perf-test: initial version of memory test infrastructure.

The idea is run some code under valgrind --leak-check=full and report
a summary, leaving the user to peruse the log file if they want.

We go to some lengths to preserve the log files from accidental
overwriting; the full corpus takes about 3 hours to run under valgrind
on my machine.

The naming of the log directories may be slightly controversial; in
the unlikely event of two runs in less than a second, the log will be
overwritten. A previous version with mktemp+timestamp was dismissed as
overkill; just mktemp alone does not sort nicely.

One new test is included, to check notmuch new for memory leaks.

performance-test/.gitignore

@@ -1,3 +1,4 @@
 tmp.*/
+log.*/
 corpus/
 notmuch.cache.*/

performance-test/M00-new

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+test_description='notmuch new'
+
+. ./perf-test-lib.sh
+
+# ensure initial 'notmuch new' is run by memory_start
+uncache_database
+
+memory_start
+
+# run 'notmuch new' a second time, to test different code paths
+memory_run "notmuch new" "notmuch new"
+
+memory_done

performance-test/Makefile.local

@@ -4,14 +4,24 @@ dir := performance-test
 
 include $(dir)/version.sh
 
+TIME_TEST_SCRIPT := ${dir}/notmuch-time-test
+MEMORY_TEST_SCRIPT := ${dir}/notmuch-memory-test
+
 CORPUS_NAME := notmuch-email-corpus-$(PERFTEST_VERSION).tar.xz
 TXZFILE := ${dir}/download/${CORPUS_NAME}
 SIGFILE := ${TXZFILE}.asc
-TEST_SCRIPT := ${dir}/notmuch-perf-test
 DEFAULT_URL :=  http://notmuchmail.org/releases/${CORPUS_NAME}
 
+perf-test: time-test memory-test
+
 time-test: setup-perf-test all
-	$(TEST_SCRIPT) $(OPTIONS)
+	@echo
+	$(TIME_TEST_SCRIPT) $(OPTIONS)
+
+memory-test: setup-perf-test all
+	@echo
+	$(MEMORY_TEST_SCRIPT) $(OPTIONS)
+
 
 .PHONY: download-corpus setup-perf-test
 
@@ -29,4 +39,4 @@ $(TXZFILE):
 download-corpus:
 	wget -O ${TXZFILE} ${DEFAULT_URL}
 
-CLEAN := $(CLEAN) $(dir)/tmp.* $(dir)/corpus $(dir)/notmuch.cache.*
+CLEAN := $(CLEAN) $(dir)/tmp.* $(dir)/log.* $(dir)/corpus $(dir)/notmuch.cache.*

performance-test/README

@@ -1,3 +1,10 @@
+Performance Tests
+-----------------
+
+This directory contains two kinds of performance tests: time tests,
+and memory tests. The former use gnu time, and the latter use
+valgrind.
+
 Pre-requisites
 --------------
 
@@ -5,9 +12,10 @@ In addition to having notmuch, you need:
 
 - gpg
 - gnu tar
-- gnu time
+- gnu time (for the time tests)
 - xz. Some speedup can be gotten by installing "pixz", but this is
   probably only worthwhile if you are debugging the tests.
+- valgrind (for the memory tests)
 
 Getting set up to run tests:
 ----------------------------
@@ -36,34 +44,47 @@ for a list of mirrors.
 Running tests
 -------------
 
-The easiest way to run performance tests is to say "make time-test", (or
+The easiest way to run performance tests is to say "make perf-test".
-simply run the notmuch-time-test script). Either command will run all
+This will run all time and memory tests.  Be aware that the memory
-available performance tests.
+tests are quite time consuming when run on the full corpus, and that
-
+depending on your interests it may be more sensible to run "make
-Alternately, you can run a specific subset of tests by simply invoking
+time-test" or "make memory-test".  You can also invoke one of the
-one of the executable scripts in this directory, (such as ./basic).
+scripts notmuch-time-test or notmuch-memory-test or run a more
-Each test script supports the following arguments
+specific subset of tests by simply invoking one of the executable
+scripts in this directory, (such as ./T00-new).  Each test script
+supports the following arguments
 
 --small / --medium / --large	Choose corpus size.
 --debug				Enable debugging. In particular don't delete
 				temporary directories.
 
+When using the make targets, you can pass arguments to all test
+scripts by defining the make variable OPTIONS.
+
 Writing tests
 -------------
 
-Have a look at "T01-dump-restore" for an example. Sourcing
+Have a look at "T01-dump-restore" for an example time test and
-"perf-test-lib.sh" is mandatory.  Utility functions include
+"M00-new" for an example memory test. In both cases sourcing
+"perf-test-lib.sh" is mandatory.
 
