{"thread":{"id":"29643","subject":"[PATCH v2 0/3] Adding a performance framework","startedAt":"2012-02-16T21:41:12Z","lastAt":"2012-02-18T10:24:40Z","messageCount":19,"participants":["Thomas Rast","Junio C Hamano","Jeff King","Jonathan Nieder"],"isPatch":true,"patchVersion":2,"patchTotal":3},"messages":[{"id":"184840","messageId":"cover.1329428159.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":null,"subject":"[PATCH v2 0/3] Adding a performance framework","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-16T21:41:12Z","receivedAt":"2012-02-16T21:41:12Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"This is a reroll of the original RFC at\n\n http://thread.gmane.org/gmane.comp.version-control.git/187127\n\nBasically, a perf framework that uses little perf scripts written in\nthe style we already use in the test suite.\n\nAs Junio pointed out, the line between GIT_BUILD_DIR and\nGIT_TEST_INSTALLED was not very clear cut or well adhered to, so I\nthrew out the overrides for the former. This also made\nGIT-TEST-OPTIONS moot.\n\nThere are no other changes, though I did a rebase on current next.\n\n\nThomas Rast (3):\n Move the user-facing test library to test-lib-functions.sh\n Introduce a performance testing framework\n Add a performance test for git-grep\n\n Makefile | 22 +-\n t/Makefile | 43 ++-\n t/perf/.gitignore | 2 +\n t/perf/Makefile | 15 +\n t/perf/README | 146 +++++++\n t/perf/aggregate.perl | 166 ++++++++\n t/perf/min_time.perl | 21 +\n t/perf/p0000-perf-lib-sanity.sh | 41 ++\n t/perf/p0001-rev-list.sh | 17 +\n t/perf/p7810-grep.sh | 23 ++\n t/perf/perf-lib.sh | 198 +++++++++\n t/perf/run | 82 ++++\n t/test-lib-functions.sh | 835 +++++++++++++++++++++++++++++++++++++++\n t/test-lib.sh | 574 ++--------------------------\n 14 files changed, 1633 insertions(+), 552 deletions(-)\n create mode 100644 t/perf/.gitignore\n create mode 100644 t/perf/Makefile\n create mode 100644 t/perf/README\n create mode 100755 t/perf/aggregate.perl\n create mode 100755 t/perf/min_time.perl\n create mode 100755 t/perf/p0000-perf-lib-sanity.sh\n create mode 100755 t/perf/p0001-rev-list.sh\n create mode 100755 t/perf/p7810-grep.sh\n create mode 100644 t/perf/perf-lib.sh\n create mode 100755 t/perf/run\n create mode 100644 t/test-lib-functions.sh\n\n-- \n1.7.9.1.334.gd1409\n"},{"id":"184842","messageId":"d5127b0051d354fc0c02666b972b853d1736d09c.1329428159.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329428159.git.trast@student.ethz.ch","subject":"[PATCH v2 1/3] Move the user-facing test library to test-lib-functions.sh","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-16T21:41:13Z","receivedAt":"2012-02-16T21:41:13Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"This just moves all the user-facing functions to a separate file and\nsources that instead.\n\nSigned-off-by: Thomas Rast \n---\n t/test-lib-functions.sh | 835 +++++++++++++++++++++++++++++++++++++++++++++++\n t/test-lib.sh | 552 +-------------------------------\n 2 files changed, 840 insertions(+), 547 deletions(-)\n create mode 100644 t/test-lib-functions.sh\n\ndiff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh\nnew file mode 100644\nindex 0000000..580f767\n--- /dev/null\n+++ b/t/test-lib-functions.sh\n@@ -0,0 +1,835 @@\n+# The semantics of the editor variables are that of invoking\n+# sh -c \"$EDITOR \\\"$@\\\"\" files ...\n+#\n+# If our trash directory contains shell metacharacters, they will be\n+# interpreted if we just set $EDITOR directly, so do a little dance with\n+# environment variables to work around this.\n+#\n+# In particular, quoting isn't enough, as the path may contain the same quote\n+# that we're using.\n+test_set_editor () {\n+\tFAKE_EDITOR=\"$1\"\n+\texport FAKE_EDITOR\n+\tEDITOR='\"$FAKE_EDITOR\"'\n+\texport EDITOR\n+}\n+\n+test_decode_color () {\n+\tawk '\n+\t\tfunction name(n) {\n+\t\t\tif (n == 0) return \"RESET\";\n+\t\t\tif (n == 1) return \"BOLD\";\n+\t\t\tif (n == 30) return \"BLACK\";\n+\t\t\tif (n == 31) return \"RED\";\n+\t\t\tif (n == 32) return \"GREEN\";\n+\t\t\tif (n == 33) return \"YELLOW\";\n+\t\t\tif (n == 34) return \"BLUE\";\n+\t\t\tif (n == 35) return \"MAGENTA\";\n+\t\t\tif (n == 36) return \"CYAN\";\n+\t\t\tif (n == 37) return \"WHITE\";\n+\t\t\tif (n == 40) return \"BLACK\";\n+\t\t\tif (n == 41) return \"BRED\";\n+\t\t\tif (n == 42) return \"BGREEN\";\n+\t\t\tif (n == 43) return \"BYELLOW\";\n+\t\t\tif (n == 44) return \"BBLUE\";\n+\t\t\tif (n == 45) return \"BMAGENTA\";\n+\t\t\tif (n == 46) return \"BCYAN\";\n+\t\t\tif (n == 47) return \"BWHITE\";\n+\t\t}\n+\t\t{\n+\t\t\twhile (match($0, /\\033\\[[0-9;]*m/) != 0) {\n+\t\t\t\tprintf \"%s<\", substr($0, 1, RSTART-1);\n+\t\t\t\tcodes = substr($0, RSTART+2, RLENGTH-3);\n+\t\t\t\tif (length(codes) == 0)\n+\t\t\t\t\tprintf \"%s\", name(0)\n+\t\t\t\telse {\n+\t\t\t\t\tn = split(codes, ary, \";\");\n+\t\t\t\t\tsep = \"\";\n+\t\t\t\t\tfor (i = 1; i <= n; i++) {\n+\t\t\t\t\t\tprintf \"%s%s\", sep, name(ary[i]);\n+\t\t\t\t\t\tsep = \";\"\n+\t\t\t\t\t}\n+\t\t\t\t}\n+\t\t\t\tprintf \">\";\n+\t\t\t\t$0 = substr($0, RSTART + RLENGTH, length($0) - RSTART - RLENGTH + 1);\n+\t\t\t}\n+\t\t\tprint\n+\t\t}\n+\t'\n+}\n+\n+nul_to_q () {\n+\tperl -pe 'y/\\000/Q/'\n+}\n+\n+q_to_nul () {\n+\tperl -pe 'y/Q/\\000/'\n+}\n+\n+q_to_cr () {\n+\ttr Q '\\015'\n+}\n+\n+q_to_tab () {\n+\ttr Q '\\011'\n+}\n+\n+append_cr () {\n+\tsed -e 's/$/Q/' | tr Q '\\015'\n+}\n+\n+remove_cr () {\n+\ttr '\\015' Q | sed -e 's/Q$//'\n+}\n+\n+# In some bourne shell implementations, the \"unset\" builtin returns\n+# nonzero status when a variable to be unset was not set in the first\n+# place.\n+#\n+# Use sane_unset when that should not be considered an error.\n+\n+sane_unset () {\n+\tunset \"$@\"\n+\treturn 0\n+}\n+\n+test_tick () {\n+\tif test -z \"${test_tick+set}\"\n+\tthen\n+\t\ttest_tick=1112911993\n+\telse\n+\t\ttest_tick=$(($test_tick + 60))\n+\tfi\n+\tGIT_COMMITTER_DATE=\"$test_tick -0700\"\n+\tGIT_AUTHOR_DATE=\"$test_tick -0700\"\n+\texport GIT_COMMITTER_DATE GIT_AUTHOR_DATE\n+}\n+\n+# Call test_commit with the arguments \" [ []]\"\n+#\n+# This will commit a file with the given contents and the given commit\n+# message. It will also add a tag with as name.\n+#\n+# Both and default to .\n+\n+test_commit () {\n+\tfile=${2:-\"$1.t\"}\n+\techo \"${3-$1}\" > \"$file\" &&\n+\tgit add \"$file\" &&\n+\ttest_tick &&\n+\tgit commit -m \"$1\" &&\n+\tgit tag \"$1\"\n+}\n+\n+# Call test_merge with the arguments \" \", where \n+# can be a tag pointing to the commit-to-merge.\n+\n+test_merge () {\n+\ttest_tick &&\n+\tgit merge -m \"$1\" \"$2\" &&\n+\tgit tag \"$1\"\n+}\n+\n+# This function helps systems where core.filemode=false is set.\n+# Use it instead of plain 'chmod +x' to set or unset the executable bit\n+# of a file in the working directory and add it to the index.\n+\n+test_chmod () {\n+\tchmod \"$@\" &&\n+\tgit update-index --add \"--chmod=$@\"\n+}\n+\n+# Unset a configuration variable, but don't fail if it doesn't exist.\n+test_unconfig () {\n+\tgit config --unset-all \"$@\"\n+\tconfig_status=$?\n+\tcase \"$config_status\" in\n+\t5) # ok, nothing to unset\n+\t\tconfig_status=0\n+\t\t;;\n+\tesac\n+\treturn $config_status\n+}\n+\n+# Set git config, automatically unsetting it after the test is over.\n+test_config () {\n+\ttest_when_finished \"test_unconfig '$1'\" &&\n+\tgit config \"$@\"\n+}\n+\n+test_config_global () {\n+\ttest_when_finished \"test_unconfig --global '$1'\" &&\n+\tgit config --global \"$@\"\n+}\n+\n+# Use test_set_prereq to tell that a particular prerequisite is available.\n+# The prerequisite can later be checked for in two ways:\n+#\n+# - Explicitly using test_have_prereq.\n+#\n+# - Implicitly by specifying the prerequisite tag in the calls to\n+# test_expect_{success,failure,code}.\n+#\n+# The single parameter is the prerequisite tag (a simple word, in all\n+# capital letters by convention).\n+\n+test_set_prereq () {\n+\tsatisfied=\"$satisfied$1 \"\n+}\n+satisfied=\" \"\n+\n+test_have_prereq () {\n+\t# prerequisites can be concatenated with ','\n+\tsave_IFS=$IFS\n+\tIFS=,\n+\tset -- $*\n+\tIFS=$save_IFS\n+\n+\ttotal_prereq=0\n+\tok_prereq=0\n+\tmissing_prereq=\n+\n+\tfor prerequisite\n+\tdo\n+\t\ttotal_prereq=$(($total_prereq + 1))\n+\t\tcase $satisfied in\n+\t\t*\" $prerequisite \"*)\n+\t\t\tok_prereq=$(($ok_prereq + 1))\n+\t\t\t;;\n+\t\t*)\n+\t\t\t# Keep a list of missing prerequisites\n+\t\t\tif test -z \"$missing_prereq\"\n+\t\t\tthen\n+\t\t\t\tmissing_prereq=$prerequisite\n+\t\t\telse\n+\t\t\t\tmissing_prereq=\"$prerequisite,$missing_prereq\"\n+\t\t\tfi\n+\t\tesac\n+\tdone\n+\n+\ttest $total_prereq = $ok_prereq\n+}\n+\n+test_declared_prereq () {\n+\tcase \",$test_prereq,\" in\n+\t*,$1,*)\n+\t\treturn 0\n+\t\t;;\n+\tesac\n+\treturn 1\n+}\n+\n+test_expect_failure () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-failure\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"checking known breakage: $2\"\n+\t\tif test_run_ \"$2\" expecting_failure\n+\t\tthen\n+\t\t\ttest_known_broken_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_known_broken_failure_ \"$1\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+test_expect_success () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"expecting success: $2\"\n+\t\tif test_run_ \"$2\"\n+\t\tthen\n+\t\t\ttest_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_failure_ \"$@\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+# test_external runs external test scripts that provide continuous\n+# test output about their progress, and succeeds/fails on\n+# zero/non-zero exit code. It outputs the test output on stdout even\n+# in non-verbose mode, and announces the external script with \"# run\n+# : ...\" before running it. When providing relative paths, keep in\n+# mind that all scripts run in \"trash directory\".\n+# Usage: test_external description command arguments...\n+# Example: test_external 'Perl API' perl ../path/to/test.pl\n+test_external () {\n+\ttest \"$#\" = 4 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 3 ||\n+\terror >&5 \"bug in the test script: not 3 or 4 parameters to test_external\"\n+\tdescr=\"$1\"\n+\tshift\n+\texport test_prereq\n+\tif ! test_skip \"$descr\" \"$@\"\n+\tthen\n+\t\t# Announce the script to reduce confusion about the\n+\t\t# test output that follows.\n+\t\tsay_color \"\" \"# run $test_count: $descr ($*)\"\n+\t\t# Export TEST_DIRECTORY, TRASH_DIRECTORY and GIT_TEST_LONG\n+\t\t# to be able to use them in script\n+\t\texport TEST_DIRECTORY TRASH_DIRECTORY GIT_TEST_LONG\n+\t\t# Run command; redirect its stderr to &4 as in\n+\t\t# test_run_, but keep its stdout on our stdout even in\n+\t\t# non-verbose mode.\n+\t\t\"$@\" 2>&4\n+\t\tif [ \"$?\" = 0 ]\n+\t\tthen\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_ok_ \"$descr\"\n+\t\t\telse\n+\t\t\t\tsay_color \"\" \"# test_external test $descr was ok\"\n+\t\t\t\ttest_success=$(($test_success + 1))\n+\t\t\tfi\n+\t\telse\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_failure_ \"$descr\" \"$@\"\n+\t\t\telse\n+\t\t\t\tsay_color error \"# test_external test $descr failed: $@\"\n+\t\t\t\ttest_failure=$(($test_failure + 1))\n+\t\t\tfi\n+\t\tfi\n+\tfi\n+}\n+\n+# Like test_external, but in addition tests that the command generated\n+# no output on stderr.\n+test_external_without_stderr () {\n+\t# The temporary file has no (and must have no) security\n+\t# implications.\n+\ttmp=${TMPDIR:-/tmp}\n+\tstderr=\"$tmp/git-external-stderr.$$.tmp\"\n+\ttest_external \"$@\" 4> \"$stderr\"\n+\t[ -f \"$stderr\" ] || error \"Internal error: $stderr disappeared.\"\n+\tdescr=\"no stderr: $1\"\n+\tshift\n+\tsay >&3 \"# expecting no stderr from previous command\"\n+\tif [ ! -s \"$stderr\" ]; then\n+\t\trm \"$stderr\"\n+\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_ok_ \"$descr\"\n+\t\telse\n+\t\t\tsay_color \"\" \"# test_external_without_stderr test $descr was ok\"\n+\t\t\ttest_success=$(($test_success + 1))\n+\t\tfi\n+\telse\n+\t\tif [ \"$verbose\" = t ]; then\n+\t\t\toutput=`echo; echo \"# Stderr is:\"; cat \"$stderr\"`\n+\t\telse\n+\t\t\toutput=\n+\t\tfi\n+\t\t# rm first in case test_failure exits.\n+\t\trm \"$stderr\"\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_failure_ \"$descr\" \"$@\" \"$output\"\n+\t\telse\n+\t\t\tsay_color error \"# test_external_without_stderr test $descr failed: $@: $output\"\n+\t\t\ttest_failure=$(($test_failure + 1))\n+\t\tfi\n+\tfi\n+}\n+\n+# debugging-friendly alternatives to \"test [-f|-d|-e]\"\n+# The commands test the existence or non-existence of $1. $2 can be\n+# given to provide a more precise diagnosis.\n+test_path_is_file () {\n+\tif ! [ -f \"$1\" ]\n+\tthen\n+\t\techo \"File $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_dir () {\n+\tif ! [ -d \"$1\" ]\n+\tthen\n+\t\techo \"Directory $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_missing () {\n+\tif [ -e \"$1\" ]\n+\tthen\n+\t\techo \"Path exists:\"\n+\t\tls -ld \"$1\"\n+\t\tif [ $# -ge 1 ]; then\n+\t\t\techo \"$*\"\n+\t\tfi\n+\t\tfalse\n+\tfi\n+}\n+\n+# test_line_count checks that a file has the number of lines it\n+# ought to. For example:\n+#\n+#\ttest_expect_success 'produce exactly one line of output' '\n+#\t\tdo something >output &&\n+#\t\ttest_line_count = 1 output\n+#\t'\n+#\n+# is like \"test $(wc -l &2 \"test_must_fail: command succeeded: $*\"\n+\t\treturn 1\n+\telif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_must_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_must_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail, but tolerates success, too. This is\n+# meant to be used in contexts like:\n+#\n+#\ttest_expect_success 'some command works without configuration' '\n+#\t\ttest_might_fail git config --unset all.configuration &&\n+#\t\tdo something\n+#\t'\n+#\n+# Writing \"git config --unset all.configuration || :\" would be wrong,\n+# because we want to notice if it fails due to segv.\n+\n+test_might_fail () {\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_might_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_might_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail and test_might_fail, but check that a\n+# given command exited with a given exit code. Meant to be used as:\n+#\n+#\ttest_expect_success 'Merge with d/f conflicts' '\n+#\t\ttest_expect_code 1 git merge \"merge msg\" B master\n+#\t'\n+\n+test_expect_code () {\n+\twant_code=$1\n+\tshift\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code = $want_code\n+\tthen\n+\t\treturn 0\n+\tfi\n+\n+\techo >&2 \"test_expect_code: command exited with $exit_code, we wanted $want_code $*\"\n+\treturn 1\n+}\n+\n+# test_cmp is a helper function to compare actual and expected output.\n+# You can use it like:\n+#\n+#\ttest_expect_success 'foo works' '\n+#\t\techo expected >expected &&\n+#\t\tfoo >actual &&\n+#\t\ttest_cmp expected actual\n+#\t'\n+#\n+# This could be written as either \"cmp\" or \"diff -u\", but:\n+# - cmp's output is not nearly as easy to read as diff -u\n+# - not all diff versions understand \"-u\"\n+\n+test_cmp() {\n+\t$GIT_TEST_CMP \"$@\"\n+}\n+\n+# This function can be used to schedule some commands to be run\n+# unconditionally at the end of the test to restore sanity:\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\ttest_when_finished \"git config --unset core.capslock\" &&\n+#\t\thello world\n+#\t'\n+#\n+# That would be roughly equivalent to\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\thello world\n+#\t\tgit config --unset core.capslock\n+#\t'\n+#\n+# except that the greeting and config --unset must both succeed for\n+# the test to pass.