[Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver

Paolo Bonzini posted 2 patches 7 years, 2 months ago
There is a newer version of this series
[Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver
Posted by Paolo Bonzini 7 years, 2 months ago
gtester is deprecated by upstream glib and it does not support tests
that call g_test_skip in some glib stable releases.

glib suggests instead using Automake's TAP support.  We do not support
Automake, but we can copy the code that beautifies the TAP output and
use it.  I chose to use the Perl copy rather than the shell/awk one,
in order to reuse Perl's TAP parsing package, but I'm open to suggestions
about which language to use.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 scripts/gtester-cat                          |  26 --
 scripts/tap-driver.pl                        | 386 +++++++++++++++++++++++++++
 scripts/tap-merge.pl                         | 108 ++++++++
 tests/Makefile.include                       |  54 ++--
 tests/docker/dockerfiles/centos7.docker      |   1 +
 tests/docker/dockerfiles/debian-amd64.docker |   1 +
 tests/docker/dockerfiles/debian-ports.docker |   1 +
 tests/docker/dockerfiles/debian-sid.docker   |   1 +
 tests/docker/dockerfiles/debian8.docker      |   1 +
 tests/docker/dockerfiles/debian9.docker      |   1 +
 tests/docker/dockerfiles/fedora.docker       |   1 +
 tests/docker/dockerfiles/ubuntu.docker       |   1 +
 12 files changed, 535 insertions(+), 47 deletions(-)
 delete mode 100755 scripts/gtester-cat
 create mode 100755 scripts/tap-driver.pl
 create mode 100755 scripts/tap-merge.pl

