1 # -*- Mode: cperl; cperl-indent-level: 4 -*-
2 # $Id: Harness.pm,v 1.11 2001/05/23 18:24:41 schwern Exp $
12 use vars qw($VERSION $Verbose $Switches $Have_Devel_Corestack $Curtest
13 $Columns $verbose $switches
14 @ISA @EXPORT @EXPORT_OK
17 # Backwards compatibility for exportable variable names.
19 *switches = \$Switches;
21 $Have_Devel_Corestack = 0;
25 $ENV{HARNESS_ACTIVE} = 1;
27 # Some experimental versions of OS/2 build have broken $?
28 my $Ignore_Exitcode = $ENV{HARNESS_IGNORE_EXITCODE};
30 my $Files_In_Dir = $ENV{HARNESS_FILELEAK_IN_DIR};
34 @EXPORT = qw(&runtests);
35 @EXPORT_OK = qw($verbose $switches);
39 $Columns = $ENV{HARNESS_COLUMNS} || $ENV{COLUMNS} || 80;
40 $Columns--; # Some shells have trouble with a full line of text.
45 Test::Harness - run perl standard test scripts with statistics
51 runtests(@test_files);
55 B<STOP!> If all you want to do is write a test script, consider using
56 Test::Simple. Otherwise, read on.
58 (By using the Test module, you can write test scripts without
59 knowing the exact output this module expects. However, if you need to
60 know the specifics, read on!)
62 Perl test scripts print to standard output C<"ok N"> for each single
63 test, where C<N> is an increasing sequence of integers. The first line
64 output by a standard test script is C<"1..M"> with C<M> being the
65 number of tests that should be run within the test
66 script. Test::Harness::runtests(@tests) runs all the testscripts
67 named as arguments and checks standard output for the expected
70 After all tests have been performed, runtests() prints some
71 performance statistics that are computed by the Benchmark module.
73 =head2 The test script output
75 The following explains how Test::Harness interprets the output of your
82 This header tells how many tests there will be. It should be the
83 first line output by your test program (but its okay if its preceded
86 In certain instanced, you may not know how many tests you will
87 ultimately be running. In this case, it is permitted (but not
88 encouraged) for the 1..M header to appear as the B<last> line output
89 by your test (again, it can be followed by further comments). But we
90 strongly encourage you to put it first.
92 Under B<no> circumstances should 1..M appear in the middle of your
93 output or more than once.
96 =item B<'ok', 'not ok'. Ok?>
98 Any output from the testscript to standard error is ignored and
99 bypassed, thus will be seen by the user. Lines written to standard
100 output containing C</^(not\s+)?ok\b/> are interpreted as feedback for
101 runtests(). All other lines are discarded.
103 C</^not ok/> indicates a failed test. C</^ok/> is a successful test.
106 =item B<test numbers>
108 Perl normally expects the 'ok' or 'not ok' to be followed by a test
109 number. It is tolerated if the test numbers after 'ok' are
110 omitted. In this case Test::Harness maintains temporarily its own
111 counter until the script supplies test numbers again. So the following
126 Failed 3/6 tests, 50.00% okay
129 =item B<$Test::Harness::verbose>
131 The global variable $Test::Harness::verbose is exportable and can be
132 used to let runtests() display the standard output of the script
133 without altering the behavior otherwise.
135 =item B<$Test::Harness::switches>
137 The global variable $Test::Harness::switches is exportable and can be
138 used to set perl command line options used for running the test
139 script(s). The default value is C<-w>.
141 =item B<Skipping tests>
143 If the standard output line contains the substring C< # Skip> (with
144 variations in spacing and case) after C<ok> or C<ok NUMBER>, it is
145 counted as a skipped test. If the whole testscript succeeds, the
146 count of skipped tests is included in the generated output.
147 C<Test::Harness> reports the text after C< # Skip\S*\s+> as a reason
150 ok 23 # skip Insufficient flogiston pressure.
