From: Austin Clements <amdragon@MIT.EDU>
To: notmuch@notmuchmail.org
Subject: [PATCH 01/10 v2] test: Test atomicity of notmuch new.
Date: Wed,  4 May 2011 16:32:03 -0400
Message-Id: <1304541123-2237-1-git-send-email-amdragon@mit.edu>
In-Reply-To: <1298015940-31986-2-git-send-email-amdragon@mit.edu>
References: <1298015940-31986-2-git-send-email-amdragon@mit.edu>
Cc: amdragon@mit.edu
Precedence: list

This tests notmuch new's ability to recover from arbitrary stopping
failures.  It interrupts notmuch new after every database commit and,
on every resulting database snapshot, re-runs notmuch new to
completion and checks that the final database state is invariant.
---
This addresses a timing bug in the atomicity test found by jrollins
and also adds additional comments.

 test/atomicity     |   96 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 test/atomicity.gdb |   50 +++++++++++++++++++++++++++
 test/basic         |    4 ++-
 test/notmuch-test  |    1 +
 4 files changed, 150 insertions(+), 1 deletions(-)
 create mode 100755 test/atomicity
 create mode 100644 test/atomicity.gdb

diff --git a/test/atomicity b/test/atomicity
new file mode 100755
index 0000000..b7ae505
--- /dev/null
+++ b/test/atomicity
@@ -0,0 +1,96 @@
+#!/bin/bash
+test_description='atomicity'
+. ./test-lib.sh
+
+# This script tests the effects of killing and restarting "notmuch
+# new" at arbitrary points.  If notmuch new is properly atomic, the
+# final database contents should be the same regardless of when (or
+# if) it is killed and restarted.
+
+# Create a maildir structure to also stress flag synchronization
+mkdir $MAIL_DIR/cur
+mkdir $MAIL_DIR/new
+mkdir $MAIL_DIR/tmp
+mkdir $MAIL_DIR/.remove-dir
+
+# Prepare the initial database
+generate_message [subject]='Duplicate' [filename]='duplicate:2,' [dir]=cur
+generate_message [subject]='Remove' [filename]='remove:2,' [dir]=cur
+generate_message [subject]='"Remove duplicate"' [filename]='remove-duplicate:2,' [dir]=cur
+cp $MAIL_DIR/cur/remove-duplicate:2, $MAIL_DIR/cur/remove-duplicate-copy:2,
+generate_message [subject]='Rename' [filename]='rename:2,' [dir]=cur
+generate_message [subject]='"Rename duplicate"' [filename]='rename-duplicate:2,' [dir]=cur
+generate_message [subject]='"Move 1"' [filename]='move1:2,' [dir]=cur
+generate_message [subject]='"Move 2"' [filename]='move2:2,' [dir]=new
+generate_message [subject]='Flag' [filename]='flag:2,' [dir]=cur
+generate_message [subject]='"Flag duplicate"' [filename]='flag-duplicate:2,' [dir]=cur
+cp $MAIL_DIR/cur/flag-duplicate:2, $MAIL_DIR/cur/flag-duplicate-copy:2,F
+generate_message [subject]='"Remove directory"' [filename]='remove-directory:2,' [dir]=.remove-dir
+generate_message [subject]='"Remove directory duplicate"' [filename]='remove-directory-duplicate:2,' [dir]=.remove-dir
+cp $MAIL_DIR/.remove-dir/remove-directory-duplicate:2, $MAIL_DIR/cur/
+notmuch new > /dev/null
+
+# Make all maildir changes, but *don't* update the database
+generate_message [subject]='Added' [filename]='added:2,' [dir]=cur
+cp $MAIL_DIR/cur/duplicate:2, $MAIL_DIR/cur/duplicate-copy:2,
+generate_message [subject]='"Add duplicate"' [filename]='add-duplicate:2,' [dir]=cur
+generate_message [subject]='"Add duplicate copy"' [filename]='add-duplicate-copy:2,' [dir]=cur
+rm $MAIL_DIR/cur/remove:2,
+rm $MAIL_DIR/cur/remove-duplicate-copy:2,
+mv $MAIL_DIR/cur/rename:2, $MAIL_DIR/cur/renamed:2,
+mv $MAIL_DIR/cur/rename-duplicate:2, $MAIL_DIR/cur/renamed-duplicate:2,
+mv $MAIL_DIR/cur/move1:2, $MAIL_DIR/new/move1:2,
+mv $MAIL_DIR/new/move2:2, $MAIL_DIR/cur/move2:2,
+mv $MAIL_DIR/cur/flag:2, $MAIL_DIR/cur/flag:2,F
+rm $MAIL_DIR/cur/flag-duplicate-copy:2,F
+rm $MAIL_DIR/.remove-dir/remove-directory:2,
+rm $MAIL_DIR/.remove-dir/remove-directory-duplicate:2,
+rmdir $MAIL_DIR/.remove-dir
+increment_mtime $MAIL_DIR/cur
+increment_mtime $MAIL_DIR/new
+increment_mtime $MAIL_DIR
+
+# Prepare a snapshot of the updated maildir.  The gdb script will
+# update the database in this snapshot as it goes.
+cp -ra $MAIL_DIR $MAIL_DIR.snap
+cp ${NOTMUCH_CONFIG} ${NOTMUCH_CONFIG}.snap
+NOTMUCH_CONFIG=${NOTMUCH_CONFIG}.snap notmuch config set database.path $MAIL_DIR.snap
