#!/usr/bin/perl -w
-# Copyright (C) 2019-2020 all contributors <meta@public-inbox.org>
+# Copyright (C) 2019-2021 all contributors <meta@public-inbox.org>
# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
#
# Parallel test runner which preloads code and reuses worker processes
# Usage: $PERL -I lib -w t/run.perl -j4
# Or via prove(1): prove -lvw t/run.perl :: -j4
use strict;
+use v5.10.1;
+use IO::Handle; # ->autoflush
use PublicInbox::TestCommon;
-use Cwd qw(getcwd);
+use PublicInbox::Spawn;
use Getopt::Long qw(:config gnu_getopt no_ignore_case auto_abbrev);
use Errno qw(EINTR);
use Fcntl qw(:seek);
-use POSIX qw(_POSIX_PIPE_BUF WNOHANG);
+use POSIX qw(WNOHANG);
+use File::Temp ();
my $jobs = 1;
my $repeat = 1;
$| = 1;
die "$0 is not compatible with TEST_RUN_MODE=0\n";
}
my @tests = scalar(@ARGV) ? @ARGV : glob('t/*.t');
-my $cwd = getcwd();
-open OLDOUT, '>&STDOUT' or die "dup STDOUT: $!";
-open OLDERR, '>&STDERR' or die "dup STDERR: $!";
-OLDOUT->autoflush(1);
-OLDERR->autoflush(1);
+open my $cwd_fh, '<', '.' or die "open .: $!";
+open my $OLDOUT, '>&STDOUT' or die "dup STDOUT: $!";
+open my $OLDERR, '>&STDERR' or die "dup STDERR: $!";
+$OLDOUT->autoflush(1);
+$OLDERR->autoflush(1);
+
+my ($run_log, $tmp_rl);
+my $rl = $ENV{TEST_RUN_LOG};
+unless ($rl) {
+ $tmp_rl = File::Temp->new(CLEANUP => 1);
+ $rl = $tmp_rl->filename;
+}
+open $run_log, '+>>', $rl or die "open $rl: $!";
+$run_log->autoflush(1); # one reader, many writers
key2sub($_) for @tests; # precache
+my ($for_destroy, $lei_env, $lei_daemon_pid, $owner_pid);
+
+# TEST_LEI_DAEMON_PERSIST is currently broken. I get ECONNRESET from
+# lei even with high kern.ipc.soacceptqueue=1073741823 or SOMAXCONN, not
+# sure why. Also, testing our internal inotify usage is unreliable
+# because lei-daemon uses a single inotify FD for all clients.
+if ($ENV{TEST_LEI_DAEMON_PERSIST} && !$ENV{TEST_LEI_DAEMON_PERSIST_DIR} &&
+ (PublicInbox::Spawn->can('recv_cmd4') ||
+ eval { require Socket::MsgHdr })) {
+ $lei_env = {};
+ ($lei_env->{XDG_RUNTIME_DIR}, $for_destroy) = tmpdir;
+ $ENV{TEST_LEI_DAEMON_PERSIST_DIR} = $lei_env->{XDG_RUNTIME_DIR};
+ run_script([qw(lei daemon-pid)], $lei_env, { 1 => \$lei_daemon_pid });
+ chomp $lei_daemon_pid;
+ $lei_daemon_pid =~ /\A[0-9]+\z/ or die "no daemon pid: $lei_daemon_pid";
+ kill(0, $lei_daemon_pid) or die "kill $lei_daemon_pid: $!";
+ if (my $t = $ENV{GNU_TAIL}) {
+ system("$t --pid=$lei_daemon_pid -F " .