diff --git a/scripts/gtester-cat b/scripts/gtester-cat
deleted file mode 100755
index 061a952..0000000
--- a/scripts/gtester-cat
+++ /dev/null
@@ -1,26 +0,0 @@
-#!/bin/sh
-#
-# Copyright IBM, Corp. 2012
-#
-# Authors:
-#  Anthony Liguori <aliguori@us.ibm.com>
-#
-# This work is licensed under the terms of the GNU GPLv2 or later.
-# See the COPYING file in the top-level directory.
-
-cat <<EOF
-<?xml version="1.0"?>
-<gtester>
- <info>
-  <package>qemu</package>
-  <version>0.0</version>
-  <revision>rev</revision>
- </info>
-EOF
-
-sed \
-  -e '/<?xml/d' \
-  -e '/^<gtester>$/d' \
-  -e '/<info>/,/<\/info>/d' \
-  -e '$b' \
-  -e '/^<\/gtester>$/d' "$@"
diff --git a/scripts/tap-driver.pl b/scripts/tap-driver.pl
new file mode 100755
index 0000000..2e674c0
--- /dev/null
+++ b/scripts/tap-driver.pl
@@ -0,0 +1,386 @@
+#! /usr/bin/env perl
+# Copyright (C) 2011-2013 Free Software Foundation, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2, or (at your option)
+# any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# As a special exception to the GNU General Public License, if you
+# distribute this file as part of a program that contains a
+# configuration script generated by Autoconf, you may include it under
+# the same distribution terms that you use for the rest of that program.
+
+# ---------------------------------- #
+#  Imports, static data, and setup.  #
+# ---------------------------------- #
+
+use warnings FATAL => 'all';
+use strict;
+use Getopt::Long ();
+use TAP::Parser;
+use Term::ANSIColor qw(:constants);
+
+my $VERSION = '2012-02-01.19'; # UTC
+
+my $ME = "tap-driver.pl";
+
+my $USAGE = <<'END';
+Usage:
+  tap-driver [--color={always|never|auto}]
+             [--verbose]
+END
+
+my $HELP = "$ME: TAP-aware test driver for QEMU testsuite harness." .
+           "\n" . $USAGE;
+
+# It's important that NO_PLAN evaluates "false" as a boolean.
+use constant NO_PLAN => 0;
+use constant EARLY_PLAN => 1;
+use constant LATE_PLAN => 2;
+
+# ------------------- #
+#  Global variables.  #
+# ------------------- #
+
+my $testno = 0;     # Number of test results seen so far.
+my $bailed_out = 0; # Whether a "Bail out!" directive has been seen.
+
+# Whether the TAP plan has been seen or not, and if yes, which kind
+# it is ("early" is seen before any test result, "late" otherwise).
+my $plan_seen = NO_PLAN;
+
+# ----------------- #
+#  Option parsing.  #
+# ----------------- #
+
+my %cfg = (
+  "color" => 0,
+  "verbose" => 0,
+);
+
+my $diag_string = "#";
+my $color = "auto";
+
+# Perl's Getopt::Long allows options to take optional arguments after a space.
+# Prevent --color by itself from consuming other arguments
+foreach (@ARGV) {
+  if ($_ eq "--color" || $_ eq "-color") {
+    $_ = "--color=$color";
+  }
+}
+
+Getopt::Long::GetOptions
+  (
+    'help' => sub { print $HELP; exit 0; },
+    'version' => sub { print "$ME $VERSION\n"; exit 0; },
+    'color=s'  => \$color,
+    'verbose' => sub { $cfg{"verbose"} = 1; },
+  ) or exit 1;
+
+if ($color =~ /^always$/i) {
+  $cfg{'color'} = 1;
+} elsif ($color =~ /^never$/i) {
+  $cfg{'color'} = 0;
+} elsif ($color =~ /^auto$/i) {
+  $cfg{'color'} = (-t STDOUT);
+} else {
+  die "Invalid color mode: $color\n";
+}
+
+# ------------- #
+#  Prototypes.  #
+# ------------- #
+
+sub bool_opt ($$);
+sub colored ($$);
+sub decorate_result ($);
+sub extract_tap_comment ($);
+sub handle_tap_bailout ($);
+sub handle_tap_plan ($);
+sub handle_tap_result ($);
+sub is_null_string ($);
+sub main (@);
+sub report ($;$);
+sub stringify_result_obj ($);
+sub testsuite_error ($);
+
+# -------------- #
+#  Subroutines.  #
+# -------------- #
+
+sub bool_opt ($$)
+{
+  my ($opt, $val) = @_;
+  if ($val =~ /^(?:y|yes)\z/i)
+    {
+      $cfg{$opt} = 1;
+    }
+  elsif ($val =~ /^(?:n|no)\z/i)
+    {
+      $cfg{$opt} = 0;
+    }
+  else
+    {
+      die "$ME: invalid argument '$val' for option '$opt'\n";
+    }
+}
+
+# If the given string is undefined or empty, return true, otherwise
+# return false.  This function is useful to avoid pitfalls like:
+#   if ($message) { print "$message\n"; }
+# which wouldn't print anything if $message is the literal "0".
+sub is_null_string ($)
+{
+  my $str = shift;
+  return ! (defined $str and length $str);
+}
+
+sub stringify_result_obj ($)
+{
+  my $result_obj = shift;
+  if ($result_obj->is_unplanned || $result_obj->number != $testno)
+    {
+      return "ERROR";
+    }
+  elsif ($plan_seen == LATE_PLAN)
+    {
+      return "ERROR";
+    }
+  elsif (!$result_obj->directive)
+    {
