1N/A chdir 't' if -d 't'; 1N/Ais('-' x 5, '-----', 'compile time x');
1N/Ais('-' x -1, '', ' x -1');
1N/Ais('ab' x 3, 'ababab', ' more than one char');
1N/Ais($a x 5, '-----', 'run time x');
1N/Ais($a x 3, 'ababab', ' more than one char');
1N/Ais($a x 0, '', ' more than one char');
1N/Ais($a x -12, '', ' more than one char');
1N/Ais(join('', (@x) x 4), '123123123123', '(@x) x Y');
1N/Ais(join('', (@x,()) x 4), '123123123123', '(@x,()) x Y');
1N/Ais(join('', (@x,1) x 4), '1231123112311231', '(@x,1) x Y');
1N/Ais(join(':', (9,9) x 4), '9:9:9:9:9:9:9:9', '(X,X) x Y');
1N/A# This test is actually testing for Digital C compiler optimizer bug, 1N/A# present in Dec C versions 5.* and 6.0 (used in Digital UNIX and VMS), 1N/A# found in December 1998. The bug was reported to Digital^WCompaq as 1N/A# DECC 2745 (21-Dec-1998) 1N/A# GEM_BUGS 7619 (23-Dec-1998) 1N/A# As of April 1999 the bug has been fixed in Tru64 UNIX 5.0 and is planned 1N/A# to be fixed also in 4.0G. 1N/A# The bug was as follows: broken code was produced for util.c:repeatcpy() 1N/A# (a utility function for the 'x' operator) in the case *all* these 1N/A# four conditions held: 1N/A# (2) "from" had the 8th bit on in its single character 1N/A# (3) count > 7 (the 'x' count > 16) 1N/A# (4) the highest optimization level was used in compilation 1N/A# (which is the default when compiling Perl) 1N/A# The bug looked like this (. being the eight-bit character and ? being \xff): 1N/A# 16 ................ 1N/A# 17 .........???????. 1N/A# 18 .........???????.. 1N/A# 19 .........???????... 1N/A# 20 .........???????.... 1N/A# 21 .........???????..... 1N/A# 22 .........???????...... 1N/A# 23 .........???????....... 1N/A# 24 .........???????.??????? 1N/A# 25 .........???????.???????. 1N/A# The bug was triggered in the "if (len == 1)" branch. The fix 1N/A# was to introduce a new temporary variable. In diff -u format: 1N/A# register char *frombase = from; 1N/A#+ register char c = *from; 1N/A# while (count-- > 0) 1N/A# The bug could also be (obscurely) avoided by changing "from" to 1N/A# be an unsigned char pointer. 1N/A# This obscure bug was not found by the then test suite but instead 1N/A# by Mark.Martinec@nsc.ijs.si while trying to install Digest-MD5-2.00. 1N/Ais("\xdd" x 24, "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd", 'Dec C bug');
1N/A# When we use a list repeat in a scalar context, it behaves like 1N/A# a scalar repeat. Make sure that works properly, and doesn't leave 1N/A# extraneous values on the stack. 1N/A# -- robin@kitsite.com 1N/Ais($x, "22", 'list repeat in scalar context');
1N/A# Make sure the stack doesn't get truncated too much - the left 1N/A# operand of the eq binop needs to remain! 1N/A# perlbug 20011113.110 works in 5.6.1, broken in 5.7.2 1N/A my $x= [("foo") x 2]; 1N/A is( join('', @$x), 'foofoo', 'list repeat in anon array ref broken [ID 20011113.110]' ); 1N/A# [ID 20010809.028] x operator not copying elements in 'for' list? 1N/A local $TODO = "x operator not copying elements in 'for' list? [ID 20010809.028]"; 1N/A for (($x =~ /./g) x 2) {