2005-04-15 19:48:58 -04:00
|
|
|
#!/usr/bin/env perl
|
|
|
|
#***************************************************************************
|
|
|
|
# _ _ ____ _
|
|
|
|
# Project ___| | | | _ \| |
|
|
|
|
# / __| | | | |_) | |
|
|
|
|
# | (__| |_| | _ <| |___
|
|
|
|
# \___|\___/|_| \_\_____|
|
|
|
|
#
|
|
|
|
# Copyright (C) 1998 - 2005, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
|
|
#
|
|
|
|
# This software is licensed as described in the file COPYING, which
|
|
|
|
# you should have received as part of this distribution. The terms
|
|
|
|
# are also available at http://curl.haxx.se/docs/copyright.html.
|
|
|
|
#
|
|
|
|
# You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
|
|
# copies of the Software, and permit persons to whom the Software is
|
|
|
|
# furnished to do so, under the terms of the COPYING file.
|
|
|
|
#
|
|
|
|
# This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
|
|
# KIND, either express or implied.
|
|
|
|
#
|
|
|
|
# $Id$
|
|
|
|
###########################################################################
|
|
|
|
|
|
|
|
use strict;
|
|
|
|
|
|
|
|
@INC=(@INC, $ENV{'srcdir'}, ".");
|
|
|
|
|
|
|
|
require "getpart.pm"; # array functions
|
|
|
|
|
|
|
|
my $srcdir = $ENV{'srcdir'} || '.';
|
|
|
|
my $TESTDIR="$srcdir/data";
|
|
|
|
|
|
|
|
# Get all commands and find out their test numbers
|
|
|
|
opendir(DIR, $TESTDIR) || die "can't opendir $TESTDIR: $!";
|
|
|
|
my @cmds = grep { /^test([0-9]+)$/ && -f "$TESTDIR/$_" } readdir(DIR);
|
|
|
|
closedir DIR;
|
|
|
|
|
|
|
|
my $TESTCASES; # start with no test cases
|
|
|
|
|
|
|
|
# cut off everything but the digits
|
|
|
|
for(@cmds) {
|
|
|
|
$_ =~ s/[a-z\/\.]*//g;
|
|
|
|
}
|
|
|
|
# the the numbers from low to high
|
|
|
|
for(sort { $a <=> $b } @cmds) {
|
|
|
|
$TESTCASES .= " $_";
|
|
|
|
}
|
|
|
|
|
|
|
|
my $t;
|
2005-04-16 08:43:32 -04:00
|
|
|
|
|
|
|
my %k; # keyword count
|
|
|
|
my %t; # keyword to test case mapping
|
|
|
|
my @miss; # test cases without keywords set
|
|
|
|
|
2005-04-16 19:15:12 -04:00
|
|
|
my $count;
|
|
|
|
|
2005-04-27 06:12:41 -04:00
|
|
|
my %errors;
|
|
|
|
|
2005-04-15 19:48:58 -04:00
|
|
|
for $t (split(/ /, $TESTCASES)) {
|
|
|
|
if(loadtest("${TESTDIR}/test${t}")) {
|
|
|
|
# bad case
|
|
|
|
next;
|
|
|
|
}
|
2005-04-27 06:12:41 -04:00
|
|
|
|
|
|
|
my @ec = getpart("verify", "errorcode");
|
|
|
|
if($ec[0]) {
|
|
|
|
# count number of check error codes
|
|
|
|
$errors{ 0 + $ec[0] } ++;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-04-15 19:48:58 -04:00
|
|
|
my @what = getpart("info", "keywords");
|
|
|
|
|
2005-04-16 08:43:32 -04:00
|
|
|
if(!$what[0]) {
|
|
|
|
push @miss, $t;
|
2005-04-16 19:15:12 -04:00
|
|
|
next;
|
2005-04-16 08:43:32 -04:00
|
|
|
}
|
|
|
|
|
2005-04-15 19:48:58 -04:00
|
|
|
for(@what) {
|
2005-04-16 08:43:32 -04:00
|
|
|
chomp;
|
|
|
|
#print "Test $t: $_\n";
|
|
|
|
$k{$_}++;
|
2005-04-22 18:29:48 -04:00
|
|
|
$t{$_} .= "$t ";
|
2005-04-15 19:48:58 -04:00
|
|
|
}
|
2005-04-27 06:12:41 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-04-16 19:15:12 -04:00
|
|
|
$count++;
|
2005-04-15 19:48:58 -04:00
|
|
|
}
|
|
|
|
|
2005-04-22 18:29:48 -04:00
|
|
|
sub show {
|
|
|
|
my ($list)=@_;
|
|
|
|
my @a = split(" ", $list);
|
|
|
|
my $ret;
|
|
|
|
|
|
|
|
my $c;
|
|
|
|
my @l = sort {rand(100) - 50} @a;
|
|
|
|
my @ll;
|
|
|
|
|
2005-04-27 06:12:41 -04:00
|
|
|
for(1 .. 11) {
|
2005-04-22 18:29:48 -04:00
|
|
|
my $v = shift @l;
|
|
|
|
if($v) {
|
|
|
|
push @ll, $v;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (sort {$a <=> $b} @ll) {
|
2005-04-27 06:12:41 -04:00
|
|
|
if($c++ == 10) {
|
2005-04-22 18:29:48 -04:00
|
|
|
$ret .= "...";
|
|
|
|
last;
|
|
|
|
}
|
|
|
|
$ret .= "$_ ";
|
|
|
|
}
|
|
|
|
return $ret;
|
|
|
|
}
|
|
|
|
|
2005-04-16 19:15:12 -04:00
|
|
|
# numerically on amount, or alphebetically if same amount
|
|
|
|
my @mtest = reverse sort { $k{$a} <=> $k{$b} || $b cmp $a } keys %k;
|
2005-04-15 19:48:58 -04:00
|
|
|
|
2005-04-16 08:43:32 -04:00
|
|
|
print <<TOP
|
2005-04-22 18:29:48 -04:00
|
|
|
<table><tr><th>Num</th><th>Keyword</th><th>Test Cases</th></tr>
|
2005-04-16 08:43:32 -04:00
|
|
|
TOP
|
|
|
|
;
|
|
|
|
for $t (@mtest) {
|
2005-04-22 18:29:48 -04:00
|
|
|
printf "<tr><td>%d</td><td>$t</td><td>%s</td></tr>\n", $k{$t},
|
|
|
|
show($t{$t});
|
2005-04-16 08:43:32 -04:00
|
|
|
}
|
2005-04-16 19:15:12 -04:00
|
|
|
printf "</table><p> $count tests (%d lack keywords)\n",
|
|
|
|
scalar(@miss);
|
2005-04-16 08:43:32 -04:00
|
|
|
|
|
|
|
for(@miss) {
|
|
|
|
print STDERR "$_ ";
|
|
|
|
}
|
2005-04-16 19:15:12 -04:00
|
|
|
|
|
|
|
print STDERR "\n";
|
2005-04-27 06:12:41 -04:00
|
|
|
|
|
|
|
print "<p>Error codes tested for:<br>\n";
|
|
|
|
|
|
|
|
# numerically on amount, or alphebetically if same amount
|
|
|
|
my @etest = sort { $a <=> $b} keys %errors;
|
|
|
|
|
|
|
|
for(@etest) {
|
|
|
|
print "$_ ";
|
|
|
|
}
|
|
|
|
print "\n";
|