\n+#\n+# Note that under --immediate mode, no clean-up is done to help diagnose\n+# what went wrong.\n+\n+test_when_finished () {\n+\ttest_cleanup=\"{ $*\n+\t\t} && (exit \\\"\\$eval_ret\\\"); eval_ret=\\$?; $test_cleanup\"\n+}\n+\n+# Most tests can use the created repository, but some may need to create more.\n+# Usage: test_create_repo \n+test_create_repo () {\n+\ttest \"$#\" = 1 ||\n+\terror \"bug in the test script: not 1 parameter to test-create-repo\"\n+\trepo=\"$1\"\n+\tmkdir -p \"$repo\"\n+\t(\n+\t\tcd \"$repo\" || error \"Cannot setup test environment\"\n+\t\t\"$GIT_EXEC_PATH/git-init\" \"--template=$GIT_BUILD_DIR/templates/blt/\" >&3 2>&4 ||\n+\t\terror \"cannot run git init -- have you built things yet?\"\n+\t\tmv .git/hooks .git/hooks-disabled\n+\t) || exit\n+}\n+\n+test_expect_failure () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-failure\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"checking known breakage: $2\"\n+\t\tif test_run_ \"$2\" expecting_failure\n+\t\tthen\n+\t\t\ttest_known_broken_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_known_broken_failure_ \"$1\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+test_expect_success () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"expecting success: $2\"\n+\t\tif test_run_ \"$2\"\n+\t\tthen\n+\t\t\ttest_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_failure_ \"$@\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+# test_external runs external test scripts that provide continuous\n+# test output about their progress, and succeeds/fails on\n+# zero/non-zero exit code. It outputs the test output on stdout even\n+# in non-verbose mode, and announces the external script with \"# run\n+# : ...\" before running it. When providing relative paths, keep in\n+# mind that all scripts run in \"trash directory\".\n+# Usage: test_external description command arguments...\n+# Example: test_external 'Perl API' perl ../path/to/test.pl\n+test_external () {\n+\ttest \"$#\" = 4 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 3 ||\n+\terror >&5 \"bug in the test script: not 3 or 4 parameters to test_external\"\n+\tdescr=\"$1\"\n+\tshift\n+\texport test_prereq\n+\tif ! test_skip \"$descr\" \"$@\"\n+\tthen\n+\t\t# Announce the script to reduce confusion about the\n+\t\t# test output that follows.\n+\t\tsay_color \"\" \"# run $test_count: $descr ($*)\"\n+\t\t# Export TEST_DIRECTORY, TRASH_DIRECTORY and GIT_TEST_LONG\n+\t\t# to be able to use them in script\n+\t\texport TEST_DIRECTORY TRASH_DIRECTORY GIT_TEST_LONG\n+\t\t# Run command; redirect its stderr to &4 as in\n+\t\t# test_run_, but keep its stdout on our stdout even in\n+\t\t# non-verbose mode.\n+\t\t\"$@\" 2>&4\n+\t\tif [ \"$?\" = 0 ]\n+\t\tthen\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_ok_ \"$descr\"\n+\t\t\telse\n+\t\t\t\tsay_color \"\" \"# test_external test $descr was ok\"\n+\t\t\t\ttest_success=$(($test_success + 1))\n+\t\t\tfi\n+\t\telse\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_failure_ \"$descr\" \"$@\"\n+\t\t\telse\n+\t\t\t\tsay_color error \"# test_external test $descr failed: $@\"\n+\t\t\t\ttest_failure=$(($test_failure + 1))\n+\t\t\tfi\n+\t\tfi\n+\tfi\n+}\n+\n+# Like test_external, but in addition tests that the command generated\n+# no output on stderr.\n+test_external_without_stderr () {\n+\t# The temporary file has no (and must have no) security\n+\t# implications.\n+\ttmp=${TMPDIR:-/tmp}\n+\tstderr=\"$tmp/git-external-stderr.$$.tmp\"\n+\ttest_external \"$@\" 4> \"$stderr\"\n+\t[ -f \"$stderr\" ] || error \"Internal error: $stderr disappeared.\"\n+\tdescr=\"no stderr: $1\"\n+\tshift\n+\tsay >&3 \"# expecting no stderr from previous command\"\n+\tif [ ! -s \"$stderr\" ]; then\n+\t\trm \"$stderr\"\n+\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_ok_ \"$descr\"\n+\t\telse\n+\t\t\tsay_color \"\" \"# test_external_without_stderr test $descr was ok\"\n+\t\t\ttest_success=$(($test_success + 1))\n+\t\tfi\n+\telse\n+\t\tif [ \"$verbose\" = t ]; then\n+\t\t\toutput=`echo; echo \"# Stderr is:\"; cat \"$stderr\"`\n+\t\telse\n+\t\t\toutput=\n+\t\tfi\n+\t\t# rm first in case test_failure exits.\n+\t\trm \"$stderr\"\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_failure_ \"$descr\" \"$@\" \"$output\"\n+\t\telse\n+\t\t\tsay_color error \"# test_external_without_stderr test $descr failed: $@: $output\"\n+\t\t\ttest_failure=$(($test_failure + 1))\n+\t\tfi\n+\tfi\n+}\n+\n+# debugging-friendly alternatives to \"test [-f|-d|-e]\"\n+# The commands test the existence or non-existence of $1. $2 can be\n+# given to provide a more precise diagnosis.\n+test_path_is_file () {\n+\tif ! [ -f \"$1\" ]\n+\tthen\n+\t\techo \"File $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_dir () {\n+\tif ! [ -d \"$1\" ]\n+\tthen\n+\t\techo \"Directory $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_missing () {\n+\tif [ -e \"$1\" ]\n+\tthen\n+\t\techo \"Path exists:\"\n+\t\tls -ld \"$1\"\n+\t\tif [ $# -ge 1 ]; then\n+\t\t\techo \"$*\"\n+\t\tfi\n+\t\tfalse\n+\tfi\n+}\n+\n+# test_line_count checks that a file has the number of lines it\n+# ought to. For example:\n+#\n+#\ttest_expect_success 'produce exactly one line of output' '\n+#\t\tdo something >output &&\n+#\t\ttest_line_count = 1 output\n+#\t'\n+#\n+# is like \"test $(wc -l &2 \"test_must_fail: command succeeded: $*\"\n+\t\treturn 1\n+\telif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_must_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_must_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail, but tolerates success, too. This is\n+# meant to be used in contexts like:\n+#\n+#\ttest_expect_success 'some command works without configuration' '\n+#\t\ttest_might_fail git config --unset all.configuration &&\n+#\t\tdo something\n+#\t'\n+#\n+# Writing \"git config --unset all.configuration || :\" would be wrong,\n+# because we want to notice if it fails due to segv.\n+\n+test_might_fail () {\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_might_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_might_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail and test_might_fail, but check that a\n+# given command exited with a given exit code. Meant to be used as:\n+#\n+#\ttest_expect_success 'Merge with d/f conflicts' '\n+#\t\ttest_expect_code 1 git merge \"merge msg\" B master\n+#\t'\n+\n+test_expect_code () {\n+\twant_code=$1\n+\tshift\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code = $want_code\n+\tthen\n+\t\treturn 0\n+\tfi\n+\n+\techo >&2 \"test_expect_code: command exited with $exit_code, we wanted $want_code $*\"\n+\treturn 1\n+}\n+\n+# test_cmp is a helper function to compare actual and expected output.\n+# You can use it like:\n+#\n+#\ttest_expect_success 'foo works' '\n+#\t\techo expected >expected &&\n+#\t\tfoo >actual &&\n+#\t\ttest_cmp expected actual\n+#\t'\n+#\n+# This could be written as either \"cmp\" or \"diff -u\", but:\n+# - cmp's output is not nearly as easy to read as diff -u\n+# - not all diff versions understand \"-u\"\n+\n+test_cmp() {\n+\t$GIT_TEST_CMP \"$@\"\n+}\n+\n+# This function can be used to schedule some commands to be run\n+# unconditionally at the end of the test to restore sanity:\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\ttest_when_finished \"git config --unset core.capslock\" &&\n+#\t\thello world\n+#\t'\n+#\n+# That would be roughly equivalent to\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\thello world\n+#\t\tgit config --unset core.capslock\n+#\t'\n+#\n+# except that the greeting and config --unset must both succeed for\n+# the test to pass.\n+#\n+# Note that under --immediate mode, no clean-up is done to help diagnose\n+# what went wrong.\n+\n+test_when_finished () {\n+\ttest_cleanup=\"{ $*\n+\t\t} && (exit \\\"\\$eval_ret\\\"); eval_ret=\\$?; $test_cleanup\"\n+}\n+\n+# Most tests can use the created repository, but some may need to create more.\n+# Usage: test_create_repo \n+test_create_repo () {\n+\ttest \"$#\" = 1 ||\n+\terror \"bug in the test script: not 1 parameter to test-create-repo\"\n+\trepo=\"$1\"\n+\tmkdir -p \"$repo\"\n+\t(\n+\t\tcd \"$repo\" || error \"Cannot setup test environment\"\n+\t\t\"$GIT_EXEC_PATH/git-init\" \"--template=$GIT_BUILD_DIR/templates/blt/\" >&3 2>&4 ||\n+\t\terror \"cannot run git init -- have you built things yet?\"\n+\t\tmv .git/hooks .git/hooks-disabled\n+\t) || exit\n+}\n+\ndiff --git a/t/test-lib.sh b/t/test-lib.sh\nindex e28d5fd..ec70ef2 100644\n--- a/t/test-lib.sh\n+++ b/t/test-lib.sh\n@@ -98,6 +98,8 @@ _z40=0000000000000000000000000000000000000000\n LF='\n '\n \n+export _x05 _x40 _z40 LF\n+\n # Each test should start with something like this, after copyright notices:\n #\n # test_description='Description of this test...\n@@ -223,248 +225,9 @@ die () {\n GIT_EXIT_OK=\n trap 'die' EXIT\n \n-# The semantics of the editor variables are that of invoking\n-# sh -c \"$EDITOR \\\"$@\\\"\" files ...\n-#\n-# If our trash directory contains shell metacharacters, they will be\n-# interpreted if we just set $EDITOR directly, so do a little dance with\n-# environment variables to work around this.\n-#\n-# In particular, quoting isn't enough, as the path may contain the same quote\n-# that we're using.\n-test_set_editor () {\n-\tFAKE_EDITOR=\"$1\"\n-\texport FAKE_EDITOR\n-\tEDITOR='\"$FAKE_EDITOR\"'\n-\texport EDITOR\n-}\n-\n-test_decode_color () {\n-\tawk '\n-\t\tfunction name(n) {\n-\t\t\tif (n == 0) return \"RESET\";\n-\t\t\tif (n == 1) return \"BOLD\";\n-\t\t\tif (n == 30) return \"BLACK\";\n-\t\t\tif (n == 31) return \"RED\";\n-\t\t\tif (n == 32) return \"GREEN\";\n-\t\t\tif (n == 33) return \"YELLOW\";\n-\t\t\tif (n == 34) return \"BLUE\";\n-\t\t\tif (n == 35) return \"MAGENTA\";\n-\t\t\tif (n == 36) return \"CYAN\";\n-\t\t\tif (n == 37) return \"WHITE\";\n-\t\t\tif (n == 40) return \"BLACK\";\n-\t\t\tif (n == 41) return \"BRED\";\n-\t\t\tif (n == 42) return \"BGREEN\";\n-\t\t\tif (n == 43) return \"BYELLOW\";\n-\t\t\tif (n == 44) return \"BBLUE\";\n-\t\t\tif (n == 45) return \"BMAGENTA\";\n-\t\t\tif (n == 46) return \"BCYAN\";\n-\t\t\tif (n == 47) return \"BWHITE\";\n-\t\t}\n-\t\t{\n-\t\t\twhile (match($0, /\\033\\[[0-9;]*m/) != 0) {\n-\t\t\t\tprintf \"%s<\", substr($0, 1, RSTART-1);\n-\t\t\t\tcodes = substr($0, RSTART+2, RLENGTH-3);\n-\t\t\t\tif (length(codes) == 0)\n-\t\t\t\t\tprintf \"%s\", name(0)\n-\t\t\t\telse {\n-\t\t\t\t\tn = split(codes, ary, \";\");\n-\t\t\t\t\tsep = \"\";\n-\t\t\t\t\tfor (i = 1; i <= n; i++) {\n-\t\t\t\t\t\tprintf \"%s%s\", sep, name(ary[i]);\n-\t\t\t\t\t\tsep = \";\"\n-\t\t\t\t\t}\n-\t\t\t\t}\n-\t\t\t\tprintf \">\";\n-\t\t\t\t$0 = substr($0, RSTART + RLENGTH, length($0) - RSTART - RLENGTH + 1);\n-\t\t\t}\n-\t\t\tprint\n-\t\t}\n-\t'\n-}\n-\n-nul_to_q () {\n-\tperl -pe 'y/\\000/Q/'\n-}\n-\n-q_to_nul () {\n-\tperl -pe 'y/Q/\\000/'\n-}\n-\n-q_to_cr () {\n-\ttr Q '\\015'\n-}\n-\n-q_to_tab () {\n-\ttr Q '\\011'\n-}\n-\n-append_cr () {\n-\tsed -e 's/$/Q/' | tr Q '\\015'\n-}\n-\n-remove_cr () {\n-\ttr '\\015' Q | sed -e 's/Q$//'\n-}\n-\n-# In some bourne shell implementations, the \"unset\" builtin returns\n-# nonzero status when a variable to be unset was not set in the first\n-# place.\n-#\n-# Use sane_unset when that should not be considered an error.\n-\n-sane_unset () {\n-\tunset \"$@\"\n-\treturn 0\n-}\n-\n-test_tick () {\n-\tif test -z \"${test_tick+set}\"\n-\tthen\n-\t\ttest_tick=1112911993\n-\telse\n-\t\ttest_tick=$(($test_tick + 60))\n-\tfi\n-\tGIT_COMMITTER_DATE=\"$test_tick -0700\"\n-\tGIT_AUTHOR_DATE=\"$test_tick -0700\"\n-\texport GIT_COMMITTER_DATE GIT_AUTHOR_DATE\n-}\n-\n-# Stop execution and start a shell. This is useful for debugging tests and\n-# only makes sense together with \"-v\".\n-#\n-# Be sure to remove all invocations of this command before submitting.\n-\n-test_pause () {\n-\tif test \"$verbose\" = t; then\n-\t\t\"$SHELL_PATH\" <&6 >&3 2>&4\n-\telse\n-\t\terror >&5 \"test_pause requires --verbose\"\n-\tfi\n-}\n-\n-# Call test_commit with the arguments \" [ []]\"\n-#\n-# This will commit a file with the given contents and the given commit\n-# message. It will also add a tag with as name.\n-#\n-# Both and default to .\n-\n-test_commit () {\n-\tfile=${2:-\"$1.t\"}\n-\techo \"${3-$1}\" > \"$file\" &&\n-\tgit add \"$file\" &&\n-\ttest_tick &&\n-\tgit commit -m \"$1\" &&\n-\tgit tag \"$1\"\n-}\n-\n-# Call test_merge with the arguments \" \", where \n-# can be a tag pointing to the commit-to-merge.\n-\n-test_merge () {\n-\ttest_tick &&\n-\tgit merge -m \"$1\" \"$2\" &&\n-\tgit tag \"$1\"\n-}\n-\n-# This function helps systems where core.filemode=false is set.\n-# Use it instead of plain 'chmod +x' to set or unset the executable bit\n-# of a file in the working directory and add it to the index.\n-\n-test_chmod () {\n-\tchmod \"$@\" &&\n-\tgit update-index --add \"--chmod=$@\"\n-}\n-\n-# Unset a configuration variable, but don't fail if it doesn't exist.\n-test_unconfig () {\n-\tgit config --unset-all \"$@\"\n-\tconfig_status=$?\n-\tcase \"$config_status\" in\n-\t5) # ok, nothing to unset\n-\t\tconfig_status=0\n-\t\t;;\n-\tesac\n-\treturn $config_status\n-}\n-\n-# Set git config, automatically unsetting it after the test is over.\n-test_config () {\n-\ttest_when_finished \"test_unconfig '$1'\" &&\n-\tgit config \"$@\"\n-}\n-\n-\n-test_config_global () {\n-\ttest_when_finished \"test_unconfig --global '$1'\" &&\n-\tgit config --global \"$@\"\n-}\n-\n-write_script () {\n-\t{\n-\t\techo \"#!${2-\"$SHELL_PATH\"}\" &&\n-\t\tcat\n-\t} >\"$1\" &&\n-\tchmod +x \"$1\"\n-}\n-\n-# Use test_set_prereq to tell that a particular prerequisite is available.\n-# The prerequisite can later be checked for in two ways:\n-#\n-# - Explicitly using test_have_prereq.\n-#\n-# - Implicitly by specifying the prerequisite tag in the calls to\n-# test_expect_{success,failure,code}.\n-#\n-# The single parameter is the prerequisite tag (a simple word, in all\n-# capital letters by convention).\n-\n-test_set_prereq () {\n-\tsatisfied=\"$satisfied$1 \"\n-}\n-satisfied=\" \"\n-\n-test_have_prereq () {\n-\t# prerequisites can be concatenated with ','\n-\tsave_IFS=$IFS\n-\tIFS=,\n-\tset -- $*\n-\tIFS=$save_IFS\n-\n-\ttotal_prereq=0\n-\tok_prereq=0\n-\tmissing_prereq=\n-\n-\tfor prerequisite\n-\tdo\n-\t\ttotal_prereq=$(($total_prereq + 1))\n-\t\tcase $satisfied in\n-\t\t*\" $prerequisite \"*)\n-\t\t\tok_prereq=$(($ok_prereq + 1))\n-\t\t\t;;\n-\t\t*)\n-\t\t\t# Keep a list of missing prerequisites\n-\t\t\tif test -z \"$missing_prereq\"\n-\t\t\tthen\n-\t\t\t\tmissing_prereq=$prerequisite\n-\t\t\telse\n-\t\t\t\tmissing_prereq=\"$prerequisite,$missing_prereq\"\n-\t\t\tfi\n-\t\tesac\n-\tdone\n-\n-\ttest $total_prereq = $ok_prereq\n-}\n-\n-test_declared_prereq () {\n-\tcase \",$test_prereq,\" in\n-\t*,$1,*)\n-\t\treturn 0\n-\t\t;;\n-\tesac\n-\treturn 1\n-}\n+# The user-facing functions are loaded from a separate file so that\n+# test_perf subshells can have them too\n+. \"${TEST_DIRECTORY:-.}\"/test-lib-functions.sh\n \n # You are not expected to call test_ok_ and test_failure_ directly, use\n # the text_expect_* functions instead.