+      return $result_obj->is_ok ? "PASS" : "FAIL";
+    }
+  elsif ($result_obj->has_todo)
+    {
+      return $result_obj->is_actual_ok ? "XPASS" : "XFAIL";
+    }
+  elsif ($result_obj->has_skip)
+    {
+      return $result_obj->is_ok ? "SKIP" : "FAIL";
+    }
+  die "$ME: INTERNAL ERROR"; # NOTREACHED
+}
+
+sub colored ($$)
+{
+  my ($color_string, $text) = @_;
+  return $color_string . $text . RESET;
+}
+
+sub decorate_result ($)
+{
+  my $result = shift;
+  return $result unless $cfg{"color"};
+  my %color_for_result =
+    (
+      "ERROR" => MAGENTA,
+      "PASS"  => GREEN,
+      "XPASS" => RED,
+      "FAIL"  => BOLD.RED,
+      "XFAIL" => BOLD.GREEN,
+      "SKIP"  => BLUE,
+    );
+  if (my $color = $color_for_result{$result})
+    {
+      return colored ($color, $result);
+    }
+  else
+    {
+      return $result; # Don't colorize unknown stuff.
+    }
+}
+
+sub report ($;$)
+{
+  my ($msg, $result, $explanation) = (undef, @_);
+  if ($result =~ /^(?:X?(?:PASS|FAIL)|SKIP|ERROR)/)
+    {
+      # Output on console might be colorized.
+      $msg = decorate_result($result);
+    }
+  elsif ($result eq "#")
+    {
+      $msg = "  ";
+    }
+  else
+    {
+      die "$ME: INTERNAL ERROR"; # NOTREACHED
+    }
+  $msg .= " $explanation" if defined $explanation;
+  print $msg . "\n";
+}
+
+sub testsuite_error ($)
+{
+  report "ERROR", "- $_[0]";
+}
+
+sub handle_tap_result ($)
+{
+  $testno++;
+  my $result_obj = shift;
+
+  my $test_result = stringify_result_obj $result_obj;
+  my $string = $result_obj->number;
+
+  my $description = $result_obj->description;
+  $string .= " $description"
+    unless is_null_string $description;
+
+  if ($plan_seen == LATE_PLAN)
+    {
+      $string .= " # AFTER LATE PLAN";
+    }
+  elsif ($result_obj->is_unplanned)
+    {
+      $string .= " # UNPLANNED";
+    }
+  elsif ($result_obj->number != $testno)
+    {
+      $string .= " # OUT-OF-ORDER (expecting $testno)";
+    }
+  elsif (my $directive = $result_obj->directive)
+    {
+      $string .= " # $directive";
+      my $explanation = $result_obj->explanation;
+      $string .= " $explanation"
+        unless is_null_string $explanation;
+    }
+
+  report $test_result, $string;
+}
+
+sub handle_tap_plan ($)
+{
+  my $plan = shift;
+  if ($plan_seen)
+    {
+      # Error, only one plan per stream is acceptable.
+      testsuite_error "multiple test plans";
+      return;
+    }
+  # The TAP plan can come before or after *all* the TAP results; we speak
+  # respectively of an "early" or a "late" plan.  If we see the plan line
+  # after at least one TAP result has been seen, assume we have a late
+  # plan; in this case, any further test result seen after the plan will
+  # be flagged as an error.
+  $plan_seen = ($testno >= 1 ? LATE_PLAN : EARLY_PLAN);
+  # If $testno > 0, we have an error ("too many tests run") that will be
+  # automatically dealt with later, so don't worry about it here.  If
+  # $plan_seen is true, we have an error due to a repeated plan, and that
+  # has already been dealt with above.  Otherwise, we have a valid "plan
+  # with SKIP" specification, and should report it as a particular kind
+  # of SKIP result.
+  if ($plan->directive && $testno == 0)
+    {
+      my $explanation = is_null_string ($plan->explanation) ?
+                        undef : "- " . $plan->explanation;
+      report "SKIP", $explanation;
+    }
+}
+
+sub handle_tap_bailout ($)
+{
+  my ($bailout, $msg) = ($_[0], "Bail out!");
+  $bailed_out = 1;
+  $msg .= " " . $bailout->explanation
+    unless is_null_string $bailout->explanation;
+  testsuite_error $msg;
+}
+
+sub extract_tap_comment ($)
+{
+  my $line = shift;
+  if (index ($line, $diag_string) == 0)
+    {
+      # Strip leading `$diag_string' from `$line'.
+      $line = substr ($line, length ($diag_string));
+      # And strip any leading and trailing whitespace left.
+      $line =~ s/(?:^\s*|\s*$)//g;
+      # Return what is left (if any).
+      return $line;
+    }
+  return "";
+}
+
+sub main (@)
+{
+  my $iterator = TAP::Parser::Iterator::Stream->new(\*STDIN);
+  my $parser = TAP::Parser->new ({iterator => $iterator });
+
+  while (defined (my $cur = $parser->next))
+    {
+      # Parsing of TAP input should stop after a "Bail out!" directive.
+      next if $bailed_out;
+
+      if ($cur->is_plan)
+        {
+          handle_tap_plan ($cur);
+        }
+      elsif ($cur->is_test)
+        {
+          handle_tap_result ($cur);
+        }
+      elsif ($cur->is_bailout)
+        {
+          handle_tap_bailout ($cur);