152 Similarly, one can include a similar explanation in a C<1..0> line
153 emitted if the test script is skipped completely:
155 1..0 # Skipped: no leverage found
159 If the standard output line contains the substring C< # TODO> after
160 C<not ok> or C<not ok NUMBER>, it is counted as a todo test. The text
161 afterwards is the thing that has to be done before this test will
164 not ok 13 # TODO harness the power of the atom
166 These tests represent a feature to be implemented or a bug to be fixed
167 and act as something of an executable "thing to do" list. They are
168 B<not> expected to succeed. Should a todo test begin succeeding,
169 Test::Harness will report it as a bonus. This indicates that whatever
170 you were supposed to do has been done and you should promote this to a
175 As an emergency measure, a test script can decide that further tests
176 are useless (e.g. missing dependencies) and testing should stop
177 immediately. In that case the test script prints the magic words
181 to standard output. Any message after these words will be displayed by
182 C<Test::Harness> as the reason why testing is stopped.
186 Additional comments may be put into the testing output on their own
187 lines. Comment lines should begin with a '#', Test::Harness will
191 # Life is good, the sun is shining, RAM is cheap.
193 # got 'Bush' expected 'Gore'
195 =item B<Anything else>
197 Any other output Test::Harness sees it will silently ignore B<BUT WE
198 PLAN TO CHANGE THIS!> If you wish to place additional output in your
199 test script, please use a comment.
206 It will happen, your tests will fail. After you mop up your ego, you
207 can begin examining the summary report:
209 t/base..............ok
210 t/nonumbers.........ok
211 t/ok................ok
212 t/test-harness......ok
213 t/waterloo..........dubious
214 Test returned status 3 (wstat 768, 0x300)
215 DIED. FAILED tests 1, 3, 5, 7, 9, 11, 13, 15, 17, 19
216 Failed 10/20 tests, 50.00% okay
217 Failed Test Stat Wstat Total Fail Failed List of Failed
218 -----------------------------------------------------------------------
219 t/waterloo.t 3 768 20 10 50.00% 1 3 5 7 9 11 13 15 17 19
220 Failed 1/5 test scripts, 80.00% okay. 10/44 subtests failed, 77.27% okay.
222 Everything passed but t/waterloo.t. It failed 10 of 20 tests and
223 exited with non-zero status indicating something dubious happened.
225 The columns in the summary report mean:
231 The test file which failed.
235 If the test exited with non-zero, this is its exit status.
239 The wait status of the test I<umm, I need a better explanation here>.
243 Total number of tests expected to run.
247 Number which failed, either from "not ok" or because they never ran.
251 Percentage of the total tests which failed.
253 =item B<List of Failed>
255 A list of the tests which failed. Successive failures may be
256 abbreviated (ie. 15-20 to indicate that tests 15, 16, 17, 18, 19 and
264 Test::Harness currently only has one function, here it is.
270 my $allok = runtests(@test_files);
272 This runs all the given @test_files and divines whether they passed
273 or failed based on their output to STDOUT (details above). It prints
274 out each individual test which failed along with a summary report and
275 a how long it all took.
277 It returns true if everything was ok, false otherwise.
280 This is just _run_all_tests() plus _show_results()
289 my($tot, $failedtests) = _run_all_tests(@tests);
290 _show_results($tot, $failedtests);
292 my $ok = ($tot->{bad} == 0 && $tot->{max});
294 die q{Assert '$ok xor keys %$failedtests' failed!}
295 unless $ok xor keys %$failedtests;
304 my @files = _globdir $dir;
306 Returns all the files in a directory. This is shorthand for backwards
307 compatibility on systems where glob() doesn't work right.
313 my @f = readdir DIRH;
319 =item B<_run_all_tests>
321 my($total, $failed) = _run_all_tests(@test_files);
323 Runs all the given @test_files (as runtests()) but does it quietly (no
324 report). $total is a hash ref summary of all the tests run. Its keys
327 bonus Number of individual todo tests unexpectedly passed
328 max Number of individual tests ran
329 ok Number of individual tests passed
330 sub_skipped Number of individual tests skipped
332 files Number of test files ran
333 good Number of test files passed
334 bad Number of test files failed
335 tests Number of test files originally given
336 skipped Number of test files skipped
338 If $total->{bad} == 0 and $total->{max} > 0, you've got a successful
341 $failed is a hash ref of all the test scripts which failed. Each key
342 is the name of a test script, each value is another hash representing
343 how that script failed. Its keys are these:
345 name Name of the test which failed
346 estat Script's exit value
347 wstat Script's wait status
348 max Number of individual tests
349 failed Number which failed
350 percent Percentage of tests which failed
351 canon List of tests which failed (as string).