\n@@ -552,311 +315,6 @@ test_skip () {\n \tesac\n }\n \n-test_expect_failure () {\n-\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 2 ||\n-\terror \"bug in the test script: not 2 or 3 parameters to test-expect-failure\"\n-\texport test_prereq\n-\tif ! test_skip \"$@\"\n-\tthen\n-\t\tsay >&3 \"checking known breakage: $2\"\n-\t\tif test_run_ \"$2\" expecting_failure\n-\t\tthen\n-\t\t\ttest_known_broken_ok_ \"$1\"\n-\t\telse\n-\t\t\ttest_known_broken_failure_ \"$1\"\n-\t\tfi\n-\tfi\n-\techo >&3 \"\"\n-}\n-\n-test_expect_success () {\n-\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 2 ||\n-\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n-\texport test_prereq\n-\tif ! test_skip \"$@\"\n-\tthen\n-\t\tsay >&3 \"expecting success: $2\"\n-\t\tif test_run_ \"$2\"\n-\t\tthen\n-\t\t\ttest_ok_ \"$1\"\n-\t\telse\n-\t\t\ttest_failure_ \"$@\"\n-\t\tfi\n-\tfi\n-\techo >&3 \"\"\n-}\n-\n-# test_external runs external test scripts that provide continuous\n-# test output about their progress, and succeeds/fails on\n-# zero/non-zero exit code. It outputs the test output on stdout even\n-# in non-verbose mode, and announces the external script with \"# run\n-# : ...\" before running it. When providing relative paths, keep in\n-# mind that all scripts run in \"trash directory\".\n-# Usage: test_external description command arguments...\n-# Example: test_external 'Perl API' perl ../path/to/test.pl\n-test_external () {\n-\ttest \"$#\" = 4 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 3 ||\n-\terror >&5 \"bug in the test script: not 3 or 4 parameters to test_external\"\n-\tdescr=\"$1\"\n-\tshift\n-\texport test_prereq\n-\tif ! test_skip \"$descr\" \"$@\"\n-\tthen\n-\t\t# Announce the script to reduce confusion about the\n-\t\t# test output that follows.\n-\t\tsay_color \"\" \"# run $test_count: $descr ($*)\"\n-\t\t# Export TEST_DIRECTORY, TRASH_DIRECTORY and GIT_TEST_LONG\n-\t\t# to be able to use them in script\n-\t\texport TEST_DIRECTORY TRASH_DIRECTORY GIT_TEST_LONG\n-\t\t# Run command; redirect its stderr to &4 as in\n-\t\t# test_run_, but keep its stdout on our stdout even in\n-\t\t# non-verbose mode.\n-\t\t\"$@\" 2>&4\n-\t\tif [ \"$?\" = 0 ]\n-\t\tthen\n-\t\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\t\ttest_ok_ \"$descr\"\n-\t\t\telse\n-\t\t\t\tsay_color \"\" \"# test_external test $descr was ok\"\n-\t\t\t\ttest_success=$(($test_success + 1))\n-\t\t\tfi\n-\t\telse\n-\t\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\t\ttest_failure_ \"$descr\" \"$@\"\n-\t\t\telse\n-\t\t\t\tsay_color error \"# test_external test $descr failed: $@\"\n-\t\t\t\ttest_failure=$(($test_failure + 1))\n-\t\t\tfi\n-\t\tfi\n-\tfi\n-}\n-\n-# Like test_external, but in addition tests that the command generated\n-# no output on stderr.\n-test_external_without_stderr () {\n-\t# The temporary file has no (and must have no) security\n-\t# implications.\n-\ttmp=${TMPDIR:-/tmp}\n-\tstderr=\"$tmp/git-external-stderr.$$.tmp\"\n-\ttest_external \"$@\" 4> \"$stderr\"\n-\t[ -f \"$stderr\" ] || error \"Internal error: $stderr disappeared.\"\n-\tdescr=\"no stderr: $1\"\n-\tshift\n-\tsay >&3 \"# expecting no stderr from previous command\"\n-\tif [ ! -s \"$stderr\" ]; then\n-\t\trm \"$stderr\"\n-\n-\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\ttest_ok_ \"$descr\"\n-\t\telse\n-\t\t\tsay_color \"\" \"# test_external_without_stderr test $descr was ok\"\n-\t\t\ttest_success=$(($test_success + 1))\n-\t\tfi\n-\telse\n-\t\tif [ \"$verbose\" = t ]; then\n-\t\t\toutput=`echo; echo \"# Stderr is:\"; cat \"$stderr\"`\n-\t\telse\n-\t\t\toutput=\n-\t\tfi\n-\t\t# rm first in case test_failure exits.\n-\t\trm \"$stderr\"\n-\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\ttest_failure_ \"$descr\" \"$@\" \"$output\"\n-\t\telse\n-\t\t\tsay_color error \"# test_external_without_stderr test $descr failed: $@: $output\"\n-\t\t\ttest_failure=$(($test_failure + 1))\n-\t\tfi\n-\tfi\n-}\n-\n-# debugging-friendly alternatives to \"test [-f|-d|-e]\"\n-# The commands test the existence or non-existence of $1. $2 can be\n-# given to provide a more precise diagnosis.\n-test_path_is_file () {\n-\tif ! [ -f \"$1\" ]\n-\tthen\n-\t\techo \"File $1 doesn't exist. $*\"\n-\t\tfalse\n-\tfi\n-}\n-\n-test_path_is_dir () {\n-\tif ! [ -d \"$1\" ]\n-\tthen\n-\t\techo \"Directory $1 doesn't exist. $*\"\n-\t\tfalse\n-\tfi\n-}\n-\n-test_path_is_missing () {\n-\tif [ -e \"$1\" ]\n-\tthen\n-\t\techo \"Path exists:\"\n-\t\tls -ld \"$1\"\n-\t\tif [ $# -ge 1 ]; then\n-\t\t\techo \"$*\"\n-\t\tfi\n-\t\tfalse\n-\tfi\n-}\n-\n-# test_line_count checks that a file has the number of lines it\n-# ought to. For example:\n-#\n-#\ttest_expect_success 'produce exactly one line of output' '\n-#\t\tdo something >output &&\n-#\t\ttest_line_count = 1 output\n-#\t'\n-#\n-# is like \"test $(wc -l &2 \"test_must_fail: command succeeded: $*\"\n-\t\treturn 1\n-\telif test $exit_code -gt 129 -a $exit_code -le 192; then\n-\t\techo >&2 \"test_must_fail: died by signal: $*\"\n-\t\treturn 1\n-\telif test $exit_code = 127; then\n-\t\techo >&2 \"test_must_fail: command not found: $*\"\n-\t\treturn 1\n-\tfi\n-\treturn 0\n-}\n-\n-# Similar to test_must_fail, but tolerates success, too. This is\n-# meant to be used in contexts like:\n-#\n-#\ttest_expect_success 'some command works without configuration' '\n-#\t\ttest_might_fail git config --unset all.configuration &&\n-#\t\tdo something\n-#\t'\n-#\n-# Writing \"git config --unset all.configuration || :\" would be wrong,\n-# because we want to notice if it fails due to segv.\n-\n-test_might_fail () {\n-\t\"$@\"\n-\texit_code=$?\n-\tif test $exit_code -gt 129 -a $exit_code -le 192; then\n-\t\techo >&2 \"test_might_fail: died by signal: $*\"\n-\t\treturn 1\n-\telif test $exit_code = 127; then\n-\t\techo >&2 \"test_might_fail: command not found: $*\"\n-\t\treturn 1\n-\tfi\n-\treturn 0\n-}\n-\n-# Similar to test_must_fail and test_might_fail, but check that a\n-# given command exited with a given exit code. Meant to be used as:\n-#\n-#\ttest_expect_success 'Merge with d/f conflicts' '\n-#\t\ttest_expect_code 1 git merge \"merge msg\" B master\n-#\t'\n-\n-test_expect_code () {\n-\twant_code=$1\n-\tshift\n-\t\"$@\"\n-\texit_code=$?\n-\tif test $exit_code = $want_code\n-\tthen\n-\t\treturn 0\n-\tfi\n-\n-\techo >&2 \"test_expect_code: command exited with $exit_code, we wanted $want_code $*\"\n-\treturn 1\n-}\n-\n-# test_cmp is a helper function to compare actual and expected output.\n-# You can use it like:\n-#\n-#\ttest_expect_success 'foo works' '\n-#\t\techo expected >expected &&\n-#\t\tfoo >actual &&\n-#\t\ttest_cmp expected actual\n-#\t'\n-#\n-# This could be written as either \"cmp\" or \"diff -u\", but:\n-# - cmp's output is not nearly as easy to read as diff -u\n-# - not all diff versions understand \"-u\"\n-\n-test_cmp() {\n-\t$GIT_TEST_CMP \"$@\"\n-}\n-\n-# This function can be used to schedule some commands to be run\n-# unconditionally at the end of the test to restore sanity:\n-#\n-#\ttest_expect_success 'test core.capslock' '\n-#\t\tgit config core.capslock true &&\n-#\t\ttest_when_finished \"git config --unset core.capslock\" &&\n-#\t\thello world\n-#\t'\n-#\n-# That would be roughly equivalent to\n-#\n-#\ttest_expect_success 'test core.capslock' '\n-#\t\tgit config core.capslock true &&\n-#\t\thello world\n-#\t\tgit config --unset core.capslock\n-#\t'\n-#\n-# except that the greeting and config --unset must both succeed for\n-# the test to pass.\n-#\n-# Note that under --immediate mode, no clean-up is done to help diagnose\n-# what went wrong.\n-\n-test_when_finished () {\n-\ttest_cleanup=\"{ $*\n-\t\t} && (exit \\\"\\$eval_ret\\\"); eval_ret=\\$?; $test_cleanup\"\n-}\n-\n-# Most tests can use the created repository, but some may need to create more.\n-# Usage: test_create_repo \n-test_create_repo () {\n-\ttest \"$#\" = 1 ||\n-\terror \"bug in the test script: not 1 parameter to test-create-repo\"\n-\trepo=\"$1\"\n-\tmkdir -p \"$repo\"\n-\t(\n-\t\tcd \"$repo\" || error \"Cannot setup test environment\"\n-\t\t\"$GIT_EXEC_PATH/git-init\" \"--template=$GIT_BUILD_DIR/templates/blt/\" >&3 2>&4 ||\n-\t\terror \"cannot run git init -- have you built things yet?\"\n-\t\tmv .git/hooks .git/hooks-disabled\n-\t) || exit\n }\n \n test_done () {\n-- \n1.7.9.1.334.gd1409\n"},{"id":"184843","messageId":"5e712370fcfe32832998c085fdf6b5a7c4e7d64b.1329428159.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329428159.git.trast@student.ethz.ch","subject":"[PATCH v2 2/3] Introduce a performance testing framework","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-16T21:41:14Z","receivedAt":"2012-02-16T21:41:14Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"This introduces a performance testing framework under t/perf/. It\ntries to be as close to the test-lib.sh infrastructure as possible,\nand thus should be easy to get used to for git developers.\n\nThe following points were considered for the implementation:\n\n1. You usually want to compare arbitrary revisions/build trees against\n each other. They may not have the performance test under\n consideration, or even the perf-lib.sh infrastructure.\n\n To cope with this, the 'run' script lets you specify arbitrary\n build dirs and revisions. It even automatically builds the revisions\n if it doesn't have them at hand yet.\n\n2. Usually you would not want to run all tests. It would take too\n long anyway. The 'run' script lets you specify which tests to run;\n or you can also do it manually. There is a Makefile for\n discoverability and 'make clean', but it is not meant for\n real-world use.\n\n3. Creating test repos from scratch in every test is extremely\n time-consuming, and shipping or downloading such large/weird repos\n is out of the question.\n\n We leave this decision to the user. Two different sizes of test\n repos can be configured, and the scripts just copy one or more of\n those (using hardlinks for the object store). By default it tries\n to use the build tree's git.git repository.\n\n This is fairly fast and versatile. Using a copy instead of a clone\n preserves many properties that the user may want to test for, such\n as lots of loose objects, unpacked refs, etc.\n\nSigned-off-by: Thomas Rast \n---\n Makefile | 22 ++++-\n t/Makefile | 43 ++++++++-\n t/perf/.gitignore | 2 +\n t/perf/Makefile | 15 +++\n t/perf/README | 146 ++++++++++++++++++++++++++++\n t/perf/aggregate.perl | 166 ++++++++++++++++++++++++++++++++\n t/perf/min_time.perl | 21 ++++\n t/perf/p0000-perf-lib-sanity.sh | 41 ++++++++\n t/perf/p0001-rev-list.sh | 17 ++++\n t/perf/perf-lib.sh | 198 +++++++++++++++++++++++++++++++++++++++\n t/perf/run | 82 ++++++++++++++++\n t/test-lib.sh | 22 ++++-\n 12 files changed, 770 insertions(+), 5 deletions(-)\n create mode 100644 t/perf/.gitignore\n create mode 100644 t/perf/Makefile\n create mode 100644 t/perf/README\n create mode 100755 t/perf/aggregate.perl\n create mode 100755 t/perf/min_time.perl\n create mode 100755 t/perf/p0000-perf-lib-sanity.sh\n create mode 100755 t/perf/p0001-rev-list.sh\n create mode 100644 t/perf/perf-lib.sh\n create mode 100755 t/perf/run\n\ndiff --git a/Makefile b/Makefile\nindex a0de4e9..1fb1705 100644\n--- a/Makefile\n+++ b/Makefile\n@@ -2361,6 +2361,10 @@ GIT-BUILD-OPTIONS: FORCE\n \t@echo USE_LIBPCRE=\\''$(subst ','\\'',$(subst ','\\'',$(USE_LIBPCRE)))'\\' >>$@\n \t@echo NO_PERL=\\''$(subst ','\\'',$(subst ','\\'',$(NO_PERL)))'\\' >>$@\n \t@echo NO_PYTHON=\\''$(subst ','\\'',$(subst ','\\'',$(NO_PYTHON)))'\\' >>$@\n+\t@echo NO_UNIX_SOCKETS=\\''$(subst ','\\'',$(subst ','\\'',$(NO_UNIX_SOCKETS)))'\\' >>$@\n+ifdef GIT_TEST_OPTS\n+\t@echo GIT_TEST_OPTS=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_TEST_OPTS)))'\\' >>$@\n+endif\n ifdef GIT_TEST_CMP\n \t@echo GIT_TEST_CMP=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_TEST_CMP)))'\\' >>$@\n endif\n@@ -2369,7 +2373,18 @@ ifdef GIT_TEST_CMP_USE_COPIED_CONTEXT\n endif\n \t@echo NO_GETTEXT=\\''$(subst ','\\'',$(subst ','\\'',$(NO_GETTEXT)))'\\' >>$@\n \t@echo GETTEXT_POISON=\\''$(subst ','\\'',$(subst ','\\'',$(GETTEXT_POISON)))'\\' >>$@\n-\t@echo NO_UNIX_SOCKETS=\\''$(subst ','\\'',$(subst ','\\'',$(NO_UNIX_SOCKETS)))'\\' >>$@\n+ifdef GIT_PERF_REPEAT_COUNT\n+\t@echo GIT_PERF_REPEAT_COUNT=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_REPEAT_COUNT)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_REPO\n+\t@echo GIT_PERF_REPO=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_REPO)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_LARGE_REPO\n+\t@echo GIT_PERF_LARGE_REPO=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_LARGE_REPO)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_MAKE_OPTS\n+\t@echo GIT_PERF_MAKE_OPTS=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_MAKE_OPTS)))'\\' >>$@\n+endif\n \n ### Detect Tck/Tk interpreter path changes\n ifndef NO_TCLTK\n@@ -2405,6 +2420,11 @@ export NO_SVN_TESTS\n test: all\n \t$(MAKE) -C t/ all\n \n+perf: all\n+\t$(MAKE) -C t/perf/ all\n+\n+.PHONY: test perf\n+\n test-ctype$X: ctype.o\n \n test-date$X: date.o ctype.o\ndiff --git a/t/Makefile b/t/Makefile\nindex b5048ab..6091211 100644\n--- a/t/Makefile\n+++ b/t/Makefile\n@@ -73,4 +73,45 @@ gitweb-test:\n valgrind:\n \t$(MAKE) GIT_TEST_OPTS=\"$(GIT_TEST_OPTS) --valgrind\"\n \n-.PHONY: pre-clean $(T) aggregate-results clean valgrind\n+perf:\n+\t$(MAKE) -C perf/ all\n+\n+# Smoke testing targets\n+-include ../GIT-VERSION-FILE\n+uname_S := $(shell sh -c 'uname -s 2>/dev/null || echo unknown')\n+uname_M := $(shell sh -c 'uname -m 2>/dev/null || echo unknown')\n+\n+test-results:\n+\tmkdir -p test-results\n+\n+test-results/git-smoke.tar.gz: test-results\n+\t$(PERL_PATH) ./harness \\\n+\t\t--archive=\"test-results/git-smoke.tar.gz\" \\\n+\t\t$(T)\n+\n+smoke: test-results/git-smoke.tar.gz\n+\n+SMOKE_UPLOAD_FLAGS =\n+ifdef SMOKE_USERNAME\n+\tSMOKE_UPLOAD_FLAGS += -F username=\"$(SMOKE_USERNAME)\" -F password=\"$(SMOKE_PASSWORD)\"\n+endif\n+ifdef SMOKE_COMMENT\n+\tSMOKE_UPLOAD_FLAGS += -F comments=\"$(SMOKE_COMMENT)\"\n+endif\n+ifdef SMOKE_TAGS\n+\tSMOKE_UPLOAD_FLAGS += -F tags=\"$(SMOKE_TAGS)\"\n+endif\n+\n+smoke_report: smoke\n+\tcurl \\\n+\t\t-H \"Expect: \" \\\n+\t\t-F project=Git \\\n+\t\t-F architecture=\"$(uname_M)\" \\\n+\t\t-F platform=\"$(uname_S)\" \\\n+\t\t-F revision=\"$(GIT_VERSION)\" \\\n+\t\t-F report_file=@test-results/git-smoke.tar.gz \\\n+\t\t$(SMOKE_UPLOAD_FLAGS) \\\n+\t\thttp://smoke.git.nix.is/app/projects/process_add_report/1 \\\n+\t| grep -v ^Redirecting\n+\n+.PHONY: pre-clean $(T) aggregate-results clean valgrind perf\ndiff --git a/t/perf/.gitignore b/t/perf/.gitignore\nnew file mode 100644\nindex 0000000..50f5cc1\n--- /dev/null\n+++ b/t/perf/.gitignore\n@@ -0,0 +1,2 @@\n+build/\n+test-results/\ndiff --git a/t/perf/Makefile b/t/perf/Makefile\nnew file mode 100644\nindex 0000000..8c47155\n--- /dev/null\n+++ b/t/perf/Makefile\n@@ -0,0 +1,15 @@\n+-include ../../config.mak\n+export GIT_TEST_OPTIONS\n+\n+all: perf\n+\n+perf: pre-clean\n+\t./run\n+\n+pre-clean:\n+\trm -rf test-results\n+\n+clean:\n+\trm -rf build \"trash directory\".* test-results\n+\n+.PHONY: all perf pre-clean clean\ndiff --git a/t/perf/README b/t/perf/README\nnew file mode 100644\nindex 0000000..fdf974a\n--- /dev/null\n+++ b/t/perf/README\n@@ -0,0 +1,146 @@\n+Git performance tests\n+=====================\n+\n+This directory holds performance testing scripts for git tools. The\n+first part of this document describes the various ways in which you\n+can run them.