+        }
+      elsif ($cfg{"verbose"})
+        {
+          my $comment = extract_tap_comment ($cur->raw);
+          report "#", "$comment" if length $comment;
+       }
+    }
+  # A "Bail out!" directive should cause us to ignore any following TAP
+  # error, as well as a non-zero exit status from the TAP producer.
+  if (!$bailed_out)
+    {
+      if (!$plan_seen)
+        {
+          testsuite_error "missing test plan";
+        }
+      elsif ($parser->tests_planned != $parser->tests_run)
+        {
+          my ($planned, $run) = ($parser->tests_planned, $parser->tests_run);
+          my $bad_amount = $run > $planned ? "many" : "few";
+          testsuite_error (sprintf "too %s tests run (expected %d, got %d)",
+                                   $bad_amount, $planned, $run);
+        }
+    }
+}
+
+# ----------- #
+#  Main code. #
+# ----------- #
+
+main @ARGV;
+
+# Local Variables:
+# perl-indent-level: 2
+# perl-continued-statement-offset: 2
+# perl-continued-brace-offset: 0
+# perl-brace-offset: 0
+# perl-brace-imaginary-offset: 0
+# perl-label-offset: -2
+# cperl-indent-level: 2
+# cperl-brace-offset: 0
+# cperl-continued-brace-offset: 0
+# cperl-label-offset: -2
+# cperl-extra-newline-before-brace: t
+# cperl-merge-trailing-else: nil
+# cperl-continued-statement-offset: 2
+# eval: (add-hook 'write-file-hooks 'time-stamp)
+# time-stamp-start: "my $VERSION = "
+# time-stamp-format: "'%:y-%02m-%02d.%02H'"
+# time-stamp-time-zone: "UTC"
+# time-stamp-end: "; # UTC"
+# End:
diff --git a/scripts/tap-merge.pl b/scripts/tap-merge.pl
new file mode 100755
index 0000000..7e0de90
--- /dev/null
+++ b/scripts/tap-merge.pl
@@ -0,0 +1,108 @@
+#! /usr/bin/env perl
+# Copyright (C) 2018 Red Hat, Inc.
+#
+# Author: Paolo Bonzini <pbonzini@redhat.com>
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2, or (at your option)
+# any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# ---------------------------------- #
+#  Imports, static data, and setup.  #
+# ---------------------------------- #
+
+use warnings FATAL => 'all';
+use strict;
+use Getopt::Long ();
+use TAP::Parser;
+
+my $ME = "tap-merge.pl";
+my $VERSION = "2018-11-29";
+
+my $HELP = "$ME: merge multiple TAP inputs from stdin.";
+
+# ----------------- #
+#  Option parsing.  #
+# ----------------- #
+
+Getopt::Long::GetOptions
+  (
+    'help' => sub { print $HELP; exit 0; },
+    'version' => sub { print "$ME $VERSION\n"; exit 0; },
+  );
+
+# -------------- #
+#  Subroutines.  #
+# -------------- #
+
+sub main ()
+{
+  my $iterator = TAP::Parser::Iterator::Stream->new(\*STDIN);
+  my $parser = TAP::Parser->new ({iterator => $iterator });
+  my $testno = 0;     # Number of test results seen so far.
+  my $bailed_out = 0; # Whether a "Bail out!" directive has been seen.
+
+  while (defined (my $cur = $parser->next))
+    {
+      if ($cur->is_bailout)
+        {
+          $bailed_out = 1;
+          next;
+        }
+      if ($cur->is_plan)
+        {
+          $bailed_out = 0;
+          next;
+        }
+      if ($cur->is_test)
+        {
+          $bailed_out = 0 if $cur->number == 1;
+          $testno++;
+          $cur = TAP::Parser::Result::Test->new({
+                          ok => $cur->ok,
+                          test_num => $testno,
+                          directive => $cur->directive,
+                          explanation => $cur->explanation,
+                          description => $cur->description
+                  });
+        }
+      print $cur->as_string . "\n" unless $bailed_out;
+    }
+  print "1..$testno\n";
+}
+
+# ----------- #
+#  Main code. #
+# ----------- #
+
+main;
+
+# Local Variables:
+# perl-indent-level: 2
+# perl-continued-statement-offset: 2
+# perl-continued-brace-offset: 0
+# perl-brace-offset: 0
+# perl-brace-imaginary-offset: 0
+# perl-label-offset: -2
+# cperl-indent-level: 2
+# cperl-brace-offset: 0
+# cperl-continued-brace-offset: 0
+# cperl-label-offset: -2
+# cperl-extra-newline-before-brace: t
+# cperl-merge-trailing-else: nil
+# cperl-continued-statement-offset: 2
+# eval: (add-hook 'write-file-hooks 'time-stamp)
+# time-stamp-start: "my $VERSION = "
+# time-stamp-format: "'%:y-%02m-%02d.%02H'"
+# time-stamp-time-zone: "UTC"
+# time-stamp-end: "; # UTC"
+# End:
diff --git a/tests/Makefile.include b/tests/Makefile.include
index fb0b449..d27e137 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -799,41 +799,53 @@ tests/test-qga$(EXESUF): qemu-ga$(EXESUF)
 tests/test-qga$(EXESUF): tests/test-qga.o $(qtest-obj-y)
 
