029 The perl4 behavior was probably more useful, if less consistent.
1033 =head2 General Regular Expression Traps using s///, etc.
1035 All types of RE traps.
1039 =item * Regular Expression
1041 C<s'$lhs'$rhs'> now does no interpolation on either side. It used to
1042 interpolate C<$lhs> but not C<$rhs>. (And still does not match a literal
1046 $string = '1 2 $a $b';
1047 $string =~ s'$a'$b';
1050 # perl4 prints: $b 2 $a $b
1051 # perl5 prints: 1 2 $a $b
1053 =item * Regular Expression
1055 C<m//g> now attaches its state to the searched string rather than the
1056 regular expression. (Once the scope of a block is left for the sub, the
1057 state of the searched string is lost)
1063 sub doit{local($_) = shift; print "Got $_ "}
1065 # perl4 prints: blah blah blah
1066 # perl5 prints: infinite loop blah...
1068 =item * Regular Expression
1070 Currently, if you use the C<m//o> qualifier on a regular expression
1071 within an anonymous sub, I<all> closures generated from that anonymous
1072 sub will use the regular expression as it was compiled when it was used
1073 the very first time in any such closure. For instance, if you say
1076 my($left,$right) = @_;
1077 return sub { $_[0] =~ /$left stuff $right/o; };
1080 build_match() will always return a sub which matches the contents of
1081 C<$left> and C<$right> as they were the I<first> time that build_match()
1082 was called, not as they are in the current call.
1084 This is probably a bug, and may change in future versions of Perl.
1086 =item * Regular Expression
1088 If no parentheses are used in a match, Perl4 sets C<$+> to
1089 the whole match, just like C<$&>. Perl5 does not.
1094 # perl4 prints: bcde
1097 =item * Regular Expression
1099 substitution now returns the null string if it fails
1102 $value = ($string =~ s/foo//);
1108 Also see L<Numerical Traps> for another example of this new feature.
1110 =item * Regular Expression
1112 C<s`lhs`rhs`> (using backticks) is now a normal substitution, with no
1116 $string =~ s`^`hostname`;
1117 print $string, "\n";
1119 # perl4 prints: <the local hostname>
1120 # perl5 prints: hostname
1122 =item * Regular Expression
1124 Stricter parsing of variables used in regular expressions
1126 s/^([^$grpc]*$grpc[$opt$plus$rep]?)//o;
1128 # perl4: compiles w/o error
1129 # perl5: with Scalar found where operator expected ..., near "$opt$plus"
1131 an added component of this example, apparently from the same script, is
1132 the actual value of the s'd string after the substitution.
1133 C<[$opt]> is a character class in perl4 and an array subscript in perl5
1138 s/^([^$grpc]*$grpc[$opt]?)/foo/;
1142 # perl5 prints: foobar
1144 =item * Regular Expression
1146 Under perl5, C<m?x?> matches only once, like C<?x?>. Under perl4, it matched
1147 repeatedly, like C</x/> or C<m!x!>.
1150 sub match { $test =~ m?once?; }
1153 # m?x? matches more then once
1156 # m?x? matches only once
1160 # perl4 prints: perl4
1161 # perl5 prints: perl5
1164 =item * Regular Expression
1166 Under perl4 and upto version 5.003, a failed C<m//g> match used to
1167 reset the internal iterator, so that subsequent C<m//g> match attempts
1168 began from the beginning of the string. In perl version 5.004 and later,
1169 failed C<m//g> matches do not reset the iterator position (which can be
1170 found using the C<pos()> function--see L<perlfunc/pos>).
1174 print $1 while ($test =~ /(o)/g);
1175 # pos $test = 0; # to get old behavior
1178 # perl4 prints: oooooo
1179 # perl5.004 prints: oo
1181 You may always reset the iterator yourself as shown in the commented line
1182 to get the old behavior.
1186 =head2 Subroutine, Signal, Sorting Traps
1188 The general group of Perl4-to-Perl5 traps having to do with
1189 Signals, Sorting, and their related subroutines, as well as
1190 general subroutine traps. Includes some OS-Specific traps.
1196 Barewords that used to look like strings to Perl will now look like subroutine
1197 calls if a subroutine by that name is defined before the compiler sees them.
1199 sub SeeYa { warn"Hasta la vista, baby!" }
1200 $SIG{'TERM'} = SeeYa;
1201 print "SIGTERM is now $SIG{'TERM'}\n";
1203 # perl4 prints: SIGTERM is main'SeeYa
1204 # perl5 prints: SIGTERM is now main::1
1206 Use B<-w> to catch this one
1208 =item * (Sort Subroutine)
1210 reverse is no longer allowed as the name of a sort subroutine.
1212 sub reverse{ print "yup "; $a <=> $b }
1213 print sort reverse a,b,c;
1215 # perl4 prints: yup yup yup yup abc
1218 =item * warn() won't let you specify a filehandle.
1220 Although it _always_ printed to STDERR, warn() would let you specify a
1221 filehandle in perl4. With perl5 it does not.
1225 # perl4 prints: Foo!
1226 # perl5 prints: String found where operator expected
1236 Under HPUX, and some other SysV OSes, one had to reset any signal handler,
1237 within the signal handler function, each time a signal was handled with
1238 perl4. With perl5, the reset is now done correctly. Any code relying
1239 on the handler _not_ being reset will have to be reworked.
1241 Since version 5.002, Perl uses sigaction() under SysV.
1246 $SIG{'INT'} = 'gotit';
1255 while (1) {sleep(10);}
1258 # perl4 (HPUX) prints: Got INT...
1259 # perl5 (HPUX) prints: Got INT... Got INT...