353 Needless to say, $failed should be empty if everything passed.
355 B<NOTE> Currently this function is still noisy. I'm working on it.
372 tests => scalar @tests,
378 # pass -I flags to children
379 my $old5lib = $ENV{PERL5LIB};
381 # VMS has a 255-byte limit on the length of %ENV entries, so
382 # toss the ones that involve perl_root, the install location
386 $new5lib = join($Config{path_sep}, grep {!/perl_root/i;} @INC);
387 $Switches =~ s/-(\S*[A-Z]\S*)/"-$1"/g;
390 $new5lib = join($Config{path_sep}, @INC);
393 local($ENV{'PERL5LIB'}) = $new5lib;
395 my @dir_files = _globdir $Files_In_Dir if defined $Files_In_Dir;
396 my $t_start = new Benchmark;
401 $maxlen = $len if $len > $maxlen;
403 # +3 : we want three dots between the test name and the "ok"
405 my $width = $maxlen + 3 - 2;
406 foreach my $tfile (@tests) {
407 my($leader, $ml) = _mk_leader($tfile, $width);
410 my $fh = _open_test($tfile);
412 # state of the current test.
421 skip_reason => undef,
425 my($seen_header, $tests_seen) = (0,0);
427 if( _parse_header($_, \%test, \%tot) ) {
428 warn "Test header seen twice!\n" if $seen_header;
432 warn "1..M can only appear at the beginning or end of tests\n"
433 if $tests_seen && $test{max} < $tests_seen;
435 elsif( _parse_test_line($_, \%test, \%tot) ) {
441 my($estatus, $wstatus) = _close_fh($fh);
443 my $allok = $test{ok} == $test{max} && $test{'next'} == $test{max}+1;
446 $failedtests{$tfile} = _dubious_return(\%test, \%tot,
448 $failedtests{$tfile}{name} = $tfile;
451 if ($test{max} and $test{skipped} + $test{bonus}) {
453 push(@msg, "$test{skipped}/$test{max} skipped: $test{skip_reason}")
455 push(@msg, "$test{bonus}/$test{max} unexpectedly succeeded")
457 print "$test{ml}ok, ".join(', ', @msg)."\n";
458 } elsif ($test{max}) {
459 print "$test{ml}ok\n";
460 } elsif (defined $test{skip_reason}) {
461 print "skipped: $test{skip_reason}\n";
464 print "skipped test on this platform\n";
471 if ($test{'next'} <= $test{max}) {
472 push @{$test{failed}}, $test{'next'}..$test{max};
474 if (@{$test{failed}}) {
475 my ($txt, $canon) = canonfailed($test{max},$test{skipped},
477 print "$test{ml}$txt";
478 $failedtests{$tfile} = { canon => $canon,
480 failed => scalar @{$test{failed}},
482 percent => 100*(scalar @{$test{failed}})/$test{max},
487 print "Don't know which tests failed: got $test{ok} ok, ".
488 "expected $test{max}\n";
489 $failedtests{$tfile} = { canon => '??',
499 } elsif ($test{'next'} == 0) {
500 print "FAILED before any test output arrived\n";
502 $failedtests{$tfile} = { canon => '??',
513 $tot{sub_skipped} += $test{skipped};
515 if (defined $Files_In_Dir) {
516 my @new_dir_files = _globdir $Files_In_Dir;
517 if (@new_dir_files != @dir_files) {
519 @f{@new_dir_files} = (1) x @new_dir_files;
520 delete @f{@dir_files};
521 my @f = sort keys %f;
522 print "LEAKED FILES: @f\n";
523 @dir_files = @new_dir_files;
527 $tot{bench} = timediff(new Benchmark, $t_start);
530 if (defined $old5lib) {
531 $ENV{PERL5LIB} = $old5lib;
533 delete $ENV{PERL5LIB};
537 return(\%tot, \%failedtests);
542 my($leader, $ml) = _mk_leader($test_file, $width);
544 Generates the 't/foo........' $leader for the given $test_file as well
545 as a similar version which will overwrite the current line (by use of
546 \r and such). $ml may be empty if Test::Harness doesn't think you're
547 on TTY. The width is the width of the "yada/blah..." string.