 SPEED = quick
-GTESTER_OPTIONS = -k $(if $(V),--verbose,-q)
-GCOV_OPTIONS = -n $(if $(V),-f,)
 
 # gtester tests, possibly with verbose output
 
+do_test_human = \
+	$(call quiet-command, \
+	  { $(foreach COMMAND, $1, \
+	    $(if $(V), echo $(notdir $(COMMAND)):;) \
+	    MALLOC_PERTURB_=$${MALLOC_PERTURB_:-$$(( $${RANDOM:-0} % 255 + 1))} \
+	      $2 $(COMMAND) -m=$(SPEED) -k $(if $(V), --tap | ./scripts/tap-driver.pl, \
+		   >/dev/null );) }, \
+	  "TEST","$@")
+
+do_test_tap = \
+	$(call quiet-command, \
+	  { $(foreach COMMAND, $1, \
+	    MALLOC_PERTURB_=$${MALLOC_PERTURB_:-$$(( $${RANDOM:-0} % 255 + 1))} \
+	      $2 $(COMMAND) -m=$(SPEED) -k --tap | sed "s/^[a-z][a-z]* [0-9]* /&$(notdir $(COMMAND)) /" ;) } \
+	      | $(SRC_PATH)/scripts/tap-merge.pl \
+	      $(if $(V), | tee "$@" | ./scripts/tap-driver.pl, >"$@"), \
+	  "TAP","$@")
+
 .PHONY: $(patsubst %, check-qtest-%, $(QTEST_TARGETS))
 $(patsubst %, check-qtest-%, $(QTEST_TARGETS)): check-qtest-%: subdir-%-softmmu $(check-qtest-y)
-	$(call quiet-command,QTEST_QEMU_BINARY=$*-softmmu/qemu-system-$* \
-		QTEST_QEMU_IMG=qemu-img$(EXESUF) \
-		MALLOC_PERTURB_=$${MALLOC_PERTURB_:-$$(( $${RANDOM:-0} % 255 + 1))} \
-		gtester $(GTESTER_OPTIONS) -m=$(SPEED) $(check-qtest-$*-y) $(check-qtest-generic-y),"GTESTER","$@")
+	$(call do_test_human,$(check-qtest-$*-y) $(check-qtest-generic-y), \
+	  QTEST_QEMU_BINARY=$*-softmmu/qemu-system-$* \
+	  QTEST_QEMU_IMG=qemu-img$(EXESUF))
 
 .PHONY: $(patsubst %, check-%, $(check-unit-y) $(check-speed-y))
 $(patsubst %, check-%, $(check-unit-y) $(check-speed-y)): check-%: %
-	$(call quiet-command, \
-		MALLOC_PERTURB_=$${MALLOC_PERTURB_:-$$(( $${RANDOM:-0} % 255 + 1))} \
-		gtester $(GTESTER_OPTIONS) -m=$(SPEED) $*,"GTESTER","$*")
+	$(call do_test_human, $*)
 
-# gtester tests with XML output
+# gtester tests with TAP output
 
-$(patsubst %, check-report-qtest-%.xml, $(QTEST_TARGETS)): check-report-qtest-%.xml: $(check-qtest-y)
-	$(call quiet-command,QTEST_QEMU_BINARY=$*-softmmu/qemu-system-$* \
-		QTEST_QEMU_IMG=qemu-img$(EXESUF) \
-	  gtester -q $(GTESTER_OPTIONS) -o $@ -m=$(SPEED) $(check-qtest-$*-y) $(check-qtest-generic-y),"GTESTER","$@")
+$(patsubst %, check-report-qtest-%.tap, $(QTEST_TARGETS)): check-report-qtest-%.tap: $(check-qtest-y)
+	$(call do_test_tap, $(check-qtest-$*-y) $(check-qtest-generic-y), \
+	  QTEST_QEMU_BINARY=$*-softmmu/qemu-system-$* \
+	  QTEST_QEMU_IMG=qemu-img$(EXESUF))
 
-check-report-unit.xml: $(check-unit-y)
-	$(call quiet-command,gtester -q $(GTESTER_OPTIONS) -o $@ -m=$(SPEED) $^,"GTESTER","$@")
+check-report-unit.tap: $(check-unit-y)
+	$(call do_test_tap,$^, \
+	  QTEST_QEMU_BINARY=$*-softmmu/qemu-system-$* \
+	  QTEST_QEMU_IMG=qemu-img$(EXESUF))
 
 # Reports and overall runs
 
-check-report.xml: $(patsubst %,check-report-qtest-%.xml, $(QTEST_TARGETS)) check-report-unit.xml
-	$(call quiet-command,$(SRC_PATH)/scripts/gtester-cat $^ > $@,"GEN","$@")
-
-check-report.html: check-report.xml
-	$(call quiet-command,gtester-report $< > $@,"GEN","$@")
+check-report.tap: $(patsubst %,check-report-qtest-%.tap, $(QTEST_TARGETS)) check-report-unit.tap
+	$(call quiet-command,$(SRC_PATH)/scripts/tap-merge.py $^ > $@,"GEN","$@")
 
 # Per guest TCG tests
 
diff --git a/tests/docker/dockerfiles/centos7.docker b/tests/docker/dockerfiles/centos7.docker
index 0a04bfb..e0f18f5 100644
--- a/tests/docker/dockerfiles/centos7.docker
+++ b/tests/docker/dockerfiles/centos7.docker
@@ -22,6 +22,7 @@ ENV PACKAGES \
     mesa-libEGL-devel \
     mesa-libgbm-devel \
     nettle-devel \
+    perl-Test-Harness \
     pixman-devel \
     SDL-devel \
     spice-glib-devel \
diff --git a/tests/docker/dockerfiles/debian-amd64.docker b/tests/docker/dockerfiles/debian-amd64.docker
index 24b113b..c66e341 100644
--- a/tests/docker/dockerfiles/debian-amd64.docker
+++ b/tests/docker/dockerfiles/debian-amd64.docker
@@ -16,6 +16,7 @@ RUN DEBIAN_FRONTEND=noninteractive eatmydata \
         liblzo2-dev \
         librdmacm-dev \
         libsnappy-dev \
+        libtest-harness-perl \
         libvte-dev
 