\n+\n+When fixing the tools or adding enhancements, you are strongly\n+encouraged to add tests in this directory to cover what you are\n+trying to fix or enhance. The later part of this short document\n+describes how your test scripts should be organized.\n+\n+\n+Running Tests\n+-------------\n+\n+The easiest way to run tests is to say \"make\". This runs all\n+the tests on the current git repository.\n+\n+ === Running 2 tests in this tree ===\n+ [...]\n+ Test this tree \n+ ---------------------------------------------------------\n+ 0001.1: rev-list --all 0.54(0.51+0.02) \n+ 0001.2: rev-list --all --objects 6.14(5.99+0.11) \n+ 7810.1: grep worktree, cheap regex 0.16(0.16+0.35) \n+ 7810.2: grep worktree, expensive regex 7.90(29.75+0.37)\n+ 7810.3: grep --cached, cheap regex 3.07(3.02+0.25) \n+ 7810.4: grep --cached, expensive regex 9.39(30.57+0.24)\n+\n+You can compare multiple repositories and even git revisions with the\n+'run' script:\n+\n+ $ ./run . origin/next /path/to/git-tree p0001-rev-list.sh\n+\n+where . stands for the current git tree. The full invocation is\n+\n+ ./run [...] [--] [...]\n+\n+A '.' argument is implied if you do not pass any other\n+revisions/directories.\n+\n+You can also manually test this or another git build tree, and then\n+call the aggregation script to summarize the results:\n+\n+ $ ./p0001-rev-list.sh\n+ [...]\n+ $ GIT_BUILD_DIR=/path/to/other/git ./p0001-rev-list.sh\n+ [...]\n+ $ ./aggregate.perl . /path/to/other/git ./p0001-rev-list.sh\n+\n+aggregate.perl has the same invocation as 'run', it just does not run\n+anything beforehand.\n+\n+You can set the following variables (also in your config.mak):\n+\n+ GIT_PERF_REPEAT_COUNT\n+\tNumber of times a test should be repeated for best-of-N\n+\tmeasurements. Defaults to 5.\n+\n+ GIT_PERF_MAKE_OPTS\n+\tOptions to use when automatically building a git tree for\n+\tperformance testing. E.g., -j6 would be useful.\n+\n+ GIT_PERF_REPO\n+ GIT_PERF_LARGE_REPO\n+\tRepositories to copy for the performance tests. The normal\n+\trepo should be at least git.git size. The large repo should\n+\tprobably be about linux-2.6.git size for optimal results.\n+\tBoth default to the git.git you are running from.\n+\n+You can also pass the options taken by ordinary git tests; the most\n+useful one is:\n+\n+--root=::\n+\tCreate \"trash\" directories used to store all temporary data during\n+\ttesting under , instead of the t/ directory.\n+\tUsing this option with a RAM-based filesystem (such as tmpfs)\n+\tcan massively speed up the test suite.\n+\n+\n+Naming Tests\n+------------\n+\n+The performance test files are named as:\n+\n+\tpNNNN-commandname-details.sh\n+\n+where N is a decimal digit. The same conventions for choosing NNNN as\n+for normal tests apply.\n+\n+\n+Writing Tests\n+-------------\n+\n+The perf script starts much like a normal test script, except it\n+sources perf-lib.sh:\n+\n+\t#!/bin/sh\n+\t#\n+\t# Copyright (c) 2005 Junio C Hamano\n+\t#\n+\n+\ttest_description='xxx performance test'\n+\t. ./perf-lib.sh\n+\n+After that you will want to use some of the following:\n+\n+\ttest_perf_default_repo # sets up a \"normal\" repository\n+\ttest_perf_large_repo # sets up a \"large\" repository\n+\n+\ttest_perf_default_repo sub # ditto, in a subdir \"sub\"\n+\n+ test_checkout_worktree # if you need the worktree too\n+\n+At least one of the first two is required!\n+\n+You can use test_expect_success as usual. For actual performance\n+tests, use\n+\n+\ttest_perf 'descriptive string' '\n+\t\tcommand1 &&\n+\t\tcommand2\n+\t'\n+\n+test_perf spawns a subshell, for lack of better options. This means\n+that\n+\n+* you _must_ export all variables that you need in the subshell\n+\n+* you _must_ flag all variables that you want to persist from the\n+ subshell with 'test_export':\n+\n+\ttest_perf 'descriptive string' '\n+\t\tfoo=$(git rev-parse HEAD) &&\n+\t\ttest_export foo\n+\t'\n+\n+ The so-exported variables are automatically marked for export in the\n+ shell executing the perf test. For your convenience, test_export is\n+ the same as export in the main shell.\n+\n+ This feature relies on a bit of magic using 'set' and 'source'.\n+ While we have tried to make sure that it can cope with embedded\n+ whitespace and other special characters, it will not work with\n+ multi-line data.\ndiff --git a/t/perf/aggregate.perl b/t/perf/aggregate.perl\nnew file mode 100755\nindex 0000000..15f7fc1\n--- /dev/null\n+++ b/t/perf/aggregate.perl\n@@ -0,0 +1,166 @@\n+#!/usr/bin/perl\n+\n+use strict;\n+use warnings;\n+use Git;\n+\n+sub get_times {\n+\tmy $name = shift;\n+\topen my $fh, \"<\", $name or return undef;\n+\tmy $line = <$fh>;\n+\treturn undef if not defined $line;\n+\tclose $fh or die \"cannot close $name: $!\";\n+\t$line =~ /^(?:(\\d+):)?(\\d+):(\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?)$/\n+\t\tor die \"bad input line: $line\";\n+\tmy $rt = ((defined $1 ? $1 : 0.0)*60+$2)*60+$3;\n+\treturn ($rt, $4, $5);\n+}\n+\n+sub format_times {\n+\tmy ($r, $u, $s, $firstr) = @_;\n+\tif (!defined $r) {\n+\t\treturn \"\";\n+\t}\n+\tmy $out = sprintf \"%.2f(%.2f+%.2f)\", $r, $u, $s;\n+\tif (defined $firstr) {\n+\t\tif ($firstr > 0) {\n+\t\t\t$out .= sprintf \" %+.1f%%\", 100.0*($r-$firstr)/$firstr;\n+\t\t} elsif ($r == 0) {\n+\t\t\t$out .= \" =\";\n+\t\t} else {\n+\t\t\t$out .= \" +inf\";\n+\t\t}\n+\t}\n+\treturn $out;\n+}\n+\n+my (@dirs, %dirnames, %dirabbrevs, %prefixes, @tests);\n+while (scalar @ARGV) {\n+\tmy $arg = $ARGV[0];\n+\tmy $dir;\n+\tlast if -f $arg or $arg eq \"--\";\n+\tif (! -d $arg) {\n+\t\tmy $rev = Git::command_oneline(qw(rev-parse --verify), $arg);\n+\t\t$dir = \"build/\".$rev;\n+\t} else {\n+\t\t$arg =~ s{/*$}{};\n+\t\t$dir = $arg;\n+\t\t$dirabbrevs{$dir} = $dir;\n+\t}\n+\tpush @dirs, $dir;\n+\t$dirnames{$dir} = $arg;\n+\tmy $prefix = $dir;\n+\t$prefix =~ tr/^a-zA-Z0-9/_/c;\n+\t$prefixes{$dir} = $prefix . '.';\n+\tshift @ARGV;\n+}\n+\n+if (not @dirs) {\n+\t@dirs = ('.');\n+}\n+$dirnames{'.'} = $dirabbrevs{'.'} = \"this tree\";\n+$prefixes{'.'} = '';\n+\n+shift @ARGV if scalar @ARGV and $ARGV[0] eq \"--\";\n+\n+@tests = @ARGV;\n+if (not @tests) {\n+\t@tests = glob \"p????-*.sh\";\n+}\n+\n+my @subtests;\n+my %shorttests;\n+for my $t (@tests) {\n+\t$t =~ s{(?:.*/)?(p(\\d+)-[^/]+)\\.sh$}{$1} or die \"bad test name: $t\";\n+\tmy $n = $2;\n+\tmy $fname = \"test-results/$t.subtests\";\n+\topen my $fp, \"<\", $fname or die \"cannot open $fname: $!\";\n+\tfor (<$fp>) {\n+\t\tchomp;\n+\t\t/^(\\d+)$/ or die \"malformed subtest line: $_\";\n+\t\tpush @subtests, \"$t.$1\";\n+\t\t$shorttests{\"$t.$1\"} = \"$n.$1\";\n+\t}\n+\tclose $fp or die \"cannot close $fname: $!\";\n+}\n+\n+sub read_descr {\n+\tmy $name = shift;\n+\topen my $fh, \"<\", $name or return \"\";\n+\tmy $line = <$fh>;\n+\tclose $fh or die \"cannot close $name\";\n+\tchomp $line;\n+\treturn $line;\n+}\n+\n+my %descrs;\n+my $descrlen = 4; # \"Test\"\n+for my $t (@subtests) {\n+\t$descrs{$t} = $shorttests{$t}.\": \".read_descr(\"test-results/$t.descr\");\n+\t$descrlen = length $descrs{$t} if length $descrs{$t}>$descrlen;\n+}\n+\n+sub have_duplicate {\n+\tmy %seen;\n+\tfor (@_) {\n+\t\treturn 1 if exists $seen{$_};\n+\t\t$seen{$_} = 1;\n+\t}\n+\treturn 0;\n+}\n+sub have_slash {\n+\tfor (@_) {\n+\t\treturn 1 if m{/};\n+\t}\n+\treturn 0;\n+}\n+\n+my %newdirabbrevs = %dirabbrevs;\n+while (!have_duplicate(values %newdirabbrevs)) {\n+\t%dirabbrevs = %newdirabbrevs;\n+\tlast if !have_slash(values %dirabbrevs);\n+\t%newdirabbrevs = %dirabbrevs;\n+\tfor (values %newdirabbrevs) {\n+\t\ts{^[^/]*/}{};\n+\t}\n+}\n+\n+my %times;\n+my @colwidth = ((0)x@dirs);\n+for my $i (0..$#dirs) {\n+\tmy $d = $dirs[$i];\n+\tmy $w = length (exists $dirabbrevs{$d} ? $dirabbrevs{$d} : $dirnames{$d});\n+\t$colwidth[$i] = $w if $w > $colwidth[$i];\n+}\n+for my $t (@subtests) {\n+\tmy $firstr;\n+\tfor my $i (0..$#dirs) {\n+\t\tmy $d = $dirs[$i];\n+\t\t$times{$prefixes{$d}.$t} = [get_times(\"test-results/$prefixes{$d}$t.times\")];\n+\t\tmy ($r,$u,$s) = @{$times{$prefixes{$d}.$t}};\n+\t\tmy $w = length format_times($r,$u,$s,$firstr);\n+\t\t$colwidth[$i] = $w if $w > $colwidth[$i];\n+\t\t$firstr = $r unless defined $firstr;\n+\t}\n+}\n+my $totalwidth = 3*@dirs+$descrlen;\n+$totalwidth += $_ for (@colwidth);\n+\n+printf \"%-${descrlen}s\", \"Test\";\n+for my $i (0..$#dirs) {\n+\tmy $d = $dirs[$i];\n+\tprintf \" %-$colwidth[$i]s\", (exists $dirabbrevs{$d} ? $dirabbrevs{$d} : $dirnames{$d});\n+}\n+print \"\\n\";\n+print \"-\"x$totalwidth, \"\\n\";\n+for my $t (@subtests) {\n+\tprintf \"%-${descrlen}s\", $descrs{$t};\n+\tmy $firstr;\n+\tfor my $i (0..$#dirs) {\n+\t\tmy $d = $dirs[$i];\n+\t\tmy ($r,$u,$s) = @{$times{$prefixes{$d}.$t}};\n+\t\tprintf \" %-$colwidth[$i]s\", format_times($r,$u,$s,$firstr);\n+\t\t$firstr = $r unless defined $firstr;\n+\t}\n+\tprint \"\\n\";\n+}\ndiff --git a/t/perf/min_time.perl b/t/perf/min_time.perl\nnew file mode 100755\nindex 0000000..c1a2717\n--- /dev/null\n+++ b/t/perf/min_time.perl\n@@ -0,0 +1,21 @@\n+#!/usr/bin/perl\n+\n+my $minrt = 1e100;\n+my $min;\n+\n+while (<>) {\n+\t# [h:]m:s.xx U.xx S.xx\n+\t/^(?:(\\d+):)?(\\d+):(\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?)$/\n+\t\tor die \"bad input line: $_\";\n+\tmy $rt = ((defined $1 ? $1 : 0.0)*60+$2)*60+$3;\n+\tif ($rt < $minrt) {\n+\t\t$min = $_;\n+\t\t$minrt = $rt;\n+\t}\n+}\n+\n+if (!defined $min) {\n+\tdie \"no input found\";\n+}\n+\n+print $min;\ndiff --git a/t/perf/p0000-perf-lib-sanity.sh b/t/perf/p0000-perf-lib-sanity.sh\nnew file mode 100755\nindex 0000000..2ca4aac\n--- /dev/null\n+++ b/t/perf/p0000-perf-lib-sanity.sh\n@@ -0,0 +1,41 @@\n+#!/bin/sh\n+\n+test_description='Tests whether perf-lib facilities work'\n+. ./perf-lib.sh\n+\n+test_perf_default_repo\n+\n+test_perf 'test_perf_default_repo works' '\n+\tfoo=$(git rev-parse HEAD) &&\n+\ttest_export foo\n+'\n+\n+test_checkout_worktree\n+\n+test_perf 'test_checkout_worktree works' '\n+\twt=$(find . | wc -l) &&\n+\tidx=$(git ls-files | wc -l) &&\n+\ttest $wt -gt $idx\n+'\n+\n+baz=baz\n+test_export baz\n+\n+test_expect_success 'test_export works' '\n+\techo \"$foo\" &&\n+\ttest \"$foo\" = \"$(git rev-parse HEAD)\" &&\n+\techo \"$baz\" &&\n+\ttest \"$baz\" = baz\n+'\n+\n+test_perf 'export a weird var' '\n+\tbar=\"weird # variable\" &&\n+\ttest_export bar\n+'\n+\n+test_expect_success 'test_export works with weird vars' '\n+\techo \"$bar\" &&\n+\ttest \"$bar\" = \"weird # variable\"\n+'\n+\n+test_done\ndiff --git a/t/perf/p0001-rev-list.sh b/t/perf/p0001-rev-list.sh\nnew file mode 100755\nindex 0000000..4f71a63\n--- /dev/null\n+++ b/t/perf/p0001-rev-list.sh\n@@ -0,0 +1,17 @@\n+#!/bin/sh\n+\n+test_description=\"Tests history walking performance\"\n+\n+. ./perf-lib.sh\n+\n+test_perf_default_repo\n+\n+test_perf 'rev-list --all' '\n+\tgit rev-list --all >/dev/null\n+'\n+\n+test_perf 'rev-list --all --objects' '\n+\tgit rev-list --all --objects >/dev/null\n+'\n+\n+test_done\ndiff --git a/t/perf/perf-lib.sh b/t/perf/perf-lib.sh\nnew file mode 100644\nindex 0000000..07e9b09\n--- /dev/null\n+++ b/t/perf/perf-lib.sh\n@@ -0,0 +1,198 @@\n+#!/bin/bash\n+#\n+# Copyright (c) 2011 Thomas Rast\n+#\n+# This program is free software: you can redistribute it and/or modify\n+# it under the terms of the GNU General Public License as published by\n+# the Free Software Foundation, either version 2 of the License, or\n+# (at your option) any later version.\n+#\n+# This program is distributed in the hope that it will be useful,\n+# but WITHOUT ANY WARRANTY; without even the implied warranty of\n+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n+# GNU General Public License for more details.\n+#\n+# You should have received a copy of the GNU General Public License\n+# along with this program. If not, see http://www.gnu.org/licenses/ .\n+\n+# do the --tee work early; it otherwise confuses our careful\n+# GIT_BUILD_DIR mangling\n+case \"$GIT_TEST_TEE_STARTED, $* \" in\n+done,*)\n+\t# do not redirect again\n+\t;;\n+*' --tee '*|*' --va'*)\n+\tmkdir -p test-results\n+\tBASE=test-results/$(basename \"$0\" .sh)\n+\t(GIT_TEST_TEE_STARTED=done ${SHELL-sh} \"$0\" \"$@\" 2>&1;\n+\t echo $? > $BASE.exit) | tee $BASE.out\n+\ttest \"$(cat $BASE.exit)\" = 0\n+\texit\n+\t;;\n+esac\n+\n+TEST_DIRECTORY=$(pwd)/..\n+TEST_OUTPUT_DIRECTORY=$(pwd)\n+if test -z \"$GIT_TEST_INSTALLED\"; then\n+\tperf_results_prefix=\n+else\n+\tperf_results_prefix=$(printf \"%s\" \"${GIT_TEST_INSTALLED%/bin-wrappers}\" | tr -c \"[a-zA-Z0-9]\" \"[_*]\")\".\"\n+\t# make the tested dir absolute\n+\tGIT_TEST_INSTALLED=$(cd \"$GIT_TEST_INSTALLED\" && pwd)\n+fi\n+\n+TEST_NO_CREATE_REPO=t\n+\n+. ../test-lib.sh\n+\n+perf_results_dir=$TEST_OUTPUT_DIRECTORY/test-results\n+mkdir -p \"$perf_results_dir\"\n+rm -f \"$perf_results_dir\"/$(basename \"$0\" .sh).subtests\n+\n+if test -z \"$GIT_PERF_REPEAT_COUNT\"; then\n+\tGIT_PERF_REPEAT_COUNT=3\n+fi\n+die_if_build_dir_not_repo () {\n+\tif ! ( cd \"$TEST_DIRECTORY/..\" &&\n+\t\t git rev-parse --build-dir >/dev/null 2>&1 ); then\n+\t\terror \"No $1 defined, and your build directory is not a repo\"\n+\tfi\n+}\n+\n+if test -z \"$GIT_PERF_REPO\"; then\n+\tdie_if_build_dir_not_repo '$GIT_PERF_REPO'\n+\tGIT_PERF_REPO=$TEST_DIRECTORY/..\n+fi\n+if test -z \"$GIT_PERF_LARGE_REPO\"; then\n+\tdie_if_build_dir_not_repo '$GIT_PERF_LARGE_REPO'\n+\tGIT_PERF_LARGE_REPO=$TEST_DIRECTORY/..\n+fi\n+\n+test_perf_create_repo_from () {\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 parameters to test-create-repo\"\n+\trepo=\"$1\"\n+\tsource=\"$2\"\n+\tsource_git=$source/$(cd \"$source\" && git rev-parse --git-dir)\n+\tmkdir -p \"$repo/.git\"\n+\t(\n+\t\tcd \"$repo/.git\" &&\n+\t\t{ cp -Rl \"$source_git/objects\" . 2>/dev/null ||\n+\t\t\tcp -R \"$source_git/objects\" .; } &&\n+\t\tfor stuff in \"$source_git\"/*; do\n+\t\t\tcase \"$stuff\" in\n+\t\t\t\t*/objects|*/hooks|*/config)\n+\t\t\t\t\t;;\n+\t\t\t\t*)\n+\t\t\t\t\tcp -R \"$stuff\" . || break\n+\t\t\t\t\t;;\n+\t\t\tesac\n+\t\tdone &&\n+\t\tcd .. &&\n+\t\tgit init -q &&\n+\t\tmv .git/hooks .git/hooks-disabled 2>/dev/null\n+\t) || error \"failed to copy repository '$source' to '$repo'\"\n+}\n+\n+# call at least one of these to establish an appropriately-sized repository\n+test_perf_default_repo () {\n+\ttest_perf_create_repo_from \"${1:-$TRASH_DIRECTORY}\" \"$GIT_PERF_REPO\"\n+}\n+test_perf_large_repo () {\n+\tif test \"$GIT_PERF_LARGE_REPO\" = \"$GIT_BUILD_DIR\"; then\n+\t\techo \"warning: \\$GIT_PERF_LARGE_REPO is \\$GIT_BUILD_DIR.\" >&2\n+\t\techo \"warning: This will work, but may not be a sufficiently large repo\" >&2\n+\t\techo \"warning: for representative measurements.\" >&2\n+\tfi\n+\ttest_perf_create_repo_from \"${1:-$TRASH_DIRECTORY}\" \"$GIT_PERF_LARGE_REPO\"\n+}\n+test_checkout_worktree () {\n+\tgit checkout-index -u -a ||\n+\terror \"git checkout-index failed\"\n+}\n+\n+# Performance tests should never fail. If they do, stop immediately\n+immediate=t\n+\n+test_run_perf_ () {\n+\ttest_cleanup=:\n+\ttest_export_=\"test_cleanup\"\n+\texport test_cleanup test_export_\n+\t/usr/bin/time -f \"%E %U %S\" -o test_time.$i \"$SHELL\" -c '\n+. '\"$TEST_DIRECTORY\"/../test-lib-functions.sh'\n+test_export () {\n+\t[ $# != 0 ] || return 0\n+\ttest_export_=\"$test_export_\\\\|$1\"\n+\tshift\n+\ttest_export \"$@\"\n+}\n+'\"$1\"'\n+ret=$?\n+set | sed -n \"s'\"/'/'\\\\\\\\''/g\"';s/^\\\\($test_export_\\\\)/export '\"'&'\"'/p\" >test_vars\n+exit $ret' >&3 2>&4\n+\teval_ret=$?