552 my ($te, $width) = @_;
554 chop($te); # XXX chomp?
556 if ($^O eq 'VMS') { $te =~ s/^.*\.t\./\[.t./s; }
557 my $blank = (' ' x 77);
558 my $leader = "$te" . '.' x ($width - length($te));
561 $ml = "\r$blank\r$leader"
562 if -t STDOUT and not $ENV{HARNESS_NOTTY} and not $Verbose;
564 return($leader, $ml);
569 my($tot, $failedtests) = @_;
572 my $bonusmsg = _bonusmsg($tot);
574 if ($tot->{bad} == 0 && $tot->{max}) {
575 print "All tests successful$bonusmsg.\n";
576 } elsif ($tot->{tests}==0){
577 die "FAILED--no tests were run for some reason.\n";
578 } elsif ($tot->{max} == 0) {
579 my $blurb = $tot->{tests}==1 ? "script" : "scripts";
580 die "FAILED--$tot->{tests} test $blurb could be run, ".
581 "alas--no output ever seen\n";
583 $pct = sprintf("%.2f", $tot->{good} / $tot->{tests} * 100);
584 my $subpct = sprintf " %d/%d subtests failed, %.2f%% okay.",
585 $tot->{max} - $tot->{ok}, $tot->{max},
586 100*$tot->{ok}/$tot->{max};
588 my($fmt_top, $fmt) = _create_fmts($failedtests);
590 # Now write to formats
591 for my $script (sort keys %$failedtests) {
592 $Curtest = $failedtests->{$script};
596 $bonusmsg =~ s/^,\s*//;
597 print "$bonusmsg.\n" if $bonusmsg;
598 die "Failed $tot->{bad}/$tot->{tests} test scripts, $pct% okay.".
603 printf("Files=%d, Tests=%d, %s\n",
604 $tot->{files}, $tot->{max}, timestr($tot->{bench}, 'nop'));
609 my($line, $test, $tot) = @_;
613 print $line if $Verbose;
616 if ($line =~ /^1\.\.([0-9]+) todo([\d\s]+);?/i) {
618 for (split(/\s+/, $2)) { $test->{todo}{$_} = 1; }
620 $tot->{max} += $test->{max};
626 # 1..0 # skip Why? Because I said so!
627 elsif ($line =~ /^1\.\.([0-9]+)
628 (\s*\#\s*[Ss]kip\S*\s* (.+))?
633 $tot->{max} += $test->{max};
635 $test->{'next'} = 1 unless $test->{'next'};
636 $test->{skip_reason} = $3 if not $test->{max} and defined $3;
651 my $s = _set_switches($test);
653 # XXX This is WAY too core specific!
654 my $cmd = ($ENV{'HARNESS_COMPILE_TEST'})
655 ? "./perl -I../lib ../utils/perlcc $test "
656 . "-r 2>> ./compilelog |"
658 $cmd = "MCR $cmd" if $^O eq 'VMS';
660 if( open(PERL, $cmd) ) {
664 print "can't run $test. $!\n";
676 sub _parse_test_line {
677 my($line, $test, $tot) = @_;
679 if ($line =~ /^(not\s+)?ok\b/i) {
680 my $this = $test->{'next'} || 1;
682 if ($line =~ /^(not )?ok\s*(\d*)(\s*#.*)?/) {
683 my($not, $tnum, $extra) = ($1, $2, $3);
685 $this = $tnum if $tnum;
687 my($type, $reason) = $extra =~ /^\s*#\s*([Ss]kip\S*|TODO)(\s+.+)?/
690 my($istodo, $isskip);
691 if( defined $type ) {
692 $istodo = $type =~ /TODO/;
693 $isskip = $type =~ /skip/i;
696 $test->{todo}{$tnum} = 1 if $istodo;
699 print "$test->{ml}NOK $this" if $test->{ml};
700 if (!$test->{todo}{$this}) {
701 push @{$test->{failed}}, $this;
708 print "$test->{ml}ok $this/$test->{max}" if $test->{ml};
711 $test->{skipped}++ if $isskip;
713 if (defined $reason and defined $test->{skip_reason}) {
714 # print "was: '$skip_reason' new '$reason'\n";
715 $test->{skip_reason} = 'various reasons'
716 if $test->{skip_reason} ne $reason;
717 } elsif (defined $reason) {
718 $test->{skip_reason} = $reason;
721 $test->{bonus}++, $tot->{bonus}++ if $test->{todo}{$this};
725 elsif ($line =~ /^ok\s*(\d*)\s*\#([^\r]*)$/) { # XXX multiline ok?