 # virgl
diff --git a/tests/docker/dockerfiles/debian-ports.docker b/tests/docker/dockerfiles/debian-ports.docker
index e05a9a9..514ab53 100644
--- a/tests/docker/dockerfiles/debian-ports.docker
+++ b/tests/docker/dockerfiles/debian-ports.docker
@@ -29,6 +29,7 @@ RUN DEBIAN_FRONTEND=noninteractive eatmydata \
         flex \
         gettext \
         git \
+        libtest-harness-perl \
         pkg-config \
         psmisc \
         python \
diff --git a/tests/docker/dockerfiles/debian-sid.docker b/tests/docker/dockerfiles/debian-sid.docker
index 9a3d168..b30cbe7 100644
--- a/tests/docker/dockerfiles/debian-sid.docker
+++ b/tests/docker/dockerfiles/debian-sid.docker
@@ -26,6 +26,7 @@ RUN DEBIAN_FRONTEND=noninteractive eatmydata \
         ca-certificates \
         flex \
         git \
+        libtest-harness-perl \
         pkg-config \
         psmisc \
         python \
diff --git a/tests/docker/dockerfiles/debian8.docker b/tests/docker/dockerfiles/debian8.docker
index 5294563..cdc3f11 100644
--- a/tests/docker/dockerfiles/debian8.docker
+++ b/tests/docker/dockerfiles/debian8.docker
@@ -29,6 +29,7 @@ RUN DEBIAN_FRONTEND=noninteractive eatmydata \
         gettext \
         git \
         gnupg \
+        libtest-harness-perl \
         pkg-config \
         python-minimal
 
diff --git a/tests/docker/dockerfiles/debian9.docker b/tests/docker/dockerfiles/debian9.docker
index 154ae2a..9561d4f 100644
--- a/tests/docker/dockerfiles/debian9.docker
+++ b/tests/docker/dockerfiles/debian9.docker
@@ -24,6 +24,7 @@ RUN DEBIAN_FRONTEND=noninteractive eatmydata \
         flex \
         gettext \
         git \
+        libtest-harness-perl \
         pkg-config \
         psmisc \
         python \
diff --git a/tests/docker/dockerfiles/fedora.docker b/tests/docker/dockerfiles/fedora.docker
index 0c4eb9e..1d0e3dc 100644
--- a/tests/docker/dockerfiles/fedora.docker
+++ b/tests/docker/dockerfiles/fedora.docker
@@ -70,6 +70,7 @@ ENV PACKAGES \
     nss-devel \
     numactl-devel \
     perl \
+    perl-Test-Harness \
     pixman-devel \
     python3 \
     PyYAML \
diff --git a/tests/docker/dockerfiles/ubuntu.docker b/tests/docker/dockerfiles/ubuntu.docker
index 36e2b17..229add5 100644
--- a/tests/docker/dockerfiles/ubuntu.docker
+++ b/tests/docker/dockerfiles/ubuntu.docker
@@ -45,6 +45,7 @@ ENV PACKAGES flex bison \
     libspice-protocol-dev \
     libspice-server-dev \
     libssh2-1-dev \
+    libtest-harness-perl \
     libusb-1.0-0-dev \
     libusbredirhost-dev \
     libvdeplug-dev \
-- 
1.8.3.1


Re: [Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver
Posted by Daniel P. Berrangé 7 years, 2 months ago
On Thu, Nov 29, 2018 at 06:45:31PM +0100, Paolo Bonzini wrote:
> gtester is deprecated by upstream glib and it does not support tests
> that call g_test_skip in some glib stable releases.
> 
> glib suggests instead using Automake's TAP support.  We do not support
> Automake, but we can copy the code that beautifies the TAP output and
> use it.  I chose to use the Perl copy rather than the shell/awk one,
> in order to reuse Perl's TAP parsing package, but I'm open to suggestions
> about which language to use.

Do we even need the beutifier ?

IIUC, this output is only seen when you with 'make check' passing V=1,
so most people won't see it.