\n+\n+\tif test $eval_ret = 0 || test -n \"$expecting_failure\"\n+\tthen\n+\t\ttest_eval_ \"$test_cleanup\"\n+\t\tsource ./test_vars || error \"failed to load updated environment\"\n+\tfi\n+\tif test \"$verbose\" = \"t\" && test -n \"$HARNESS_ACTIVE\"; then\n+\t\techo \"\"\n+\tfi\n+\treturn \"$eval_ret\"\n+}\n+\n+\n+test_perf () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tbase=$(basename \"$0\" .sh)\n+\t\techo \"$test_count\" >>\"$perf_results_dir\"/$base.subtests\n+\t\techo \"$1\" >\"$perf_results_dir\"/$base.$test_count.descr\n+\t\tif test -z \"$verbose\"; then\n+\t\t\techo -n \"perf $test_count - $1:\"\n+\t\telse\n+\t\t\techo \"perf $test_count - $1:\"\n+\t\tfi\n+\t\tfor i in $(seq 1 $GIT_PERF_REPEAT_COUNT); do\n+\t\t\tsay >&3 \"running: $2\"\n+\t\t\tif test_run_perf_ \"$2\"\n+\t\t\tthen\n+\t\t\t\tif test -z \"$verbose\"; then\n+\t\t\t\t\techo -n \" $i\"\n+\t\t\t\telse\n+\t\t\t\t\techo \"* timing run $i/$GIT_PERF_REPEAT_COUNT:\"\n+\t\t\t\tfi\n+\t\t\telse\n+\t\t\t\ttest -z \"$verbose\" && echo\n+\t\t\t\ttest_failure_ \"$@\"\n+\t\t\t\tbreak\n+\t\t\tfi\n+\t\tdone\n+\t\tif test -z \"$verbose\"; then\n+\t\t\techo \" ok\"\n+\t\telse\n+\t\t\ttest_ok_ \"$1\"\n+\t\tfi\n+\t\tbase=\"$perf_results_dir\"/\"$perf_results_prefix$(basename \"$0\" .sh)\".\"$test_count\"\n+\t\t\"$TEST_DIRECTORY\"/perf/min_time.perl test_time.* >\"$base\".times\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+# We extend test_done to print timings at the end (./run disables this\n+# and does it after running everything)\n+test_at_end_hook_ () {\n+\tif test -z \"$GIT_PERF_AGGREGATING_LATER\"; then\n+\t\t( cd \"$TEST_DIRECTORY\"/perf && ./aggregate.perl $(basename \"$0\") )\n+\tfi\n+}\n+\n+test_export () {\n+\texport \"$@\"\n+}\ndiff --git a/t/perf/run b/t/perf/run\nnew file mode 100755\nindex 0000000..cfd7012\n--- /dev/null\n+++ b/t/perf/run\n@@ -0,0 +1,82 @@\n+#!/bin/sh\n+\n+case \"$1\" in\n+\t--help)\n+\t\techo \"usage: $0 [other_git_tree...] [--] [test_scripts]\"\n+\t\texit 0\n+\t\t;;\n+esac\n+\n+die () {\n+\techo >&2 \"error: $*\"\n+\texit 1\n+}\n+\n+run_one_dir () {\n+\tif test $# -eq 0; then\n+\t\tset -- p????-*.sh\n+\tfi\n+\techo \"=== Running $# tests in ${GIT_TEST_INSTALLED:-this tree} ===\"\n+\tfor t in \"$@\"; do\n+\t\t./$t $GIT_TEST_OPTS\n+\tdone\n+}\n+\n+unpack_git_rev () {\n+\trev=$1\n+\tmkdir -p build/$rev\n+\t(cd \"$(git rev-parse --show-cdup)\" && git archive --format=tar $rev) |\n+\t(cd build/$rev && tar x)\n+}\n+build_git_rev () {\n+\trev=$1\n+\tcp ../../config.mak build/$rev/config.mak\n+\t(cd build/$rev && make $GIT_PERF_MAKE_OPTS) ||\n+\tdie \"failed to build revision '$mydir'\"\n+}\n+\n+run_dirs_helper () {\n+\tmydir=${1%/}\n+\tshift\n+\twhile test $# -gt 0 -a \"$1\" != -- -a ! -f \"$1\"; do\n+\t\tshift\n+\tdone\n+\tif test $# -gt 0 -a \"$1\" = --; then\n+\t\tshift\n+\tfi\n+\tif [ ! -d \"$mydir\" ]; then\n+\t\trev=$(git rev-parse --verify \"$mydir\" 2>/dev/null) ||\n+\t\tdie \"'$mydir' is neither a directory nor a valid revision\"\n+\t\tif [ ! -d build/$rev ]; then\n+\t\t\tunpack_git_rev $rev\n+\t\tfi\n+\t\tbuild_git_rev $rev\n+\t\tmydir=build/$rev\n+\tfi\n+\tif test \"$mydir\" = .; then\n+\t\tunset GIT_TEST_INSTALLED\n+\telse\n+\t\tGIT_TEST_INSTALLED=\"$mydir/bin-wrappers\"\n+\t\texport GIT_TEST_INSTALLED\n+\tfi\n+\trun_one_dir \"$@\"\n+}\n+\n+run_dirs () {\n+\twhile test $# -gt 0 -a \"$1\" != -- -a ! -f \"$1\"; do\n+\t\trun_dirs_helper \"$@\"\n+\t\tshift\n+\tdone\n+}\n+\n+GIT_PERF_AGGREGATING_LATER=t\n+export GIT_PERF_AGGREGATING_LATER\n+\n+cd \"$(dirname $0)\"\n+. ../../GIT-BUILD-OPTIONS\n+\n+if test $# = 0 -o \"$1\" = -- -o -f \"$1\"; then\n+\tset -- . \"$@\"\n+fi\n+run_dirs \"$@\"\n+./aggregate.perl \"$@\"\ndiff --git a/t/test-lib.sh b/t/test-lib.sh\nindex ec70ef2..d75766a 100644\n--- a/t/test-lib.sh\n+++ b/t/test-lib.sh\n@@ -55,6 +55,7 @@ unset $(perl -e '\n \t\t.*_TEST\n \t\tPROVE\n \t\tVALGRIND\n+\t\tPERF_AGGREGATING_LATER\n \t));\n \tmy @vars = grep(/^GIT_/ && !/^GIT_($ok)/o, @env);\n \tprint join(\"\\n\", @vars);\n@@ -315,13 +316,16 @@ test_skip () {\n \tesac\n }\n \n+# stub; perf-lib overrides it\n+test_at_end_hook_ () {\n+\t:\n }\n \n test_done () {\n \tGIT_EXIT_OK=t\n \n \tif test -z \"$HARNESS_ACTIVE\"; then\n-\t\ttest_results_dir=\"$TEST_DIRECTORY/test-results\"\n+\t\ttest_results_dir=\"$TEST_OUTPUT_DIRECTORY/test-results\"\n \t\tmkdir -p \"$test_results_dir\"\n \t\ttest_results_path=\"$test_results_dir/${0%.sh}-$$.counts\"\n \n@@ -360,6 +364,8 @@ test_done () {\n \t\tcd \"$(dirname \"$remove_trash\")\" &&\n \t\trm -rf \"$(basename \"$remove_trash\")\"\n \n+\t\ttest_at_end_hook_\n+\n \t\texit 0 ;;\n \n \t*)\n@@ -382,6 +388,12 @@ then\n \t# itself.\n \tTEST_DIRECTORY=$(pwd)\n fi\n+if test -z \"$TEST_OUTPUT_DIRECTORY\"\n+then\n+\t# Similarly, override this to store the test-results subdir\n+\t# elsewhere\n+\tTEST_OUTPUT_DIRECTORY=$TEST_DIRECTORY\n+fi\n GIT_BUILD_DIR=\"$TEST_DIRECTORY\"/..\n \n if test -n \"$valgrind\"\n@@ -517,7 +529,7 @@ test=\"trash directory.$(basename \"$0\" .sh)\"\n test -n \"$root\" && test=\"$root/$test\"\n case \"$test\" in\n /*) TRASH_DIRECTORY=\"$test\" ;;\n- *) TRASH_DIRECTORY=\"$TEST_DIRECTORY/$test\" ;;\n+ *) TRASH_DIRECTORY=\"$TEST_OUTPUT_DIRECTORY/$test\" ;;\n esac\n test ! -z \"$debug\" || remove_trash=$TRASH_DIRECTORY\n rm -fr \"$test\" || {\n@@ -529,7 +541,11 @@ rm -fr \"$test\" || {\n HOME=\"$TRASH_DIRECTORY\"\n export HOME\n \n-test_create_repo \"$test\"\n+if test -z \"$TEST_NO_CREATE_REPO\"; then\n+\ttest_create_repo \"$test\"\n+else\n+\tmkdir -p \"$test\"\n+fi\n # Use -P to resolve symlinks in our working directory so that the cwd\n # in subprocesses like git equals our $PWD (for pathname comparisons).\n cd -P \"$test\" || exit 1\n-- \n1.7.9.1.334.gd1409\n"},{"id":"184841","messageId":"9186cd4a628dc240f8519cddd3ca4e6fad0fce9f.1329428159.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329428159.git.trast@student.ethz.ch","subject":"[PATCH v2 3/3] Add a performance test for git-grep","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-16T21:41:15Z","receivedAt":"2012-02-16T21:41:15Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"The only catch is that we don't really know what our repo contains, so\nwe have to ignore any possible \"not found\" status from git-grep.\n\nSigned-off-by: Thomas Rast \n---\n t/perf/p7810-grep.sh | 23 +++++++++++++++++++++++\n 1 file changed, 23 insertions(+)\n create mode 100755 t/perf/p7810-grep.sh\n\ndiff --git a/t/perf/p7810-grep.sh b/t/perf/p7810-grep.sh\nnew file mode 100755\nindex 0000000..9f4ade6\n--- /dev/null\n+++ b/t/perf/p7810-grep.sh\n@@ -0,0 +1,23 @@\n+#!/bin/sh\n+\n+test_description=\"git-grep performance in various modes\"\n+\n+. ./perf-lib.sh\n+\n+test_perf_large_repo\n+test_checkout_worktree\n+\n+test_perf 'grep worktree, cheap regex' '\n+\tgit grep some_nonexistent_string || :\n+'\n+test_perf 'grep worktree, expensive regex' '\n+\tgit grep \"^.* *some_nonexistent_string$\" || :\n+'\n+test_perf 'grep --cached, cheap regex' '\n+\tgit grep --cached some_nonexistent_string || :\n+'\n+test_perf 'grep --cached, expensive regex' '\n+\tgit grep --cached \"^.* *some_nonexistent_string$\" || :\n+'\n+\n+test_done\n-- \n1.7.9.1.334.gd1409\n"},{"id":"184846","messageId":"7v7gzmxw78.fsf@alter.siamese.dyndns.org","threadId":"29643","inReplyTo":"d5127b0051d354fc0c02666b972b853d1736d09c.1329428159.git.trast@student.ethz.ch","subject":"Re: [PATCH v2 1/3] Move the user-facing test library to test-lib-functions.sh","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2012-02-16T22:14:19Z","receivedAt":"2012-02-16T22:14:19Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Thomas Rast writes:\n\n> This just moves all the user-facing functions to a separate file and\n> sources that instead.\n>\n> Signed-off-by: Thomas Rast \n> ---\n> t/test-lib-functions.sh | 835 +++++++++++++++++++++++++++++++++++++++++++++++\n> t/test-lib.sh | 552 +-------------------------------\n> 2 files changed, 840 insertions(+), 547 deletions(-)\n> create mode 100644 t/test-lib-functions.sh\n\nI would have expected from the log description that the number of deleted\nlines would be about the same as the number of added lines, and the\ndifference would primarily come from the addition of \"include\" aka \"dot\"\n\". ./test-lib-functions.sh\" that becomes necessary in t/test-lib.sh, some\nboilerplate material at the beginning of the new file e.g. \"#!/bin/sh\",\nand copying (not moving) the same Copyright block to the new file.\n\nBut 835-552 = 283 feels way way more than that. What else is going on?\n"},{"id":"184877","messageId":"20120217074543.GA25213@sigill.intra.peff.net","threadId":"29643","inReplyTo":"5e712370fcfe32832998c085fdf6b5a7c4e7d64b.1329428159.git.trast@student.ethz.ch","subject":"Re: [PATCH v2 2/3] Introduce a performance testing framework","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2012-02-17T07:45:43Z","receivedAt":"2012-02-17T07:45:43Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Thu, Feb 16, 2012 at 10:41:14PM +0100, Thomas Rast wrote:\n\n> +\tif test $eval_ret = 0 || test -n \"$expecting_failure\"\n> +\tthen\n> +\t\ttest_eval_ \"$test_cleanup\"\n> +\t\tsource ./test_vars || error \"failed to load updated environment\"\n> +\tfi\n\n\"source\" is a bash-ism (actually, it is a csh-ism as far as I know, but\nperhaps it predates even that). The correct POSIX spelling is \".\".\n\nAfter tweaking this line, the suite seems to run fine for me with dash\nas /bin/sh. Patch is below for your convenience.\n\n-Peff\n\n---\ndiff --git a/t/perf/perf-lib.sh b/t/perf/perf-lib.sh\nindex 07e9b09..629d7d5 100644\n--- a/t/perf/perf-lib.sh\n+++ b/t/perf/perf-lib.sh\n@@ -135,7 +135,7 @@ exit $ret' >&3 2>&4\n \tif test $eval_ret = 0 || test -n \"$expecting_failure\"\n \tthen\n \t\ttest_eval_ \"$test_cleanup\"\n-\t\tsource ./test_vars || error \"failed to load updated environment\"\n+\t\t. ./test_vars || error \"failed to load updated environment\"\n \tfi\n \tif test \"$verbose\" = \"t\" && test -n \"$HARNESS_ACTIVE\"; then\n \t\techo \"\"\n"},{"id":"184869","messageId":"8762f5rg14.fsf@thomas.inf.ethz.ch","threadId":"29643","inReplyTo":"7v7gzmxw78.fsf@alter.siamese.dyndns.org","subject":"Re: [PATCH v2 1/3] Move the user-facing test library to test-lib-functions.sh","fromName":"Thomas Rast","fromEmail":"trast@inf.ethz.ch","sentAt":"2012-02-17T09:00:07Z","receivedAt":"2012-02-17T09:00:07Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"Junio C Hamano writes:\n\n> Thomas Rast writes:\n>\n>> This just moves all the user-facing functions to a separate file and\n>> sources that instead.\n>>\n>> Signed-off-by: Thomas Rast \n>> ---\n>> t/test-lib-functions.sh | 835 +++++++++++++++++++++++++++++++++++++++++++++++\n>> t/test-lib.sh | 552 +-------------------------------\n>> 2 files changed, 840 insertions(+), 547 deletions(-)\n>> create mode 100644 t/test-lib-functions.sh\n>\n> I would have expected from the log description that the number of deleted\n> lines would be about the same as the number of added lines, and the\n> difference would primarily come from the addition of \"include\" aka \"dot\"\n> \". ./test-lib-functions.sh\" that becomes necessary in t/test-lib.sh, some\n> boilerplate material at the beginning of the new file e.g. \"#!/bin/sh\",\n> and copying (not moving) the same Copyright block to the new file.\n>\n> But 835-552 = 283 feels way way more than that. What else is going on?\n\nHum, you're right. I checked with blame -C -C that I introduced no new\nlines, but I must accidentally have duplicated parts of it during\nconflict resolution.\n\n-- \nThomas Rast\ntrast@{inf,student}.ethz.ch\n"},{"id":"184871","messageId":"cover.1329472405.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"7v7gzmxw78.fsf@alter.siamese.dyndns.org","subject":"[PATCH v3 0/3]","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-17T10:25:07Z","receivedAt":"2012-02-17T10:25:07Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"Junio C Hamano wrote:\n> > ---\n> > t/test-lib-functions.sh | 835 +++++++++++++++++++++++++++++++++++++++++++++++\n> > t/test-lib.sh | 552 +-------------------------------\n> > 2 files changed, 840 insertions(+), 547 deletions(-)\n> > create mode 100644 t/test-lib-functions.sh\n> \n> I would have expected from the log description that the number of deleted\n> lines would be about the same as the number of added lines, and the\n> difference would primarily come from the addition of \"include\" aka \"dot\"\n> \". ./test-lib-functions.sh\" that becomes necessary in t/test-lib.sh, some\n> boilerplate material at the beginning of the new file e.g. \"#!/bin/sh\",\n> and copying (not moving) the same Copyright block to the new file.\n\nThere were actually more mistakes lurking :-( so I am resending the\nwhole series. I also put in the copyright that you asked for. I\nverified the results by looking at the diff between a reverse git-show\nfor test-lib.sh and a forward git-show for test-lib-functions.sh,\nwhich looks as follows:\n\n --- /dev/fd/63\t2012-02-17 10:55:32.994197654 +0100\n +++ /dev/fd/62\t2012-02-17 10:55:32.994197654 +0100\n @@ -9,17 +9,29 @@\n \n Signed-off-by: Thomas Rast \n \n -diff --git b/t/test-lib.sh a/t/test-lib.sh\n -index 1da3f40..e28d5fd 100644\n ---- b/t/test-lib.sh\n -+++ a/t/test-lib.sh\n -@@ -223,9 +223,248 @@ die () {\n - GIT_EXIT_OK=\n - trap 'die' EXIT\n - \n --# The user-facing functions are loaded from a separate file so that\n --# test_perf subshells can have them too\n --. \"${TEST_DIRECTORY:-.}\"/test-lib-functions.sh\n +diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh\n +new file mode 100644\n +index 0000000..7b3b4be\n +--- /dev/null\n ++++ b/t/test-lib-functions.sh\n +@@ -0,0 +1,565 @@\n ++#!/bin/sh\n ++#\n ++# Copyright (c) 2005 Junio C Hamano\n ++#\n ++# This program is free software: you can redistribute it and/or modify\n ++# it under the terms of the GNU General Public License as published by\n ++# the Free Software Foundation, either version 2 of the License, or\n ++# (at your option) any later version.\n ++#\n ++# This program is distributed in the hope that it will be useful,\n ++# but WITHOUT ANY WARRANTY; without even the implied warranty of\n ++# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n ++# GNU General Public License for more details.\n ++#\n ++# You should have received a copy of the GNU General Public License\n ++# along with this program. If not, see http://www.gnu.org/licenses/ .\n ++\n +# The semantics of the editor variables are that of invoking\n +# sh -c \"$EDITOR \\\"$@\\\"\" files ...\n +#\n @@ -192,7 +204,6 @@\n +\tgit config \"$@\"\n +}\n +\n -+\n +test_config_global () {\n +\ttest_when_finished \"test_unconfig --global '$1'\" &&\n +\tgit config --global \"$@\"\n @@ -262,13 +273,7 @@\n +\tesac\n +\treturn 1\n +}\n - \n - # You are not expected to call test_ok_ and test_failure_ directly, use\n - # the text_expect_* functions instead.