726 $this = $1 if $1 > 0;
727 print "$test->{ml}ok $this/$test->{max}" if $test->{ml};
732 # an ok or not ok not matching the 3 cases above...
733 # just ignore it for compatibility with TEST
737 if ($this > $test->{'next'}) {
738 # print "Test output counter mismatch [test $this]\n";
739 # no need to warn probably
740 push @{$test->{failed}}, $test->{'next'}..$this-1;
742 elsif ($this < $test->{'next'}) {
743 #we have seen more "ok" lines than the number suggests
744 print "Confused test output: test $this answered after ".
745 "test ", $test->{'next'}-1, "\n";
746 $test->{'next'} = $this;
748 $test->{'next'} = $this + 1;
751 elsif ($line =~ /^Bail out!\s*(.*)/i) { # magic words
752 die "FAILED--Further testing stopped" .
753 ($1 ? ": $1\n" : ".\n");
762 $bonusmsg = (" ($tot->{bonus} subtest".($tot->{bonus} > 1 ? 's' : '').
763 " UNEXPECTEDLY SUCCEEDED)")
766 if ($tot->{skipped}) {
767 $bonusmsg .= ", $tot->{skipped} test"
768 . ($tot->{skipped} != 1 ? 's' : '');
769 if ($tot->{sub_skipped}) {
770 $bonusmsg .= " and $tot->{sub_skipped} subtest"
771 . ($tot->{sub_skipped} != 1 ? 's' : '');
773 $bonusmsg .= ' skipped';
775 elsif ($tot->{sub_skipped}) {
776 $bonusmsg .= ", $tot->{sub_skipped} subtest"
777 . ($tot->{sub_skipped} != 1 ? 's' : '')
784 # VMS has some subtle nastiness with closing the test files.
788 close($fh); # must close to reap child resource values
790 my $wstatus = $Ignore_Exitcode ? 0 : $?; # Can trust $? ?
792 $estatus = ($^O eq 'VMS'
793 ? eval 'use vmsish "status"; $estatus = $?'
796 return($estatus, $wstatus);
800 # Set up the command-line switches to run perl as.
805 open(TEST, $test) or print "can't open $test. $!\n";
808 $s .= " $ENV{'HARNESS_PERL_SWITCHES'}"
809 if exists $ENV{'HARNESS_PERL_SWITCHES'};
810 $s .= join " ", q[ "-T"], map {qq["-I$_"]} @INC
811 if $first =~ /^#!.*\bperl.*-\w*T/;
813 close(TEST) or print "can't close $test. $!\n";
819 # Test program go boom.
820 sub _dubious_return {
821 my($test, $tot, $estatus, $wstatus) = @_;
822 my ($failed, $canon, $percent) = ('??', '??');
824 printf "$test->{ml}dubious\n\tTest returned status $estatus ".
825 "(wstat %d, 0x%x)\n",
827 print "\t\t(VMS status is $estatus)\n" if $^O eq 'VMS';
829 if (corestatus($wstatus)) { # until we have a wait module
830 if ($Have_Devel_Corestack) {
831 Devel::CoreStack::stack($^X);
833 print "\ttest program seems to have generated a core\n";
840 if ($test->{'next'} == $test->{max} + 1 and not @{$test->{failed}}) {
841 print "\tafter all the subtests completed successfully\n";
843 $failed = 0; # But we do not set $canon!