When V=1 is set, it just changes:

$ ./tests/test-io-channel-socket -k --tap
# random seed: R02Sb9d02d2a9658ae1be211540f5ceebc01
1..9
# Start of io tests
# Start of channel tests
# Start of socket tests
ok 1 /io/channel/socket/ipv4-sync
ok 2 /io/channel/socket/ipv4-async
ok 3 /io/channel/socket/ipv4-fd
ok 4 /io/channel/socket/ipv6-sync
ok 5 /io/channel/socket/ipv6-async
ok 6 /io/channel/socket/unix-sync
ok 7 /io/channel/socket/unix-async
**
ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
Bail out! ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
Aborted (core dumped)


To look like:

$ ./tests/test-io-channel-socket -k --tap | ./scripts/tap-driver.pl 
PASS 1 /io/channel/socket/ipv4-sync
PASS 2 /io/channel/socket/ipv4-async
PASS 3 /io/channel/socket/ipv4-fd
PASS 4 /io/channel/socket/ipv6-sync
PASS 5 /io/channel/socket/ipv6-async
PASS 6 /io/channel/socket/unix-sync
PASS 7 /io/channel/socket/unix-async
ERROR - Bail out! ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)

where 'PASS' is coloured in green and 'ERROR' is red. I'm not
convinced this change justifies having the tap-driver.pl script

Since this commit already proposes a dep on the perl-Test-Harness
package, the host will have the 'prove' command, which can be used
for invoking tests.

Prove just needs a wrapper to know how to invoke glib tests:

  $ cat glibharness.sh 
  #!/bin/sh

  exec "$@" -k --tap

Now you can run :

$ prove -v -e ./glibharness.sh  tests/test-io-channel-socket
tests/test-io-channel-socket .. 
# random seed: R02S7b47ca0de1bc27cbbcf22dcc130c2b7f
1..9
# Start of io tests
# Start of channel tests
# Start of socket tests
ok 1 /io/channel/socket/ipv4-sync
ok 2 /io/channel/socket/ipv4-async
ok 3 /io/channel/socket/ipv4-fd
ok 4 /io/channel/socket/ipv6-sync
ok 5 /io/channel/socket/ipv6-async
ok 6 /io/channel/socket/unix-sync
ok 7 /io/channel/socket/unix-async
**
ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
Bailout called.  Further testing stopped:  ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
Bail out! ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
FAILED--Further testing stopped: ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)


Where the failure line is now coloured in Red (or sucess is coloured
in green)

prove can also run tests in batches, giving progress of execution of
each test as it runs

$ prove  -e ./glibharness.sh  tests/test-io-channel-{buffer,command,file,socket,tls} 
tests/test-io-channel-buffer ... ok   
tests/test-io-channel-command .. ok   
tests/test-io-channel-file ..... ok   
tests/test-io-channel-socket ... 1/? **
ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
Bailout called.  Further testing stopped:  ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)
FAILED--Further testing stopped: ERROR:tests/test-io-channel-socket.c:424:test_io_channel_unix_fd_pass: assertion failed: (0)


Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: [Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver
Posted by Paolo Bonzini 7 years, 2 months ago
On 30/11/18 16:50, Daniel P. Berrangé wrote:
> IIUC, this output is only seen when you with 'make check' passing V=1,
> so most people won't see it.

V=1 should be enabled by default on Patchew, so it will be visible in
the CI logs.  But anyway that's minor, there are more reasons why a
custom beautifier is better than prove.

First, at the very least, failures should be included in the output even
if V=1 is not provided.  Therefore, for failing tests we'd have:

  $ scripts/tap-driver.pl < f.tap
  FAIL 2 bbb
  XPASS 3 ccc # TODO
  FAIL 6 fff # SKIP cannot find frobnicator

And the output of prove in that case would be inferior for developers,
unless you use "V=1":

  $ prove -e cat f.tap
  f.tap .. Failed 2/6 subtests
	(less 2 skipped subtests: 2 okay)
	(1 TODO test unexpectedly succeeded)

  Test Summary Report
  -------------------
  f.tap (Wstat: 0 Tests: 6 Failed: 2)
    Failed tests:  2, 6
    TODO passed:   3
  Files=1, Tests=6,  0 wallclock secs ( 0.02 usr +  0.00 sys =  0.02 CPU)
  Result: FAIL

where you don't really know what tests 2 and 6 are without looking at
the TAP output.  The part before the summary provides a nice progress
report, as you point out, but it adds a bunch of noise with respect to
skipped subtests (skipped tests will be more common once patch 1 is
merged; qgraph also adds some) that I personally dislike.

You can hide it with -Q, but then you cannot invoke anymore more than
one test with a single invocation of prove, and the Makefile output
becomes much bigger:

   TEST    check-foo

vs.

   TEST    check-foo
  All tests successful.
  Files=1, Tests=3,  0 wallclock secs ( 0.01 usr +  0.00 sys =  0.01 CPU)
  Result: PASS

There are two other smaller "philosophical" differences.  First, prove
treats XPASS as passes, while tap-driver.pl treats them as failures.
Second, skipped tests really stand out in the "prove -v" output, with
white-on-blue text.

I'd rather avoid getting into the business of beautifying the output of
prove.