\n -@@ -313,6 +552,313 @@ test_skip () {\n - \tesac\n - }\n - \n ++\n +test_expect_failure () {\n +\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n +\ttest \"$#\" = 2 ||\n @@ -575,7 +580,3 @@\n +\t\tmv .git/hooks .git/hooks-disabled\n +\t) || exit\n +}\n -+\n - test_done () {\n - \tGIT_EXIT_OK=t\n - \n\n\n\nThomas Rast (3):\n Move the user-facing test library to test-lib-functions.sh\n Introduce a performance testing framework\n Add a performance test for git-grep\n\n Makefile | 22 +-\n t/Makefile | 43 ++-\n t/perf/.gitignore | 2 +\n t/perf/Makefile | 15 +\n t/perf/README | 146 ++++++++++\n t/perf/aggregate.perl | 166 +++++++++++\n t/perf/min_time.perl | 21 ++\n t/perf/p0000-perf-lib-sanity.sh | 41 +++\n t/perf/p0001-rev-list.sh | 17 ++\n t/perf/p7810-grep.sh | 23 ++\n t/perf/perf-lib.sh | 198 ++++++++++++++\n t/perf/run | 82 ++++++\n t/test-lib-functions.sh | 565 ++++++++++++++++++++++++++++++++++++++\n t/test-lib.sh | 574 ++-------------------------------------\n 14 files changed, 1363 insertions(+), 552 deletions(-)\n create mode 100644 t/perf/.gitignore\n create mode 100644 t/perf/Makefile\n create mode 100644 t/perf/README\n create mode 100755 t/perf/aggregate.perl\n create mode 100755 t/perf/min_time.perl\n create mode 100755 t/perf/p0000-perf-lib-sanity.sh\n create mode 100755 t/perf/p0001-rev-list.sh\n create mode 100755 t/perf/p7810-grep.sh\n create mode 100644 t/perf/perf-lib.sh\n create mode 100755 t/perf/run\n create mode 100644 t/test-lib-functions.sh\n\n-- \n1.7.9.1.365.ge223f\n"},{"id":"184872","messageId":"0f0f88c03b879ad9f952e8168750974723073aba.1329472405.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329472405.git.trast@student.ethz.ch","subject":"[PATCH v3 1/3] Move the user-facing test library to test-lib-functions.sh","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-17T10:25:08Z","receivedAt":"2012-02-17T10:25:08Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"This just moves all the user-facing functions to a separate file and\nsources that instead.\n\nSigned-off-by: Thomas Rast \n---\n t/test-lib-functions.sh | 565 +++++++++++++++++++++++++++++++++++++++++++++++\n t/test-lib.sh | 552 +--------------------------------------------\n 2 files changed, 568 insertions(+), 549 deletions(-)\n create mode 100644 t/test-lib-functions.sh\n\ndiff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh\nnew file mode 100644\nindex 0000000..7b3b4be\n--- /dev/null\n+++ b/t/test-lib-functions.sh\n@@ -0,0 +1,565 @@\n+#!/bin/sh\n+#\n+# Copyright (c) 2005 Junio C Hamano\n+#\n+# This program is free software: you can redistribute it and/or modify\n+# it under the terms of the GNU General Public License as published by\n+# the Free Software Foundation, either version 2 of the License, or\n+# (at your option) any later version.\n+#\n+# This program is distributed in the hope that it will be useful,\n+# but WITHOUT ANY WARRANTY; without even the implied warranty of\n+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n+# GNU General Public License for more details.\n+#\n+# You should have received a copy of the GNU General Public License\n+# along with this program. If not, see http://www.gnu.org/licenses/ .\n+\n+# The semantics of the editor variables are that of invoking\n+# sh -c \"$EDITOR \\\"$@\\\"\" files ...\n+#\n+# If our trash directory contains shell metacharacters, they will be\n+# interpreted if we just set $EDITOR directly, so do a little dance with\n+# environment variables to work around this.\n+#\n+# In particular, quoting isn't enough, as the path may contain the same quote\n+# that we're using.\n+test_set_editor () {\n+\tFAKE_EDITOR=\"$1\"\n+\texport FAKE_EDITOR\n+\tEDITOR='\"$FAKE_EDITOR\"'\n+\texport EDITOR\n+}\n+\n+test_decode_color () {\n+\tawk '\n+\t\tfunction name(n) {\n+\t\t\tif (n == 0) return \"RESET\";\n+\t\t\tif (n == 1) return \"BOLD\";\n+\t\t\tif (n == 30) return \"BLACK\";\n+\t\t\tif (n == 31) return \"RED\";\n+\t\t\tif (n == 32) return \"GREEN\";\n+\t\t\tif (n == 33) return \"YELLOW\";\n+\t\t\tif (n == 34) return \"BLUE\";\n+\t\t\tif (n == 35) return \"MAGENTA\";\n+\t\t\tif (n == 36) return \"CYAN\";\n+\t\t\tif (n == 37) return \"WHITE\";\n+\t\t\tif (n == 40) return \"BLACK\";\n+\t\t\tif (n == 41) return \"BRED\";\n+\t\t\tif (n == 42) return \"BGREEN\";\n+\t\t\tif (n == 43) return \"BYELLOW\";\n+\t\t\tif (n == 44) return \"BBLUE\";\n+\t\t\tif (n == 45) return \"BMAGENTA\";\n+\t\t\tif (n == 46) return \"BCYAN\";\n+\t\t\tif (n == 47) return \"BWHITE\";\n+\t\t}\n+\t\t{\n+\t\t\twhile (match($0, /\\033\\[[0-9;]*m/) != 0) {\n+\t\t\t\tprintf \"%s<\", substr($0, 1, RSTART-1);\n+\t\t\t\tcodes = substr($0, RSTART+2, RLENGTH-3);\n+\t\t\t\tif (length(codes) == 0)\n+\t\t\t\t\tprintf \"%s\", name(0)\n+\t\t\t\telse {\n+\t\t\t\t\tn = split(codes, ary, \";\");\n+\t\t\t\t\tsep = \"\";\n+\t\t\t\t\tfor (i = 1; i <= n; i++) {\n+\t\t\t\t\t\tprintf \"%s%s\", sep, name(ary[i]);\n+\t\t\t\t\t\tsep = \";\"\n+\t\t\t\t\t}\n+\t\t\t\t}\n+\t\t\t\tprintf \">\";\n+\t\t\t\t$0 = substr($0, RSTART + RLENGTH, length($0) - RSTART - RLENGTH + 1);\n+\t\t\t}\n+\t\t\tprint\n+\t\t}\n+\t'\n+}\n+\n+nul_to_q () {\n+\tperl -pe 'y/\\000/Q/'\n+}\n+\n+q_to_nul () {\n+\tperl -pe 'y/Q/\\000/'\n+}\n+\n+q_to_cr () {\n+\ttr Q '\\015'\n+}\n+\n+q_to_tab () {\n+\ttr Q '\\011'\n+}\n+\n+append_cr () {\n+\tsed -e 's/$/Q/' | tr Q '\\015'\n+}\n+\n+remove_cr () {\n+\ttr '\\015' Q | sed -e 's/Q$//'\n+}\n+\n+# In some bourne shell implementations, the \"unset\" builtin returns\n+# nonzero status when a variable to be unset was not set in the first\n+# place.\n+#\n+# Use sane_unset when that should not be considered an error.\n+\n+sane_unset () {\n+\tunset \"$@\"\n+\treturn 0\n+}\n+\n+test_tick () {\n+\tif test -z \"${test_tick+set}\"\n+\tthen\n+\t\ttest_tick=1112911993\n+\telse\n+\t\ttest_tick=$(($test_tick + 60))\n+\tfi\n+\tGIT_COMMITTER_DATE=\"$test_tick -0700\"\n+\tGIT_AUTHOR_DATE=\"$test_tick -0700\"\n+\texport GIT_COMMITTER_DATE GIT_AUTHOR_DATE\n+}\n+\n+# Stop execution and start a shell. This is useful for debugging tests and\n+# only makes sense together with \"-v\".\n+#\n+# Be sure to remove all invocations of this command before submitting.\n+\n+test_pause () {\n+\tif test \"$verbose\" = t; then\n+\t\t\"$SHELL_PATH\" <&6 >&3 2>&4\n+\telse\n+\t\terror >&5 \"test_pause requires --verbose\"\n+\tfi\n+}\n+\n+# Call test_commit with the arguments \" [ []]\"\n+#\n+# This will commit a file with the given contents and the given commit\n+# message. It will also add a tag with as name.\n+#\n+# Both and default to .\n+\n+test_commit () {\n+\tfile=${2:-\"$1.t\"}\n+\techo \"${3-$1}\" > \"$file\" &&\n+\tgit add \"$file\" &&\n+\ttest_tick &&\n+\tgit commit -m \"$1\" &&\n+\tgit tag \"$1\"\n+}\n+\n+# Call test_merge with the arguments \" \", where \n+# can be a tag pointing to the commit-to-merge.\n+\n+test_merge () {\n+\ttest_tick &&\n+\tgit merge -m \"$1\" \"$2\" &&\n+\tgit tag \"$1\"\n+}\n+\n+# This function helps systems where core.filemode=false is set.\n+# Use it instead of plain 'chmod +x' to set or unset the executable bit\n+# of a file in the working directory and add it to the index.\n+\n+test_chmod () {\n+\tchmod \"$@\" &&\n+\tgit update-index --add \"--chmod=$@\"\n+}\n+\n+# Unset a configuration variable, but don't fail if it doesn't exist.\n+test_unconfig () {\n+\tgit config --unset-all \"$@\"\n+\tconfig_status=$?\n+\tcase \"$config_status\" in\n+\t5) # ok, nothing to unset\n+\t\tconfig_status=0\n+\t\t;;\n+\tesac\n+\treturn $config_status\n+}\n+\n+# Set git config, automatically unsetting it after the test is over.\n+test_config () {\n+\ttest_when_finished \"test_unconfig '$1'\" &&\n+\tgit config \"$@\"\n+}\n+\n+test_config_global () {\n+\ttest_when_finished \"test_unconfig --global '$1'\" &&\n+\tgit config --global \"$@\"\n+}\n+\n+write_script () {\n+\t{\n+\t\techo \"#!${2-\"$SHELL_PATH\"}\" &&\n+\t\tcat\n+\t} >\"$1\" &&\n+\tchmod +x \"$1\"\n+}\n+\n+# Use test_set_prereq to tell that a particular prerequisite is available.\n+# The prerequisite can later be checked for in two ways:\n+#\n+# - Explicitly using test_have_prereq.\n+#\n+# - Implicitly by specifying the prerequisite tag in the calls to\n+# test_expect_{success,failure,code}.\n+#\n+# The single parameter is the prerequisite tag (a simple word, in all\n+# capital letters by convention).\n+\n+test_set_prereq () {\n+\tsatisfied=\"$satisfied$1 \"\n+}\n+satisfied=\" \"\n+\n+test_have_prereq () {\n+\t# prerequisites can be concatenated with ','\n+\tsave_IFS=$IFS\n+\tIFS=,\n+\tset -- $*\n+\tIFS=$save_IFS\n+\n+\ttotal_prereq=0\n+\tok_prereq=0\n+\tmissing_prereq=\n+\n+\tfor prerequisite\n+\tdo\n+\t\ttotal_prereq=$(($total_prereq + 1))\n+\t\tcase $satisfied in\n+\t\t*\" $prerequisite \"*)\n+\t\t\tok_prereq=$(($ok_prereq + 1))\n+\t\t\t;;\n+\t\t*)\n+\t\t\t# Keep a list of missing prerequisites\n+\t\t\tif test -z \"$missing_prereq\"\n+\t\t\tthen\n+\t\t\t\tmissing_prereq=$prerequisite\n+\t\t\telse\n+\t\t\t\tmissing_prereq=\"$prerequisite,$missing_prereq\"\n+\t\t\tfi\n+\t\tesac\n+\tdone\n+\n+\ttest $total_prereq = $ok_prereq\n+}\n+\n+test_declared_prereq () {\n+\tcase \",$test_prereq,\" in\n+\t*,$1,*)\n+\t\treturn 0\n+\t\t;;\n+\tesac\n+\treturn 1\n+}\n+\n+test_expect_failure () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-failure\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"checking known breakage: $2\"\n+\t\tif test_run_ \"$2\" expecting_failure\n+\t\tthen\n+\t\t\ttest_known_broken_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_known_broken_failure_ \"$1\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+test_expect_success () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tsay >&3 \"expecting success: $2\"\n+\t\tif test_run_ \"$2\"\n+\t\tthen\n+\t\t\ttest_ok_ \"$1\"\n+\t\telse\n+\t\t\ttest_failure_ \"$@\"\n+\t\tfi\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+# test_external runs external test scripts that provide continuous\n+# test output about their progress, and succeeds/fails on\n+# zero/non-zero exit code. It outputs the test output on stdout even\n+# in non-verbose mode, and announces the external script with \"# run\n+# : ...\" before running it. When providing relative paths, keep in\n+# mind that all scripts run in \"trash directory\".\n+# Usage: test_external description command arguments...\n+# Example: test_external 'Perl API' perl ../path/to/test.pl\n+test_external () {\n+\ttest \"$#\" = 4 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 3 ||\n+\terror >&5 \"bug in the test script: not 3 or 4 parameters to test_external\"\n+\tdescr=\"$1\"\n+\tshift\n+\texport test_prereq\n+\tif ! test_skip \"$descr\" \"$@\"\n+\tthen\n+\t\t# Announce the script to reduce confusion about the\n+\t\t# test output that follows.\n+\t\tsay_color \"\" \"# run $test_count: $descr ($*)\"\n+\t\t# Export TEST_DIRECTORY, TRASH_DIRECTORY and GIT_TEST_LONG\n+\t\t# to be able to use them in script\n+\t\texport TEST_DIRECTORY TRASH_DIRECTORY GIT_TEST_LONG\n+\t\t# Run command; redirect its stderr to &4 as in\n+\t\t# test_run_, but keep its stdout on our stdout even in\n+\t\t# non-verbose mode.\n+\t\t\"$@\" 2>&4\n+\t\tif [ \"$?\" = 0 ]\n+\t\tthen\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_ok_ \"$descr\"\n+\t\t\telse\n+\t\t\t\tsay_color \"\" \"# test_external test $descr was ok\"\n+\t\t\t\ttest_success=$(($test_success + 1))\n+\t\t\tfi\n+\t\telse\n+\t\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\t\ttest_failure_ \"$descr\" \"$@\"\n+\t\t\telse\n+\t\t\t\tsay_color error \"# test_external test $descr failed: $@\"\n+\t\t\t\ttest_failure=$(($test_failure + 1))\n+\t\t\tfi\n+\t\tfi\n+\tfi\n+}\n+\n+# Like test_external, but in addition tests that the command generated\n+# no output on stderr.\n+test_external_without_stderr () {\n+\t# The temporary file has no (and must have no) security\n+\t# implications.\n+\ttmp=${TMPDIR:-/tmp}\n+\tstderr=\"$tmp/git-external-stderr.$$.tmp\"\n+\ttest_external \"$@\" 4> \"$stderr\"\n+\t[ -f \"$stderr\" ] || error \"Internal error: $stderr disappeared.\"\n+\tdescr=\"no stderr: $1\"\n+\tshift\n+\tsay >&3 \"# expecting no stderr from previous command\"\n+\tif [ ! -s \"$stderr\" ]; then\n+\t\trm \"$stderr\"\n+\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_ok_ \"$descr\"\n+\t\telse\n+\t\t\tsay_color \"\" \"# test_external_without_stderr test $descr was ok\"\n+\t\t\ttest_success=$(($test_success + 1))\n+\t\tfi\n+\telse\n+\t\tif [ \"$verbose\" = t ]; then\n+\t\t\toutput=`echo; echo \"# Stderr is:\"; cat \"$stderr\"`\n+\t\telse\n+\t\t\toutput=\n+\t\tfi\n+\t\t# rm first in case test_failure exits.\n+\t\trm \"$stderr\"\n+\t\tif test $test_external_has_tap -eq 0; then\n+\t\t\ttest_failure_ \"$descr\" \"$@\" \"$output\"\n+\t\telse\n+\t\t\tsay_color error \"# test_external_without_stderr test $descr failed: $@: $output\"\n+\t\t\ttest_failure=$(($test_failure + 1))\n+\t\tfi\n+\tfi\n+}\n+\n+# debugging-friendly alternatives to \"test [-f|-d|-e]\"\n+# The commands test the existence or non-existence of $1. $2 can be\n+# given to provide a more precise diagnosis.\n+test_path_is_file () {\n+\tif ! [ -f \"$1\" ]\n+\tthen\n+\t\techo \"File $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_dir () {\n+\tif ! [ -d \"$1\" ]\n+\tthen\n+\t\techo \"Directory $1 doesn't exist. $*\"\n+\t\tfalse\n+\tfi\n+}\n+\n+test_path_is_missing () {\n+\tif [ -e \"$1\" ]\n+\tthen\n+\t\techo \"Path exists:\"\n+\t\tls -ld \"$1\"\n+\t\tif [ $# -ge 1 ]; then\n+\t\t\techo \"$*\"\n+\t\tfi\n+\t\tfalse\n+\tfi\n+}\n+\n+# test_line_count checks that a file has the number of lines it\n+# ought to. For example:\n+#\n+#\ttest_expect_success 'produce exactly one line of output' '\n+#\t\tdo something >output &&\n+#\t\ttest_line_count = 1 output\n+#\t'\n+#\n+# is like \"test $(wc -l &2 \"test_must_fail: command succeeded: $*\"\n+\t\treturn 1\n+\telif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_must_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_must_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail, but tolerates success, too. This is\n+# meant to be used in contexts like:\n+#\n+#\ttest_expect_success 'some command works without configuration' '\n+#\t\ttest_might_fail git config --unset all.configuration &&\n+#\t\tdo something\n+#\t'\n+#\n+# Writing \"git config --unset all.configuration || :\" would be wrong,\n+# because we want to notice if it fails due to segv.\n+\n+test_might_fail () {\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code -gt 129 -a $exit_code -le 192; then\n+\t\techo >&2 \"test_might_fail: died by signal: $*\"\n+\t\treturn 1\n+\telif test $exit_code = 127; then\n+\t\techo >&2 \"test_might_fail: command not found: $*\"\n+\t\treturn 1\n+\tfi\n+\treturn 0\n+}\n+\n+# Similar to test_must_fail and test_might_fail, but check that a\n+# given command exited with a given exit code. Meant to be used as:\n+#\n+#\ttest_expect_success 'Merge with d/f conflicts' '\n+#\t\ttest_expect_code 1 git merge \"merge msg\" B master\n+#\t'\n+\n+test_expect_code () {\n+\twant_code=$1\n+\tshift\n+\t\"$@\"\n+\texit_code=$?\n+\tif test $exit_code = $want_code\n+\tthen\n+\t\treturn 0\n+\tfi\n+\n+\techo >&2 \"test_expect_code: command exited with $exit_code, we wanted $want_code $*\"\n+\treturn 1\n+}\n+\n+# test_cmp is a helper function to compare actual and expected output.\n+# You can use it like:\n+#\n+#\ttest_expect_success 'foo works' '\n+#\t\techo expected >expected &&\n+#\t\tfoo >actual &&\n+#\t\ttest_cmp expected actual\n+#\t'\n+#\n+# This could be written as either \"cmp\" or \"diff -u\", but:\n+# - cmp's output is not nearly as easy to read as diff -u\n+# - not all diff versions understand \"-u\"\n+\n+test_cmp() {\n+\t$GIT_TEST_CMP \"$@\"\n+}\n+\n+# This function can be used to schedule some commands to be run\n+# unconditionally at the end of the test to restore sanity:\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\ttest_when_finished \"git config --unset core.capslock\" &&\n+#\t\thello world\n+#\t'\n+#\n+# That would be roughly equivalent to\n+#\n+#\ttest_expect_success 'test core.capslock' '\n+#\t\tgit config core.capslock true &&\n+#\t\thello world\n+#\t\tgit config --unset core.capslock\n+#\t'\n+#\n+# except that the greeting and config --unset must both succeed for\n+# the test to pass.