846 push @{$test->{failed}}, $test->{'next'}..$test->{max};
847 $failed = @{$test->{failed}};
848 (my $txt, $canon) = canonfailed($test->{max},$test->{skipped},@{$test->{failed}});
849 $percent = 100*(scalar @{$test->{failed}})/$test->{max};
854 return { canon => $canon, max => $test->{max} || '??',
857 estat => $estatus, wstat => $wstatus,
862 sub _garbled_output {
863 my($gibberish) = shift;
864 warn "Confusing test output: '$gibberish'\n";
869 my($failedtests) = @_;
871 my $failed_str = "Failed Test";
872 my $middle_str = " Stat Wstat Total Fail Failed ";
873 my $list_str = "List of Failed";
875 # Figure out our longest name string for formatting purposes.
876 my $max_namelen = length($failed_str);
877 foreach my $script (keys %$failedtests) {
878 my $namelen = length $failedtests->{$script}->{name};
879 $max_namelen = $namelen if $namelen > $max_namelen;
882 my $list_len = $Columns - length($middle_str) - $max_namelen;
883 if ($list_len < length($list_str)) {
884 $list_len = length($list_str);
885 $max_namelen = $Columns - length($middle_str) - $list_len;
886 if ($max_namelen < length($failed_str)) {
887 $max_namelen = length($failed_str);
888 $Columns = $max_namelen + length($middle_str) + $list_len;
892 my $fmt_top = "format STDOUT_TOP =\n"
893 . sprintf("%-${max_namelen}s", $failed_str)
899 my $fmt = "format STDOUT =\n"
900 . "@" . "<" x ($max_namelen - 1)
901 . " @>> @>>>> @>>>> @>>> ^##.##% "
902 . "^" . "<" x ($list_len - 1) . "\n"
903 . '{ $Curtest->{name}, $Curtest->{estat},'
904 . ' $Curtest->{wstat}, $Curtest->{max},'
905 . ' $Curtest->{failed}, $Curtest->{percent},'
906 . ' $Curtest->{canon}'
908 . "~~" . " " x ($Columns - $list_len - 2) . "^"
909 . "<" x ($list_len - 1) . "\n"
910 . '$Curtest->{canon}'
918 return($fmt_top, $fmt);
922 my $tried_devel_corestack;
927 eval {require 'wait.ph'};
928 my $ret = defined &WCOREDUMP ? WCOREDUMP($st) : $st & 0200;
930 eval { require Devel::CoreStack; $Have_Devel_Corestack++ }
931 unless $tried_devel_corestack++;
937 sub canonfailed ($@) {
938 my($max,$skipped,@failed) = @_;
940 @failed = sort {$a <=> $b} grep !$seen{$_}++, @failed;
941 my $failed = @failed;
945 my $last = $min = shift @failed;
948 for (@failed, $failed[-1]) { # don't forget the last one
949 if ($_ > $last+1 || $_ == $last) {
953 push @canon, "$min-$last";
960 push @result, "FAILED tests @canon\n";
961 $canon = join ' ', @canon;
963 push @result, "FAILED test $last\n";
967 push @result, "\tFailed $failed/$max tests, ";
968 push @result, sprintf("%.2f",100*(1-$failed/$max)), "% okay";
969 my $ender = 's' x ($skipped > 1);
970 my $good = $max - $failed - $skipped;
971 my $goodper = sprintf("%.2f",100*($good/$max));
972 push @result, " (-$skipped skipped test$ender: $good okay, ".
976 my $txt = join "", @result;
993 C<&runtests> is exported by Test::Harness per default.
995 C<$verbose> and C<$switches> are exported upon request.
1002 =item C<All tests successful.\nFiles=%d, Tests=%d, %s>
1004 If all tests are successful some statistics about the performance are
1007 =item C<FAILED tests %s\n\tFailed %d/%d tests, %.2f%% okay.>
1009 For any single script that has failing subtests statistics like the
1012 =item C<Test returned status %d (wstat %d)>
1014 Scripts that return a non-zero exit status, both C<$? E<gt>E<gt> 8>
1015 and C<$?> are printed in a message similar to the above.
1017 =item C<Failed 1 test, %.2f%% okay. %s>
1019 =item C<Failed %d/%d tests, %.2f%% okay. %s>
1021 If not all tests were successful, the script dies with one of the
1024 =item C<FAILED--Further testing stopped%s>
1026 If a single subtest decides that further testing will not make sense,
1027 the script dies with this message.