1263 Under SysV OSes, C<seek()> on a file opened to append C<E<gt>E<gt>> now does
1264 the right thing w.r.t. the fopen() manpage. e.g., - When a file is opened
1265 for append, it is impossible to overwrite information already in
1268 open(TEST,">>seek.test");
1269 $start = tell TEST ;
1274 seek(TEST,$start,0);
1275 print TEST "18 characters here";
1277 # perl4 (solaris) seek.test has: 18 characters here
1278 # perl5 (solaris) seek.test has: 1 2 3 4 5 6 7 8 9 18 characters here
1284 =head2 Interpolation Traps
1286 Perl4-to-Perl5 traps having to do with how things get interpolated
1287 within certain expressions, statements, contexts, or whatever.
1291 =item * Interpolation
1293 @ now always interpolates an array in double-quotish strings.
1295 print "To: someone@somewhere.com\n";
1297 # perl4 prints: To:someone@somewhere.com
1298 # perl5 errors : In string, @somewhere now must be written as \@somewhere
1300 =item * Interpolation
1302 Double-quoted strings may no longer end with an unescaped $ or @.
1306 print "foo is $foo, bar is $bar\n";
1308 # perl4 prints: foo is foo$, bar is bar@
1309 # perl5 errors: Final $ should be \$ or $name
1311 Note: perl5 DOES NOT error on the terminating @ in $bar
1313 =item * Interpolation
1315 Perl now sometimes evaluates arbitrary expressions inside braces that occur
1316 within double quotes (usually when the opening brace is preceded by C<$>
1322 sub foo { return "bar" };
1323 print "|@{w.w.w}|${main'foo}|";
1325 # perl4 prints: |@{w.w.w}|foo|
1326 # perl5 prints: |buz|bar|
1328 Note that you can C<use strict;> to ward off such trappiness under perl5.
1330 =item * Interpolation
1332 The construct "this is $$x" used to interpolate the pid at that
1333 point, but now apparently tries to dereference C<$x>. C<$$> by itself still
1334 works fine, however.
1336 print "this is $$x\n";
1338 # perl4 prints: this is XXXx (XXX is the current pid)
1339 # perl5 prints: this is
1341 =item * Interpolation
1343 Creation of hashes on the fly with C<eval "EXPR"> now requires either both
1344 C<$>'s to be protected in the specification of the hash name, or both curlies
1345 to be protected. If both curlies are protected, the result will be compatible
1346 with perl4 and perl5. This is a very common practice, and should be changed
1347 to use the block form of C<eval{}> if possible.
1349 $hashname = "foobar";
1352 eval "\$$hashname{'$key'} = q|$value|";
1353 (defined($foobar{'baz'})) ? (print "Yup") : (print "Nope");
1356 # perl5 prints: Nope
1360 eval "\$$hashname{'$key'} = q|$value|";
1364 eval "\$\$hashname{'$key'} = q|$value|";
1366 causes the following result:
1368 # perl4 prints: Nope
1373 eval "\$$hashname\{'$key'\} = q|$value|";
1375 causes the following result:
1379 # and is compatible for both versions
1382 =item * Interpolation
1384 perl4 programs which unconsciously rely on the bugs in earlier perl versions.
1386 perl -e '$bar=q/not/; print "This is $foo{$bar} perl5"'
1388 # perl4 prints: This is not perl5
1389 # perl5 prints: This is perl5
1391 =item * Interpolation
1393 You also have to be careful about array references.
1398 perl 5 prints: syntax error
1400 =item * Interpolation
1402 Similarly, watch out for:
1405 print "\$$foo{bar}\n";
1407 # perl4 prints: $array{bar}
1410 Perl 5 is looking for C<$array{bar}> which doesn't exist, but perl 4 is
1411 happy just to expand $foo to "array" by itself. Watch out for this
1412 especially in C<eval>'s.
1414 =item * Interpolation
1416 C<qq()> string passed to C<eval>
1419 foreach \$y (keys %\$x\) {
1424 # perl4 runs this ok
1425 # perl5 prints: Can't find string terminator ")"
1437 Existing dbm databases created under perl4 (or any other dbm/ndbm tool)
1438 may cause the same script, run under perl5, to fail. The build of perl5
1439 must have been linked with the same dbm/ndbm as the default for C<dbmopen()>
1440 to function properly without C<tie>'ing to an extension dbm implementation.
1442 dbmopen (%dbm, "file", undef);
1446 # perl5 prints: ok (IFF linked with -ldbm or -lndbm)
1451 Existing dbm databases created under perl4 (or any other dbm/ndbm tool)
1452 may cause the same script, run under perl5, to fail. The error generated
1453 when exceeding the limit on the key/value size will cause perl5 to exit
1456 dbmopen(DB, "testdb",0600) || die "couldn't open db! $!";
1457 $DB{'trap'} = "x" x 1024; # value too large for most dbm/ndbm
1461 dbm store returned -1, errno 28, key "trap" at - line 3.
1465 dbm store returned -1, errno 28, key "trap" at - line 3.
1469 =head2 Unclassified Traps
1475 =item * C<require>/C<do> trap using returned value
1477 If the file doit.pl has:
1485 And the do.pl file has the following single line:
1489 Running doit.pl gives the following:
1491 # perl 4 prints: 3 (aborts the subroutine early)
1494 Same behavior if you replace C<do> with C<require>.
1496 =item * C<split> on empty string with LIMIT specified
1499 @list = split(/foo/, $string, 2)
1501 Perl4 returns a one element list containing the empty string but Perl5
1502 returns an empty list.
1506 As always, if any of these are ever officially declared as bugs,
1507 they'll be fixed and removed.