\n+#\n+# Note that under --immediate mode, no clean-up is done to help diagnose\n+# what went wrong.\n+\n+test_when_finished () {\n+\ttest_cleanup=\"{ $*\n+\t\t} && (exit \\\"\\$eval_ret\\\"); eval_ret=\\$?; $test_cleanup\"\n+}\n+\n+# Most tests can use the created repository, but some may need to create more.\n+# Usage: test_create_repo \n+test_create_repo () {\n+\ttest \"$#\" = 1 ||\n+\terror \"bug in the test script: not 1 parameter to test-create-repo\"\n+\trepo=\"$1\"\n+\tmkdir -p \"$repo\"\n+\t(\n+\t\tcd \"$repo\" || error \"Cannot setup test environment\"\n+\t\t\"$GIT_EXEC_PATH/git-init\" \"--template=$GIT_BUILD_DIR/templates/blt/\" >&3 2>&4 ||\n+\t\terror \"cannot run git init -- have you built things yet?\"\n+\t\tmv .git/hooks .git/hooks-disabled\n+\t) || exit\n+}\ndiff --git a/t/test-lib.sh b/t/test-lib.sh\nindex e28d5fd..1da3f40 100644\n--- a/t/test-lib.sh\n+++ b/t/test-lib.sh\n@@ -223,248 +223,9 @@ die () {\n GIT_EXIT_OK=\n trap 'die' EXIT\n \n-# The semantics of the editor variables are that of invoking\n-# sh -c \"$EDITOR \\\"$@\\\"\" files ...\n-#\n-# If our trash directory contains shell metacharacters, they will be\n-# interpreted if we just set $EDITOR directly, so do a little dance with\n-# environment variables to work around this.\n-#\n-# In particular, quoting isn't enough, as the path may contain the same quote\n-# that we're using.\n-test_set_editor () {\n-\tFAKE_EDITOR=\"$1\"\n-\texport FAKE_EDITOR\n-\tEDITOR='\"$FAKE_EDITOR\"'\n-\texport EDITOR\n-}\n-\n-test_decode_color () {\n-\tawk '\n-\t\tfunction name(n) {\n-\t\t\tif (n == 0) return \"RESET\";\n-\t\t\tif (n == 1) return \"BOLD\";\n-\t\t\tif (n == 30) return \"BLACK\";\n-\t\t\tif (n == 31) return \"RED\";\n-\t\t\tif (n == 32) return \"GREEN\";\n-\t\t\tif (n == 33) return \"YELLOW\";\n-\t\t\tif (n == 34) return \"BLUE\";\n-\t\t\tif (n == 35) return \"MAGENTA\";\n-\t\t\tif (n == 36) return \"CYAN\";\n-\t\t\tif (n == 37) return \"WHITE\";\n-\t\t\tif (n == 40) return \"BLACK\";\n-\t\t\tif (n == 41) return \"BRED\";\n-\t\t\tif (n == 42) return \"BGREEN\";\n-\t\t\tif (n == 43) return \"BYELLOW\";\n-\t\t\tif (n == 44) return \"BBLUE\";\n-\t\t\tif (n == 45) return \"BMAGENTA\";\n-\t\t\tif (n == 46) return \"BCYAN\";\n-\t\t\tif (n == 47) return \"BWHITE\";\n-\t\t}\n-\t\t{\n-\t\t\twhile (match($0, /\\033\\[[0-9;]*m/) != 0) {\n-\t\t\t\tprintf \"%s<\", substr($0, 1, RSTART-1);\n-\t\t\t\tcodes = substr($0, RSTART+2, RLENGTH-3);\n-\t\t\t\tif (length(codes) == 0)\n-\t\t\t\t\tprintf \"%s\", name(0)\n-\t\t\t\telse {\n-\t\t\t\t\tn = split(codes, ary, \";\");\n-\t\t\t\t\tsep = \"\";\n-\t\t\t\t\tfor (i = 1; i <= n; i++) {\n-\t\t\t\t\t\tprintf \"%s%s\", sep, name(ary[i]);\n-\t\t\t\t\t\tsep = \";\"\n-\t\t\t\t\t}\n-\t\t\t\t}\n-\t\t\t\tprintf \">\";\n-\t\t\t\t$0 = substr($0, RSTART + RLENGTH, length($0) - RSTART - RLENGTH + 1);\n-\t\t\t}\n-\t\t\tprint\n-\t\t}\n-\t'\n-}\n-\n-nul_to_q () {\n-\tperl -pe 'y/\\000/Q/'\n-}\n-\n-q_to_nul () {\n-\tperl -pe 'y/Q/\\000/'\n-}\n-\n-q_to_cr () {\n-\ttr Q '\\015'\n-}\n-\n-q_to_tab () {\n-\ttr Q '\\011'\n-}\n-\n-append_cr () {\n-\tsed -e 's/$/Q/' | tr Q '\\015'\n-}\n-\n-remove_cr () {\n-\ttr '\\015' Q | sed -e 's/Q$//'\n-}\n-\n-# In some bourne shell implementations, the \"unset\" builtin returns\n-# nonzero status when a variable to be unset was not set in the first\n-# place.\n-#\n-# Use sane_unset when that should not be considered an error.\n-\n-sane_unset () {\n-\tunset \"$@\"\n-\treturn 0\n-}\n-\n-test_tick () {\n-\tif test -z \"${test_tick+set}\"\n-\tthen\n-\t\ttest_tick=1112911993\n-\telse\n-\t\ttest_tick=$(($test_tick + 60))\n-\tfi\n-\tGIT_COMMITTER_DATE=\"$test_tick -0700\"\n-\tGIT_AUTHOR_DATE=\"$test_tick -0700\"\n-\texport GIT_COMMITTER_DATE GIT_AUTHOR_DATE\n-}\n-\n-# Stop execution and start a shell. This is useful for debugging tests and\n-# only makes sense together with \"-v\".\n-#\n-# Be sure to remove all invocations of this command before submitting.\n-\n-test_pause () {\n-\tif test \"$verbose\" = t; then\n-\t\t\"$SHELL_PATH\" <&6 >&3 2>&4\n-\telse\n-\t\terror >&5 \"test_pause requires --verbose\"\n-\tfi\n-}\n-\n-# Call test_commit with the arguments \" [ []]\"\n-#\n-# This will commit a file with the given contents and the given commit\n-# message. It will also add a tag with as name.\n-#\n-# Both and default to .\n-\n-test_commit () {\n-\tfile=${2:-\"$1.t\"}\n-\techo \"${3-$1}\" > \"$file\" &&\n-\tgit add \"$file\" &&\n-\ttest_tick &&\n-\tgit commit -m \"$1\" &&\n-\tgit tag \"$1\"\n-}\n-\n-# Call test_merge with the arguments \" \", where \n-# can be a tag pointing to the commit-to-merge.\n-\n-test_merge () {\n-\ttest_tick &&\n-\tgit merge -m \"$1\" \"$2\" &&\n-\tgit tag \"$1\"\n-}\n-\n-# This function helps systems where core.filemode=false is set.\n-# Use it instead of plain 'chmod +x' to set or unset the executable bit\n-# of a file in the working directory and add it to the index.\n-\n-test_chmod () {\n-\tchmod \"$@\" &&\n-\tgit update-index --add \"--chmod=$@\"\n-}\n-\n-# Unset a configuration variable, but don't fail if it doesn't exist.\n-test_unconfig () {\n-\tgit config --unset-all \"$@\"\n-\tconfig_status=$?\n-\tcase \"$config_status\" in\n-\t5) # ok, nothing to unset\n-\t\tconfig_status=0\n-\t\t;;\n-\tesac\n-\treturn $config_status\n-}\n-\n-# Set git config, automatically unsetting it after the test is over.\n-test_config () {\n-\ttest_when_finished \"test_unconfig '$1'\" &&\n-\tgit config \"$@\"\n-}\n-\n-\n-test_config_global () {\n-\ttest_when_finished \"test_unconfig --global '$1'\" &&\n-\tgit config --global \"$@\"\n-}\n-\n-write_script () {\n-\t{\n-\t\techo \"#!${2-\"$SHELL_PATH\"}\" &&\n-\t\tcat\n-\t} >\"$1\" &&\n-\tchmod +x \"$1\"\n-}\n-\n-# Use test_set_prereq to tell that a particular prerequisite is available.\n-# The prerequisite can later be checked for in two ways:\n-#\n-# - Explicitly using test_have_prereq.\n-#\n-# - Implicitly by specifying the prerequisite tag in the calls to\n-# test_expect_{success,failure,code}.\n-#\n-# The single parameter is the prerequisite tag (a simple word, in all\n-# capital letters by convention).\n-\n-test_set_prereq () {\n-\tsatisfied=\"$satisfied$1 \"\n-}\n-satisfied=\" \"\n-\n-test_have_prereq () {\n-\t# prerequisites can be concatenated with ','\n-\tsave_IFS=$IFS\n-\tIFS=,\n-\tset -- $*\n-\tIFS=$save_IFS\n-\n-\ttotal_prereq=0\n-\tok_prereq=0\n-\tmissing_prereq=\n-\n-\tfor prerequisite\n-\tdo\n-\t\ttotal_prereq=$(($total_prereq + 1))\n-\t\tcase $satisfied in\n-\t\t*\" $prerequisite \"*)\n-\t\t\tok_prereq=$(($ok_prereq + 1))\n-\t\t\t;;\n-\t\t*)\n-\t\t\t# Keep a list of missing prerequisites\n-\t\t\tif test -z \"$missing_prereq\"\n-\t\t\tthen\n-\t\t\t\tmissing_prereq=$prerequisite\n-\t\t\telse\n-\t\t\t\tmissing_prereq=\"$prerequisite,$missing_prereq\"\n-\t\t\tfi\n-\t\tesac\n-\tdone\n-\n-\ttest $total_prereq = $ok_prereq\n-}\n-\n-test_declared_prereq () {\n-\tcase \",$test_prereq,\" in\n-\t*,$1,*)\n-\t\treturn 0\n-\t\t;;\n-\tesac\n-\treturn 1\n-}\n+# The user-facing functions are loaded from a separate file so that\n+# test_perf subshells can have them too\n+. \"${TEST_DIRECTORY:-.}\"/test-lib-functions.sh\n \n # You are not expected to call test_ok_ and test_failure_ directly, use\n # the text_expect_* functions instead.\n@@ -552,313 +313,6 @@ test_skip () {\n \tesac\n }\n \n-test_expect_failure () {\n-\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 2 ||\n-\terror \"bug in the test script: not 2 or 3 parameters to test-expect-failure\"\n-\texport test_prereq\n-\tif ! test_skip \"$@\"\n-\tthen\n-\t\tsay >&3 \"checking known breakage: $2\"\n-\t\tif test_run_ \"$2\" expecting_failure\n-\t\tthen\n-\t\t\ttest_known_broken_ok_ \"$1\"\n-\t\telse\n-\t\t\ttest_known_broken_failure_ \"$1\"\n-\t\tfi\n-\tfi\n-\techo >&3 \"\"\n-}\n-\n-test_expect_success () {\n-\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 2 ||\n-\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n-\texport test_prereq\n-\tif ! test_skip \"$@\"\n-\tthen\n-\t\tsay >&3 \"expecting success: $2\"\n-\t\tif test_run_ \"$2\"\n-\t\tthen\n-\t\t\ttest_ok_ \"$1\"\n-\t\telse\n-\t\t\ttest_failure_ \"$@\"\n-\t\tfi\n-\tfi\n-\techo >&3 \"\"\n-}\n-\n-# test_external runs external test scripts that provide continuous\n-# test output about their progress, and succeeds/fails on\n-# zero/non-zero exit code. It outputs the test output on stdout even\n-# in non-verbose mode, and announces the external script with \"# run\n-# : ...\" before running it. When providing relative paths, keep in\n-# mind that all scripts run in \"trash directory\".\n-# Usage: test_external description command arguments...\n-# Example: test_external 'Perl API' perl ../path/to/test.pl\n-test_external () {\n-\ttest \"$#\" = 4 && { test_prereq=$1; shift; } || test_prereq=\n-\ttest \"$#\" = 3 ||\n-\terror >&5 \"bug in the test script: not 3 or 4 parameters to test_external\"\n-\tdescr=\"$1\"\n-\tshift\n-\texport test_prereq\n-\tif ! test_skip \"$descr\" \"$@\"\n-\tthen\n-\t\t# Announce the script to reduce confusion about the\n-\t\t# test output that follows.\n-\t\tsay_color \"\" \"# run $test_count: $descr ($*)\"\n-\t\t# Export TEST_DIRECTORY, TRASH_DIRECTORY and GIT_TEST_LONG\n-\t\t# to be able to use them in script\n-\t\texport TEST_DIRECTORY TRASH_DIRECTORY GIT_TEST_LONG\n-\t\t# Run command; redirect its stderr to &4 as in\n-\t\t# test_run_, but keep its stdout on our stdout even in\n-\t\t# non-verbose mode.\n-\t\t\"$@\" 2>&4\n-\t\tif [ \"$?\" = 0 ]\n-\t\tthen\n-\t\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\t\ttest_ok_ \"$descr\"\n-\t\t\telse\n-\t\t\t\tsay_color \"\" \"# test_external test $descr was ok\"\n-\t\t\t\ttest_success=$(($test_success + 1))\n-\t\t\tfi\n-\t\telse\n-\t\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\t\ttest_failure_ \"$descr\" \"$@\"\n-\t\t\telse\n-\t\t\t\tsay_color error \"# test_external test $descr failed: $@\"\n-\t\t\t\ttest_failure=$(($test_failure + 1))\n-\t\t\tfi\n-\t\tfi\n-\tfi\n-}\n-\n-# Like test_external, but in addition tests that the command generated\n-# no output on stderr.\n-test_external_without_stderr () {\n-\t# The temporary file has no (and must have no) security\n-\t# implications.\n-\ttmp=${TMPDIR:-/tmp}\n-\tstderr=\"$tmp/git-external-stderr.$$.tmp\"\n-\ttest_external \"$@\" 4> \"$stderr\"\n-\t[ -f \"$stderr\" ] || error \"Internal error: $stderr disappeared.\"\n-\tdescr=\"no stderr: $1\"\n-\tshift\n-\tsay >&3 \"# expecting no stderr from previous command\"\n-\tif [ ! -s \"$stderr\" ]; then\n-\t\trm \"$stderr\"\n-\n-\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\ttest_ok_ \"$descr\"\n-\t\telse\n-\t\t\tsay_color \"\" \"# test_external_without_stderr test $descr was ok\"\n-\t\t\ttest_success=$(($test_success + 1))\n-\t\tfi\n-\telse\n-\t\tif [ \"$verbose\" = t ]; then\n-\t\t\toutput=`echo; echo \"# Stderr is:\"; cat \"$stderr\"`\n-\t\telse\n-\t\t\toutput=\n-\t\tfi\n-\t\t# rm first in case test_failure exits.\n-\t\trm \"$stderr\"\n-\t\tif test $test_external_has_tap -eq 0; then\n-\t\t\ttest_failure_ \"$descr\" \"$@\" \"$output\"\n-\t\telse\n-\t\t\tsay_color error \"# test_external_without_stderr test $descr failed: $@: $output\"\n-\t\t\ttest_failure=$(($test_failure + 1))\n-\t\tfi\n-\tfi\n-}\n-\n-# debugging-friendly alternatives to \"test [-f|-d|-e]\"\n-# The commands test the existence or non-existence of $1. $2 can be\n-# given to provide a more precise diagnosis.\n-test_path_is_file () {\n-\tif ! [ -f \"$1\" ]\n-\tthen\n-\t\techo \"File $1 doesn't exist. $*\"\n-\t\tfalse\n-\tfi\n-}\n-\n-test_path_is_dir () {\n-\tif ! [ -d \"$1\" ]\n-\tthen\n-\t\techo \"Directory $1 doesn't exist. $*\"\n-\t\tfalse\n-\tfi\n-}\n-\n-test_path_is_missing () {\n-\tif [ -e \"$1\" ]\n-\tthen\n-\t\techo \"Path exists:\"\n-\t\tls -ld \"$1\"\n-\t\tif [ $# -ge 1 ]; then\n-\t\t\techo \"$*\"\n-\t\tfi\n-\t\tfalse\n-\tfi\n-}\n-\n-# test_line_count checks that a file has the number of lines it\n-# ought to. For example:\n-#\n-#\ttest_expect_success 'produce exactly one line of output' '\n-#\t\tdo something >output &&\n-#\t\ttest_line_count = 1 output\n-#\t'\n-#\n-# is like \"test $(wc -l &2 \"test_must_fail: command succeeded: $*\"\n-\t\treturn 1\n-\telif test $exit_code -gt 129 -a $exit_code -le 192; then\n-\t\techo >&2 \"test_must_fail: died by signal: $*\"\n-\t\treturn 1\n-\telif test $exit_code = 127; then\n-\t\techo >&2 \"test_must_fail: command not found: $*\"\n-\t\treturn 1\n-\tfi\n-\treturn 0\n-}\n-\n-# Similar to test_must_fail, but tolerates success, too. This is\n-# meant to be used in contexts like:\n-#\n-#\ttest_expect_success 'some command works without configuration' '\n-#\t\ttest_might_fail git config --unset all.configuration &&\n-#\t\tdo something\n-#\t'\n-#\n-# Writing \"git config --unset all.configuration || :\" would be wrong,\n-# because we want to notice if it fails due to segv.\n-\n-test_might_fail () {\n-\t\"$@\"\n-\texit_code=$?\n-\tif test $exit_code -gt 129 -a $exit_code -le 192; then\n-\t\techo >&2 \"test_might_fail: died by signal: $*\"\n-\t\treturn 1\n-\telif test $exit_code = 127; then\n-\t\techo >&2 \"test_might_fail: command not found: $*\"\n-\t\treturn 1\n-\tfi\n-\treturn 0\n-}\n-\n-# Similar to test_must_fail and test_might_fail, but check that a\n-# given command exited with a given exit code. Meant to be used as:\n-#\n-#\ttest_expect_success 'Merge with d/f conflicts' '\n-#\t\ttest_expect_code 1 git merge \"merge msg\" B master\n-#\t'\n-\n-test_expect_code () {\n-\twant_code=$1\n-\tshift\n-\t\"$@\"\n-\texit_code=$?\n-\tif test $exit_code = $want_code\n-\tthen\n-\t\treturn 0\n-\tfi\n-\n-\techo >&2 \"test_expect_code: command exited with $exit_code, we wanted $want_code $*\"\n-\treturn 1\n-}\n-\n-# test_cmp is a helper function to compare actual and expected output.\n-# You can use it like:\n-#\n-#\ttest_expect_success 'foo works' '\n-#\t\techo expected >expected &&\n-#\t\tfoo >actual &&\n-#\t\ttest_cmp expected actual\n-#\t'\n-#\n-# This could be written as either \"cmp\" or \"diff -u\", but:\n-# - cmp's output is not nearly as easy to read as diff -u\n-# - not all diff versions understand \"-u\"\n-\n-test_cmp() {\n-\t$GIT_TEST_CMP \"$@\"\n-}\n-\n-# This function can be used to schedule some commands to be run\n-# unconditionally at the end of the test to restore sanity:\n-#\n-#\ttest_expect_success 'test core.capslock' '\n-#\t\tgit config core.capslock true &&\n-#\t\ttest_when_finished \"git config --unset core.capslock\" &&\n-#\t\thello world\n-#\t'\n-#\n-# That would be roughly equivalent to\n-#\n-#\ttest_expect_success 'test core.capslock' '\n-#\t\tgit config core.capslock true &&\n-#\t\thello world\n-#\t\tgit config --unset core.capslock\n-#\t'\n-#\n-# except that the greeting and config --unset must both succeed for\n-# the test to pass.\n-#\n-# Note that under --immediate mode, no clean-up is done to help diagnose\n-# what went wrong.\n-\n-test_when_finished () {\n-\ttest_cleanup=\"{ $*\n-\t\t} && (exit \\\"\\$eval_ret\\\"); eval_ret=\\$?; $test_cleanup\"\n-}\n-\n-# Most tests can use the created repository, but some may need to create more.