1035 =item C<HARNESS_IGNORE_EXITCODE>
1037 Makes harness ignore the exit status of child processes when defined.
1039 =item C<HARNESS_NOTTY>
1041 When set to a true value, forces it to behave as though STDOUT were
1042 not a console. You may need to set this if you don't want harness to
1043 output more frequent progress messages using carriage returns. Some
1044 consoles may not handle carriage returns properly (which results in a
1045 somewhat messy output).
1047 =item C<HARNESS_COMPILE_TEST>
1049 When true it will make harness attempt to compile the test using
1050 C<perlcc> before running it.
1052 B<NOTE> This currently only works when sitting in the perl source
1055 =item C<HARNESS_FILELEAK_IN_DIR>
1057 When set to the name of a directory, harness will check after each
1058 test whether new files appeared in that directory, and report them as
1060 LEAKED FILES: scr.tmp 0 my.db
1062 If relative, directory name is with respect to the current directory at
1063 the moment runtests() was called. Putting absolute path into
1064 C<HARNESS_FILELEAK_IN_DIR> may give more predicatable results.
1066 =item C<HARNESS_PERL_SWITCHES>
1068 Its value will be prepended to the switches used to invoke perl on
1069 each test. For example, setting C<HARNESS_PERL_SWITCHES> to C<-W> will
1070 run all tests with all warnings enabled.
1072 =item C<HARNESS_COLUMNS>
1074 This value will be used for the width of the terminal. If it is not
1075 set then it will default to C<COLUMNS>. If this is not set, it will
1076 default to 80. Note that users of Bourne-sh based shells will need to
1077 C<export COLUMNS> for this module to use that variable.
1079 =item C<HARNESS_ACTIVE>
1081 Harness sets this before executing the individual tests. This allows
1082 the tests to determine if they are being executed through the harness
1083 or by any other means.
1089 Here's how Test::Harness tests itself
1091 $ cd ~/src/devel/Test-Harness
1092 $ perl -Mblib -e 'use Test::Harness qw(&runtests $verbose);
1093 $verbose=0; runtests @ARGV;' t/*.t
1094 Using /home/schwern/src/devel/Test-Harness/blib
1095 t/base..............ok
1096 t/nonumbers.........ok
1097 t/ok................ok
1098 t/test-harness......ok
1099 All tests successful.
1100 Files=4, Tests=24, 2 wallclock secs ( 0.61 cusr + 0.41 csys = 1.02 CPU)
1104 L<Test> and L<Test::Simple> for writing test scripts, L<Benchmark> for
1105 the underlying timing routines, L<Devel::CoreStack> to generate core
1106 dumps from failed tests and L<Devel::Cover> for test coverage
1111 Either Tim Bunce or Andreas Koenig, we don't know. What we know for
1112 sure is, that it was inspired by Larry Wall's TEST script that came
1113 with perl distributions for ages. Numerous anonymous contributors
1114 exist. Andreas Koenig held the torch for many years.
1116 Current maintainer is Michael G Schwern E<lt>schwern@pobox.comE<gt>
1120 Provide a way of running tests quietly (ie. no printing) for automated
1121 validation of tests. This will probably take the form of a version
1122 of runtests() which rather than printing its output returns raw data
1123 on the state of the tests.
1125 Fix HARNESS_COMPILE_TEST without breaking its core usage.
1127 Figure a way to report test names in the failure summary.
1129 Rework the test summary so long test names are not truncated as badly.
1131 Merge back into bleadperl.
1133 Deal with VMS's "not \nok 4\n" mistake.
1135 Add option for coverage analysis.
1138 Keeping whittling away at _run_all_tests()
1141 Clean up how the summary is printed. Get rid of those damned formats.
1145 Test::Harness uses $^X to determine the perl binary to run the tests
1146 with. Test scripts running via the shebang (C<#!>) line may not be
1147 portable because $^X is not consistent for shebang scripts across
1148 platforms. This is no problem when Test::Harness is run with an
1149 absolute path to the perl binary or when $^X can be found in the path.
1151 HARNESS_COMPILE_TEST currently assumes its run from the Perl source