Paolo

Re: [Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver
Posted by Eric Blake 7 years, 2 months ago
On 11/29/18 11:45 AM, Paolo Bonzini wrote:
> gtester is deprecated by upstream glib and it does not support tests
> that call g_test_skip in some glib stable releases.
> 
> glib suggests instead using Automake's TAP support.  We do not support
> Automake, but we can copy the code that beautifies the TAP output and
> use it.  I chose to use the Perl copy rather than the shell/awk one,
> in order to reuse Perl's TAP parsing package, but I'm open to suggestions
> about which language to use.

Maybe a reference to a URL documenting the glib deprecation would be in 
order? I found https://blog.gtk.org/2018/07/11/news-from-glib-2-58/

> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---

> +++ b/scripts/tap-driver.pl
> @@ -0,0 +1,386 @@
> +#! /usr/bin/env perl
> +# Copyright (C) 2011-2013 Free Software Foundation, Inc.

This is not the latest version of automake.git/contrib/tap-driver.pl - 
which Automake version is it from?  Automake moved its perl driver out 
to pasture in 2013, stating that the awk+shell version is preferred in 
new automake projects.  I don't have a strong preference for which one 
you pick, but do worry that if automake adds future enhancements to the 
awk+shell, then the perl version won't keep up and we'll be stuck 
redoing things again in a few years.  On the other hand, TAP doesn't 
seem to be gaining new features at a very fast rate.

> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.

Among other things, the most recent version of tap-driver.pl switched 
all references to https://.

> +
> +# As a special exception to the GNU General Public License, if you
> +# distribute this file as part of a program that contains a
> +# configuration script generated by Autoconf, you may include it under
> +# the same distribution terms that you use for the rest of that program.

We do not use Autoconf, so this exception does not apply to our use of 
this file.  But since our project is GPL, I don't see including this 
file as a problem, nor do I find any problem with leaving the exception 
in place.

> +
> +my $VERSION = '2012-02-01.19'; # UTC

Okay, so this matches the version in automake.git commit 6b819187, which 
was shipped in automake 1.11b.  Six years old, but still does what we 
need, at least for today.

Otherwise, looks good to me, although I'm weak enough with TAP in 
general that this is a fairly weak:
Reviewed-by: Eric Blake <eblake@redhat.com>

(The real test is if all the CI systems are happy with it...)

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

Re: [Qemu-devel] [PATCH 2/2] test: replace gtester with a TAP driver
Posted by Paolo Bonzini 7 years, 2 months ago
On 29/11/18 22:06, Eric Blake wrote:
> On 11/29/18 11:45 AM, Paolo Bonzini wrote:
>> gtester is deprecated by upstream glib and it does not support tests
>> that call g_test_skip in some glib stable releases.
>>
>> glib suggests instead using Automake's TAP support.  We do not support
>> Automake, but we can copy the code that beautifies the TAP output and
>> use it.  I chose to use the Perl copy rather than the shell/awk one,
>> in order to reuse Perl's TAP parsing package, but I'm open to suggestions
>> about which language to use.
> 
> Maybe a reference to a URL documenting the glib deprecation would be in
> order? I found https://blog.gtk.org/2018/07/11/news-from-glib-2-58/

Good idea.

>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
> 
>> +++ b/scripts/tap-driver.pl
>> @@ -0,0 +1,386 @@
>> +#! /usr/bin/env perl
>> +# Copyright (C) 2011-2013 Free Software Foundation, Inc.
> 
> This is not the latest version of automake.git/contrib/tap-driver.pl -
> which Automake version is it from?  Automake moved its perl driver out
> to pasture in 2013, stating that the awk+shell version is preferred in
> new automake projects.  I don't have a strong preference for which one
> you pick, but do worry that if automake adds future enhancements to the
> awk+shell, then the perl version won't keep up and we'll be stuck
> redoing things again in a few years.  On the other hand, TAP doesn't
> seem to be gaining new features at a very fast rate.

It does get some new features from time to time, and that's why I
preferred an external parser to a home-grown one---especially if we
would need to duplicate the awk parser in tap-merge.pl.

The version I used is just the one in the machine where I developed it
(RHEL 7), but there have been no relevant changes after that commit, so
it's not important.

>> +# You should have received a copy of the GNU General Public License
>> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> 
> Among other things, the most recent version of tap-driver.pl switched
> all references to https://.

I can change this too.

>> +# As a special exception to the GNU General Public License, if you
>> +# distribute this file as part of a program that contains a
>> +# configuration script generated by Autoconf, you may include it under
>> +# the same distribution terms that you use for the rest of that program.
> 
> We do not use Autoconf, so this exception does not apply to our use of
> this file.  But since our project is GPL, I don't see including this
> file as a problem, nor do I find any problem with leaving the exception
> in place.

Yes, I find removing exceptions to be a bit unkind.

Paolo