-- 'add_email_corpus' unpacks a set of messages and adds them to the database.
+Basics:
-- 'cache_database': makes a snapshot of the current database
+
-- 'uncache_database': forces the next 'add_email_corpus' to rebuild the
+- '(time|memory)_start' unpacks the mail corpus and calls notmuch new if it
-  database.
-- 'time_start' unpacks the mail corpus and calls notmuch new if it
    cannot find a cache of the appropriate corpus.
-- 'time_done' does the cleanup; comment it out or pass --debug to the
+- '(time|memory)_run' runs the command under time or valgrind. Currently
+  "memory_run" does not support i/o redirection in the command.
+- '(time|memory)_done' does the cleanup; comment it out or pass --debug to the
   script to leave the temporary files around.
 
+Utility functions include
+
+- 'add_email_corpus' unpacks a set of messages and tags
+- 'cache_database': makes a snapshot of the current database
+- 'uncache_database': forces the next '(time|memory)_start' to rebuild the
+  database.
+
 Scripts are run in the order specified in notmuch-perf-test. In the
 future this order might be chosen automatically so please follow the
-convention of starting the name with 'T' followed by two digits to
+convention of starting the name with 'T' or 'M' followed by two digits
-specify the order.
+to specify the order.

performance-test/perf-test-lib.sh

@@ -89,24 +89,59 @@ add_email_corpus ()
 
     cp -lr $TAG_CORPUS $TMP_DIRECTORY/corpus.tags
     cp -lr $MAIL_CORPUS $MAIL_DIR
-
 }
 
-time_start () {
+notmuch_new_with_cache ()
-
+{
-    add_email_corpus
-
-    print_header
-
     if [ -d $DB_CACHE_DIR ]; then
 	cp -r $DB_CACHE_DIR ${MAIL_DIR}/.notmuch
     else
-	time_run 'Initial notmuch new' "notmuch new"
+	"$1" 'Initial notmuch new' "notmuch new"
 	cache_database
     fi
 }
 
-cache_database () {
+time_start ()
+{
+    add_email_corpus
+
+    print_header
+
+    notmuch_new_with_cache time_run
+}
+
+memory_start ()
+{
+    add_email_corpus
+
+    local timestamp=$(date +%Y%m%dT%H%M%S)
+    log_dir="${TEST_DIRECTORY}/log.$(basename $0)-$corpus_size-${timestamp}"
+    mkdir -p ${log_dir}
+
+    notmuch_new_with_cache memory_run
+}
+
+memory_run ()
+{
+    test_count=$(($test_count+1))
+
+    log_file=$log_dir/$test_count.log
+
+    printf "[ %d ]\t%s\n" $test_count "$1"
+
+    valgrind --leak-check=full --log-file="$log_file" $2
+
+    awk '/LEAK SUMMARY/,/suppressed/ { sub(/^==[0-9]*==/," "); print }' "$log_file"
+    echo
+}
+
+memory_done ()
+{
+    time_done
+}
+
+cache_database ()
+{
     if [ -d $MAIL_DIR/.notmuch ]; then
 	cp -r $MAIL_DIR/.notmuch $DB_CACHE_DIR
     else
@@ -114,17 +149,20 @@ cache_database () {
     fi
 }
 
-uncache_database () {
+uncache_database ()
+{
     rm -rf $DB_CACHE_DIR
 }
 
-print_header () {
+print_header ()
-    printf "[v%4s %6s]          Wall(s)\tUsr(s)\tSys(s)\tRes(K)\tIn/Out(512B)\n" \
+{
-	   ${PERFTEST_VERSION} ${corpus_size}
+    printf "\t\t\tWall(s)\tUsr(s)\tSys(s)\tRes(K)\tIn/Out(512B)\n"
 }
 
-time_run () {
+time_run ()
+{
     printf "  %-22s" "$1"
+    test_count=$(($test_count+1))
     if test "$verbose" != "t"; then exec 4>test.output 3>&4; fi
     if ! eval >&3 "/usr/bin/time -f '%e\t%U\t%S\t%M\t%I/%O' $2" ; then
 	test_failure=$(($test_failure + 1))
@@ -133,7 +171,8 @@ time_run () {
     return 0
 }
 
-time_done () {
+time_done ()
+{
     if [ "$test_failure" = "0" ]; then
 	rm -rf "$remove_tmp"
 	exit 0
@@ -144,6 +183,8 @@ time_done () {
 
 cd -P "$test" || error "Cannot setup test environment"
 test_failure=0
+test_count=0
 
-echo
+printf "\n%-55s [%s %s]\n"  \
-echo $(basename "$0"): "Testing ${test_description:-notmuch performance}"
+    "$(basename "$0"): Testing ${test_description:-notmuch performance}" \
+    "${PERFTEST_VERSION}"  "${corpus_size}"