+
+
+test_begin_subtest '"notmuch new" is idempotent under arbitrary aborts'
+
+# Execute notmuch new and, at every call to rename, snapshot the
+# database, run notmuch new again on the snapshot, and capture the
+# results of search.
+export MAIL_DIR
+gdb -batch -x ../atomicity.gdb notmuch > /dev/null 2>&1
+
+# Get the final, golden output
+notmuch search '*' > expected
+
+# Check output against golden output
+outcount=$(cat outcount)
+echo -n > searchall
+echo -n > expectall
+for ((i = 0; i < $outcount; i++)); do
+    if ! cmp -s search.$i expected; then
+        # Find the range of interruptions that match this output
+        for ((end = $i + 1 ; end < $outcount; end++)); do
+            if ! cmp -s search.$i search.$end; then
+                break
+            fi
+        done
+        echo "When interrupted after $test/backtrace.$(expr $i - 1) (abort points $i-$(expr $end - 1))" >> searchall
+        cat search.$i >> searchall
+        cat expected >> expectall
+        echo >> searchall
+        echo >> expectall
+
+        i=$(expr $end - 1)
+    fi
+done
+test_expect_equal_failure "$(cat searchall)" "$(cat expectall)"
+
+test_expect_success "detected $outcount>10 abort points" "test $outcount -gt 10"
+
+test_done
diff --git a/test/atomicity.gdb b/test/atomicity.gdb
new file mode 100644
index 0000000..fd67525
--- /dev/null
+++ b/test/atomicity.gdb
@@ -0,0 +1,50 @@
+# This gdb script runs notmuch new and simulates killing and
+# restarting notmuch new after every Xapian commit.  To simulate this
+# more efficiently, this script runs notmuch new and, immediately
+# after every Xapian commit, it *pauses* the running notmuch new,
+# copies the entire database and maildir to a snapshot directory, and
+# executes a full notmuch new on that snapshot, comparing the final
+# results with the expected output.  It can then resume the paused
+# notmuch new, which is still running on the original maildir, and
+# repeat this process.
+
+set args new
+
+# Make Xapian commit after every operation instead of batching
+set environment XAPIAN_FLUSH_THRESHOLD = 1
+
+# gdb can't keep track of a simple integer.  This is me weeping.
+shell echo 0 > outcount
+
+shell touch inodes
+
+break rename
+commands
+# As an optimization, only consider snapshots after a Xapian commit.
+# Xapian overwrites record.base? as the last step in the commit.
+shell echo > gdbcmd
+shell stat -c %i $MAIL_DIR/.notmuch/xapian/record.base* > inodes.new
+shell if cmp inodes inodes.new; then echo cont > gdbcmd; fi
+shell mv inodes.new inodes
+source gdbcmd
+
+# Save a backtrace in case the test does fail
+set logging file backtrace
+set logging on
+backtrace
+set logging off
+shell mv backtrace backtrace.`cat outcount`
+
+# Snapshot the database
+shell rm -r $MAIL_DIR.snap/.notmuch
+shell cp -r $MAIL_DIR/.notmuch $MAIL_DIR.snap/.notmuch
+# Restore the mtime of $MAIL_DIR.snap, which we just changed
+shell touch -r $MAIL_DIR $MAIL_DIR.snap
+# Run notmuch new to completion on the snapshot
+shell NOTMUCH_CONFIG=${NOTMUCH_CONFIG}.snap XAPIAN_FLUSH_THRESHOLD=1000 notmuch new > /dev/null
+shell NOTMUCH_CONFIG=${NOTMUCH_CONFIG}.snap notmuch search '*' > search.`cat outcount` 2>&1
+shell echo $(expr $(cat outcount) + 1) > outcount
+cont
+end
+
+run
diff --git a/test/basic b/test/basic
index 3b43ad9..f9d5fc3 100755
--- a/test/basic
+++ b/test/basic
@@ -57,7 +57,9 @@ available=$(ls -1 ../ | \
     sed -r -e "/^(aggregate-results.sh|Makefile|Makefile.local|notmuch-test)/d" \
 	   -e "/^(README|test-lib.sh|test-results|tmp.*|valgrind|corpus*)/d" \
 	   -e "/^(emacs.expected-output|smtp-dummy|smtp-dummy.c|test-verbose)/d" \
-	   -e "/^(test.expected-output|.*~)/d" | sort)
+	   -e "/^(test.expected-output|.*~)/d" \
+	   -e "/^(atomicity.gdb)/d" \
+	   | sort)
 test_expect_equal "$tests_in_suite" "$available"
 
 EXPECTED=../test.expected-output
diff --git a/test/notmuch-test b/test/notmuch-test
index 8dd9c1e..5b22fa8 100755
--- a/test/notmuch-test
+++ b/test/notmuch-test
@@ -38,6 +38,7 @@ TESTS="
   emacs
   emacs-large-search-buffer
   maildir-sync
+  atomicity
 "
 
 # Clean up any results from a previous run
-- 
1.7.4.1