+ "$lei_env->{XDG_RUNTIME_DIR}/lei/errors.log >&2 &");
+ }
+ if (my $strace_cmd = $ENV{STRACE_CMD}) {
+ system("$strace_cmd -p $lei_daemon_pid &");
+ }
+ $owner_pid = $$;
+}
+
if ($shuffle) {
require List::Util;
} elsif (open(my $prove_state, '<', '.prove') && eval { require YAML::XS }) {
our $tb = Test::More->builder;
sub DIE (;$) {
- print OLDERR @_;
+ print $OLDERR @_;
exit(1);
}
sub test_status () {
$? = 255 if $? == 0 && !$tb->is_passing;
my $status = $? ? 'not ok' : 'ok';
- chdir($cwd) or DIE "chdir($cwd): $!";
+ chdir($cwd_fh) or DIE "fchdir: $!";
if ($log_suffix ne '') {
my $log = $worker_test;
$log =~ s/\.t\z/$log_suffix/;
my $skip = '';
if (open my $fh, '<', $log) {
my @not_ok = grep(!/^(?:ok |[ \t]*#)/ms, <$fh>);
- pop @not_ok if $not_ok[-1] =~ /^[0-9]+\.\.[0-9]+$/;
+ my $last = $not_ok[-1] // '';
+ pop @not_ok if $last =~ /^[0-9]+\.\.[0-9]+$/;
my $pfx = "# $log: ";
- print OLDERR map { $pfx.$_ } @not_ok;
+ print $OLDERR map { $pfx.$_ } @not_ok;
seek($fh, 0, SEEK_SET) or die "seek: $!";
# show unique skip texts and the number of times
my @sk = (<$fh> =~ m/^ok [0-9]+ (# skip [^\n]+)/mgs);
if (@sk) {
my %nr;
- $nr{$_}++ for @sk;
- for (@sk) {
- my $n = delete $nr{$_} or next;
- print OLDERR "$pfx$_ ($n)\n";
- }
+ my @err = grep { !$nr{$_}++ } @sk;
+ print $OLDERR "$pfx$_ ($nr{$_})\n" for @err;
$skip = ' # total skipped: '.scalar(@sk);
}
} else {
- print OLDERR "could not open: $log: $!\n";
+ print $OLDERR "could not open: $log: $!\n";
}
- print OLDOUT "$status $worker_test$skip\n";
+ print $OLDOUT "$status $worker_test$skip\n";
}
}
sub run_test ($) {
my ($test) = @_;
+ syswrite($run_log, "$$ $test\n");
my $log_fh;
if ($log_suffix ne '') {
my $log = $test;
my $eof; # we stop respawning if true
my $start_worker = sub {
- my ($i, $j, $rd, $todo) = @_;
- defined(my $pid = fork) or DIE "fork: $!";
+ my ($j, $rd, $wr, $todo) = @_;
+ my $pid = fork // DIE "fork: $!";
if ($pid == 0) {
+ close $wr if $wr;
+ $SIG{USR1} = undef; # undo parent $SIG{USR1}
$worker = $$;
while (1) {
my $r = sysread($rd, my $buf, UINT_SIZE);
$tb->reset;
}
kill 'USR1', $producer if !$eof; # sets $eof in $producer
- DIE join('', map { "E: $_\n" } @err) if @err;
+ if (@err) { # write to run_log for $sigchld handler
+ syswrite($run_log, "$$ @err\n");
+ DIE join('', map { "E: $_\n" } @err);
+ }
exit(0);
} else {
$pids{$pid} = $j;
for (my $i = $repeat; $i != 0; $i--) {
my @todo = $shuffle ? List::Util::shuffle(@tests) : @tests;
- # single-producer, multi-consumer queue relying on POSIX semantics
+ # single-producer, multi-consumer queue relying on POSIX pipe semantics
+ # POSIX.1-2008 stipulates a regular file should work, but Linux <3.14
+ # had broken read(2) semantics according to the read(2) manpage
pipe(my ($rd, $wr)) or DIE "pipe: $!";
# fill the queue before forking so children can start earlier
- my $n = (_POSIX_PIPE_BUF / UINT_SIZE);
+ my $n = (POSIX::PIPE_BUF / UINT_SIZE);
if ($n >= $#todo) {
print $wr join('', map { pack('I', $_) } (0..$#todo)) or DIE;
- close $wr or die;
- $wr = undef;
+ undef $wr;
} else { # write what we can...
$wr->autoflush(1);
print $wr join('', map { pack('I', $_) } (0..$n)) or DIE;
push @err, "reaped unknown $pid ($?)";
next;
}
- push @err, "job[$j] ($?)" if $?;
+ if ($?) {
+ seek($run_log, 0, SEEK_SET);
+ chomp(my @t = grep(/^$pid /, <$run_log>));
+ $t[0] //= "$pid unknown";
+ push @err, "job[$j] ($?) PID=$t[-1]";
+ }
# skip_all can exit(0), respawn if needed:
if (!$eof) {
- print OLDERR "# respawning job[$j]\n";
- $start_worker->($i, $j, $rd, \@todo);
+ print $OLDERR "# respawning job[$j]\n";
+ $start_worker->($j, $rd, $wr, \@todo);
}
}
};
# start the workers to consume the queue
for (my $j = 0; $j < $jobs; $j++) {
- $start_worker->($i, $j, $rd, \@todo);
+ $start_worker->($j, $rd, $wr, \@todo);
}
-
if ($wr) {
local $SIG{CHLD} = $sigchld;
# too many tests to fit in the pipe before starting workers,
# send the rest now the workers are running
print $wr join('', map { pack('I', $_) } ($n..$#todo)) or DIE;
- close $wr or die;
+ undef $wr;
}
$sigchld->(0) while scalar(keys(%pids));
DIE join('', map { "E: $_\n" } @err) if @err;
}
-print OLDOUT "1..".($repeat * scalar(@tests))."\n" if $repeat >= 0;
+print $OLDOUT "1..".($repeat * scalar(@tests))."\n" if $repeat >= 0;
+if ($lei_env && $$ == $owner_pid) {
+ my $opt = { 1 => $OLDOUT, 2 => $OLDERR };
+ my $cur_daemon_pid;
+ run_script([qw(lei daemon-pid)], $lei_env, { 1 => \$cur_daemon_pid });
+ run_script([qw(lei daemon-kill)], $lei_env, $opt);
+ DIE "lei daemon restarted\n" if $cur_daemon_pid != $lei_daemon_pid;
+}