\n-# Usage: test_create_repo \n-test_create_repo () {\n-\ttest \"$#\" = 1 ||\n-\terror \"bug in the test script: not 1 parameter to test-create-repo\"\n-\trepo=\"$1\"\n-\tmkdir -p \"$repo\"\n-\t(\n-\t\tcd \"$repo\" || error \"Cannot setup test environment\"\n-\t\t\"$GIT_EXEC_PATH/git-init\" \"--template=$GIT_BUILD_DIR/templates/blt/\" >&3 2>&4 ||\n-\t\terror \"cannot run git init -- have you built things yet?\"\n-\t\tmv .git/hooks .git/hooks-disabled\n-\t) || exit\n-}\n-\n test_done () {\n \tGIT_EXIT_OK=t\n \n-- \n1.7.9.1.365.ge223f\n"},{"id":"184874","messageId":"c862838275745584956bfd57c8c91f3048f8113c.1329472405.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329472405.git.trast@student.ethz.ch","subject":"[PATCH v3 2/3] Introduce a performance testing framework","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-17T10:25:09Z","receivedAt":"2012-02-17T10:25:09Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"This introduces a performance testing framework under t/perf/. It\ntries to be as close to the test-lib.sh infrastructure as possible,\nand thus should be easy to get used to for git developers.\n\nThe following points were considered for the implementation:\n\n1. You usually want to compare arbitrary revisions/build trees against\n each other. They may not have the performance test under\n consideration, or even the perf-lib.sh infrastructure.\n\n To cope with this, the 'run' script lets you specify arbitrary\n build dirs and revisions. It even automatically builds the revisions\n if it doesn't have them at hand yet.\n\n2. Usually you would not want to run all tests. It would take too\n long anyway. The 'run' script lets you specify which tests to run;\n or you can also do it manually. There is a Makefile for\n discoverability and 'make clean', but it is not meant for\n real-world use.\n\n3. Creating test repos from scratch in every test is extremely\n time-consuming, and shipping or downloading such large/weird repos\n is out of the question.\n\n We leave this decision to the user. Two different sizes of test\n repos can be configured, and the scripts just copy one or more of\n those (using hardlinks for the object store). By default it tries\n to use the build tree's git.git repository.\n\n This is fairly fast and versatile. Using a copy instead of a clone\n preserves many properties that the user may want to test for, such\n as lots of loose objects, unpacked refs, etc.\n\nSigned-off-by: Thomas Rast \n---\n Makefile | 22 ++++-\n t/Makefile | 43 ++++++++-\n t/perf/.gitignore | 2 +\n t/perf/Makefile | 15 +++\n t/perf/README | 146 +++++++++++++++++++++++++++++\n t/perf/aggregate.perl | 166 ++++++++++++++++++++++++++++++++\n t/perf/min_time.perl | 21 +++++\n t/perf/p0000-perf-lib-sanity.sh | 41 ++++++++\n t/perf/p0001-rev-list.sh | 17 ++++\n t/perf/perf-lib.sh | 198 +++++++++++++++++++++++++++++++++++++++\n t/perf/run | 82 ++++++++++++++++\n t/test-lib.sh | 26 ++++-\n 12 files changed, 774 insertions(+), 5 deletions(-)\n create mode 100644 t/perf/.gitignore\n create mode 100644 t/perf/Makefile\n create mode 100644 t/perf/README\n create mode 100755 t/perf/aggregate.perl\n create mode 100755 t/perf/min_time.perl\n create mode 100755 t/perf/p0000-perf-lib-sanity.sh\n create mode 100755 t/perf/p0001-rev-list.sh\n create mode 100644 t/perf/perf-lib.sh\n create mode 100755 t/perf/run\n\ndiff --git a/Makefile b/Makefile\nindex a0de4e9..1fb1705 100644\n--- a/Makefile\n+++ b/Makefile\n@@ -2361,6 +2361,10 @@ GIT-BUILD-OPTIONS: FORCE\n \t@echo USE_LIBPCRE=\\''$(subst ','\\'',$(subst ','\\'',$(USE_LIBPCRE)))'\\' >>$@\n \t@echo NO_PERL=\\''$(subst ','\\'',$(subst ','\\'',$(NO_PERL)))'\\' >>$@\n \t@echo NO_PYTHON=\\''$(subst ','\\'',$(subst ','\\'',$(NO_PYTHON)))'\\' >>$@\n+\t@echo NO_UNIX_SOCKETS=\\''$(subst ','\\'',$(subst ','\\'',$(NO_UNIX_SOCKETS)))'\\' >>$@\n+ifdef GIT_TEST_OPTS\n+\t@echo GIT_TEST_OPTS=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_TEST_OPTS)))'\\' >>$@\n+endif\n ifdef GIT_TEST_CMP\n \t@echo GIT_TEST_CMP=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_TEST_CMP)))'\\' >>$@\n endif\n@@ -2369,7 +2373,18 @@ ifdef GIT_TEST_CMP_USE_COPIED_CONTEXT\n endif\n \t@echo NO_GETTEXT=\\''$(subst ','\\'',$(subst ','\\'',$(NO_GETTEXT)))'\\' >>$@\n \t@echo GETTEXT_POISON=\\''$(subst ','\\'',$(subst ','\\'',$(GETTEXT_POISON)))'\\' >>$@\n-\t@echo NO_UNIX_SOCKETS=\\''$(subst ','\\'',$(subst ','\\'',$(NO_UNIX_SOCKETS)))'\\' >>$@\n+ifdef GIT_PERF_REPEAT_COUNT\n+\t@echo GIT_PERF_REPEAT_COUNT=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_REPEAT_COUNT)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_REPO\n+\t@echo GIT_PERF_REPO=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_REPO)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_LARGE_REPO\n+\t@echo GIT_PERF_LARGE_REPO=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_LARGE_REPO)))'\\' >>$@\n+endif\n+ifdef GIT_PERF_MAKE_OPTS\n+\t@echo GIT_PERF_MAKE_OPTS=\\''$(subst ','\\'',$(subst ','\\'',$(GIT_PERF_MAKE_OPTS)))'\\' >>$@\n+endif\n \n ### Detect Tck/Tk interpreter path changes\n ifndef NO_TCLTK\n@@ -2405,6 +2420,11 @@ export NO_SVN_TESTS\n test: all\n \t$(MAKE) -C t/ all\n \n+perf: all\n+\t$(MAKE) -C t/perf/ all\n+\n+.PHONY: test perf\n+\n test-ctype$X: ctype.o\n \n test-date$X: date.o ctype.o\ndiff --git a/t/Makefile b/t/Makefile\nindex b5048ab..6091211 100644\n--- a/t/Makefile\n+++ b/t/Makefile\n@@ -73,4 +73,45 @@ gitweb-test:\n valgrind:\n \t$(MAKE) GIT_TEST_OPTS=\"$(GIT_TEST_OPTS) --valgrind\"\n \n-.PHONY: pre-clean $(T) aggregate-results clean valgrind\n+perf:\n+\t$(MAKE) -C perf/ all\n+\n+# Smoke testing targets\n+-include ../GIT-VERSION-FILE\n+uname_S := $(shell sh -c 'uname -s 2>/dev/null || echo unknown')\n+uname_M := $(shell sh -c 'uname -m 2>/dev/null || echo unknown')\n+\n+test-results:\n+\tmkdir -p test-results\n+\n+test-results/git-smoke.tar.gz: test-results\n+\t$(PERL_PATH) ./harness \\\n+\t\t--archive=\"test-results/git-smoke.tar.gz\" \\\n+\t\t$(T)\n+\n+smoke: test-results/git-smoke.tar.gz\n+\n+SMOKE_UPLOAD_FLAGS =\n+ifdef SMOKE_USERNAME\n+\tSMOKE_UPLOAD_FLAGS += -F username=\"$(SMOKE_USERNAME)\" -F password=\"$(SMOKE_PASSWORD)\"\n+endif\n+ifdef SMOKE_COMMENT\n+\tSMOKE_UPLOAD_FLAGS += -F comments=\"$(SMOKE_COMMENT)\"\n+endif\n+ifdef SMOKE_TAGS\n+\tSMOKE_UPLOAD_FLAGS += -F tags=\"$(SMOKE_TAGS)\"\n+endif\n+\n+smoke_report: smoke\n+\tcurl \\\n+\t\t-H \"Expect: \" \\\n+\t\t-F project=Git \\\n+\t\t-F architecture=\"$(uname_M)\" \\\n+\t\t-F platform=\"$(uname_S)\" \\\n+\t\t-F revision=\"$(GIT_VERSION)\" \\\n+\t\t-F report_file=@test-results/git-smoke.tar.gz \\\n+\t\t$(SMOKE_UPLOAD_FLAGS) \\\n+\t\thttp://smoke.git.nix.is/app/projects/process_add_report/1 \\\n+\t| grep -v ^Redirecting\n+\n+.PHONY: pre-clean $(T) aggregate-results clean valgrind perf\ndiff --git a/t/perf/.gitignore b/t/perf/.gitignore\nnew file mode 100644\nindex 0000000..50f5cc1\n--- /dev/null\n+++ b/t/perf/.gitignore\n@@ -0,0 +1,2 @@\n+build/\n+test-results/\ndiff --git a/t/perf/Makefile b/t/perf/Makefile\nnew file mode 100644\nindex 0000000..8c47155\n--- /dev/null\n+++ b/t/perf/Makefile\n@@ -0,0 +1,15 @@\n+-include ../../config.mak\n+export GIT_TEST_OPTIONS\n+\n+all: perf\n+\n+perf: pre-clean\n+\t./run\n+\n+pre-clean:\n+\trm -rf test-results\n+\n+clean:\n+\trm -rf build \"trash directory\".* test-results\n+\n+.PHONY: all perf pre-clean clean\ndiff --git a/t/perf/README b/t/perf/README\nnew file mode 100644\nindex 0000000..fdf974a\n--- /dev/null\n+++ b/t/perf/README\n@@ -0,0 +1,146 @@\n+Git performance tests\n+=====================\n+\n+This directory holds performance testing scripts for git tools. The\n+first part of this document describes the various ways in which you\n+can run them.\n+\n+When fixing the tools or adding enhancements, you are strongly\n+encouraged to add tests in this directory to cover what you are\n+trying to fix or enhance. The later part of this short document\n+describes how your test scripts should be organized.\n+\n+\n+Running Tests\n+-------------\n+\n+The easiest way to run tests is to say \"make\". This runs all\n+the tests on the current git repository.\n+\n+ === Running 2 tests in this tree ===\n+ [...]\n+ Test this tree \n+ ---------------------------------------------------------\n+ 0001.1: rev-list --all 0.54(0.51+0.02) \n+ 0001.2: rev-list --all --objects 6.14(5.99+0.11) \n+ 7810.1: grep worktree, cheap regex 0.16(0.16+0.35) \n+ 7810.2: grep worktree, expensive regex 7.90(29.75+0.37)\n+ 7810.3: grep --cached, cheap regex 3.07(3.02+0.25) \n+ 7810.4: grep --cached, expensive regex 9.39(30.57+0.24)\n+\n+You can compare multiple repositories and even git revisions with the\n+'run' script:\n+\n+ $ ./run . origin/next /path/to/git-tree p0001-rev-list.sh\n+\n+where . stands for the current git tree. The full invocation is\n+\n+ ./run [...] [--] [...]\n+\n+A '.' argument is implied if you do not pass any other\n+revisions/directories.\n+\n+You can also manually test this or another git build tree, and then\n+call the aggregation script to summarize the results:\n+\n+ $ ./p0001-rev-list.sh\n+ [...]\n+ $ GIT_BUILD_DIR=/path/to/other/git ./p0001-rev-list.sh\n+ [...]\n+ $ ./aggregate.perl . /path/to/other/git ./p0001-rev-list.sh\n+\n+aggregate.perl has the same invocation as 'run', it just does not run\n+anything beforehand.\n+\n+You can set the following variables (also in your config.mak):\n+\n+ GIT_PERF_REPEAT_COUNT\n+\tNumber of times a test should be repeated for best-of-N\n+\tmeasurements. Defaults to 5.\n+\n+ GIT_PERF_MAKE_OPTS\n+\tOptions to use when automatically building a git tree for\n+\tperformance testing. E.g., -j6 would be useful.\n+\n+ GIT_PERF_REPO\n+ GIT_PERF_LARGE_REPO\n+\tRepositories to copy for the performance tests. The normal\n+\trepo should be at least git.git size. The large repo should\n+\tprobably be about linux-2.6.git size for optimal results.\n+\tBoth default to the git.git you are running from.\n+\n+You can also pass the options taken by ordinary git tests; the most\n+useful one is:\n+\n+--root=::\n+\tCreate \"trash\" directories used to store all temporary data during\n+\ttesting under , instead of the t/ directory.\n+\tUsing this option with a RAM-based filesystem (such as tmpfs)\n+\tcan massively speed up the test suite.\n+\n+\n+Naming Tests\n+------------\n+\n+The performance test files are named as:\n+\n+\tpNNNN-commandname-details.sh\n+\n+where N is a decimal digit. The same conventions for choosing NNNN as\n+for normal tests apply.\n+\n+\n+Writing Tests\n+-------------\n+\n+The perf script starts much like a normal test script, except it\n+sources perf-lib.sh:\n+\n+\t#!/bin/sh\n+\t#\n+\t# Copyright (c) 2005 Junio C Hamano\n+\t#\n+\n+\ttest_description='xxx performance test'\n+\t. ./perf-lib.sh\n+\n+After that you will want to use some of the following:\n+\n+\ttest_perf_default_repo # sets up a \"normal\" repository\n+\ttest_perf_large_repo # sets up a \"large\" repository\n+\n+\ttest_perf_default_repo sub # ditto, in a subdir \"sub\"\n+\n+ test_checkout_worktree # if you need the worktree too\n+\n+At least one of the first two is required!\n+\n+You can use test_expect_success as usual. For actual performance\n+tests, use\n+\n+\ttest_perf 'descriptive string' '\n+\t\tcommand1 &&\n+\t\tcommand2\n+\t'\n+\n+test_perf spawns a subshell, for lack of better options. This means\n+that\n+\n+* you _must_ export all variables that you need in the subshell\n+\n+* you _must_ flag all variables that you want to persist from the\n+ subshell with 'test_export':\n+\n+\ttest_perf 'descriptive string' '\n+\t\tfoo=$(git rev-parse HEAD) &&\n+\t\ttest_export foo\n+\t'\n+\n+ The so-exported variables are automatically marked for export in the\n+ shell executing the perf test. For your convenience, test_export is\n+ the same as export in the main shell.\n+\n+ This feature relies on a bit of magic using 'set' and 'source'.\n+ While we have tried to make sure that it can cope with embedded\n+ whitespace and other special characters, it will not work with\n+ multi-line data.\ndiff --git a/t/perf/aggregate.perl b/t/perf/aggregate.perl\nnew file mode 100755\nindex 0000000..15f7fc1\n--- /dev/null\n+++ b/t/perf/aggregate.perl\n@@ -0,0 +1,166 @@\n+#!/usr/bin/perl\n+\n+use strict;\n+use warnings;\n+use Git;\n+\n+sub get_times {\n+\tmy $name = shift;\n+\topen my $fh, \"<\", $name or return undef;\n+\tmy $line = <$fh>;\n+\treturn undef if not defined $line;\n+\tclose $fh or die \"cannot close $name: $!\";\n+\t$line =~ /^(?:(\\d+):)?(\\d+):(\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?)$/\n+\t\tor die \"bad input line: $line\";\n+\tmy $rt = ((defined $1 ? $1 : 0.0)*60+$2)*60+$3;\n+\treturn ($rt, $4, $5);\n+}\n+\n+sub format_times {\n+\tmy ($r, $u, $s, $firstr) = @_;\n+\tif (!defined $r) {\n+\t\treturn \"\";\n+\t}\n+\tmy $out = sprintf \"%.2f(%.2f+%.2f)\", $r, $u, $s;\n+\tif (defined $firstr) {\n+\t\tif ($firstr > 0) {\n+\t\t\t$out .= sprintf \" %+.1f%%\", 100.0*($r-$firstr)/$firstr;\n+\t\t} elsif ($r == 0) {\n+\t\t\t$out .= \" =\";\n+\t\t} else {\n+\t\t\t$out .= \" +inf\";\n+\t\t}\n+\t}\n+\treturn $out;\n+}\n+\n+my (@dirs, %dirnames, %dirabbrevs, %prefixes, @tests);\n+while (scalar @ARGV) {\n+\tmy $arg = $ARGV[0];\n+\tmy $dir;\n+\tlast if -f $arg or $arg eq \"--\";\n+\tif (! -d $arg) {\n+\t\tmy $rev = Git::command_oneline(qw(rev-parse --verify), $arg);\n+\t\t$dir = \"build/\".$rev;\n+\t} else {\n+\t\t$arg =~ s{/*$}{};\n+\t\t$dir = $arg;\n+\t\t$dirabbrevs{$dir} = $dir;\n+\t}\n+\tpush @dirs, $dir;\n+\t$dirnames{$dir} = $arg;\n+\tmy $prefix = $dir;\n+\t$prefix =~ tr/^a-zA-Z0-9/_/c;\n+\t$prefixes{$dir} = $prefix . '.';\n+\tshift @ARGV;\n+}\n+\n+if (not @dirs) {\n+\t@dirs = ('.');\n+}\n+$dirnames{'.'} = $dirabbrevs{'.'} = \"this tree\";\n+$prefixes{'.'} = '';\n+\n+shift @ARGV if scalar @ARGV and $ARGV[0] eq \"--\";\n+\n+@tests = @ARGV;\n+if (not @tests) {\n+\t@tests = glob \"p????-*.sh\";\n+}\n+\n+my @subtests;\n+my %shorttests;\n+for my $t (@tests) {\n+\t$t =~ s{(?:.*/)?(p(\\d+)-[^/]+)\\.sh$}{$1} or die \"bad test name: $t\";\n+\tmy $n = $2;\n+\tmy $fname = \"test-results/$t.subtests\";\n+\topen my $fp, \"<\", $fname or die \"cannot open $fname: $!\";\n+\tfor (<$fp>) {\n+\t\tchomp;\n+\t\t/^(\\d+)$/ or die \"malformed subtest line: $_\";\n+\t\tpush @subtests, \"$t.$1\";\n+\t\t$shorttests{\"$t.$1\"} = \"$n.$1\";\n+\t}\n+\tclose $fp or die \"cannot close $fname: $!\";\n+}\n+\n+sub read_descr {\n+\tmy $name = shift;\n+\topen my $fh, \"<\", $name or return \"\";\n+\tmy $line = <$fh>;\n+\tclose $fh or die \"cannot close $name\";\n+\tchomp $line;\n+\treturn $line;\n+}\n+\n+my %descrs;\n+my $descrlen = 4; # \"Test\"\n+for my $t (@subtests) {\n+\t$descrs{$t} = $shorttests{$t}.\": \".read_descr(\"test-results/$t.descr\");\n+\t$descrlen = length $descrs{$t} if length $descrs{$t}>$descrlen;\n+}\n+\n+sub have_duplicate {\n+\tmy %seen;\n+\tfor (@_) {\n+\t\treturn 1 if exists $seen{$_};\n+\t\t$seen{$_} = 1;\n+\t}\n+\treturn 0;\n+}\n+sub have_slash {\n+\tfor (@_) {\n+\t\treturn 1 if m{/};\n+\t}\n+\treturn 0;\n+}\n+\n+my %newdirabbrevs = %dirabbrevs;\n+while (!have_duplicate(values %newdirabbrevs)) {\n+\t%dirabbrevs = %newdirabbrevs;\n+\tlast if !have_slash(values %dirabbrevs);\n+\t%newdirabbrevs = %dirabbrevs;\n+\tfor (values %newdirabbrevs) {\n+\t\ts{^[^/]*/}{};\n+\t}\n+}\n+\n+my %times;\n+my @colwidth = ((0)x@dirs);\n+for my $i (0..$#dirs) {\n+\tmy $d = $dirs[$i];\n+\tmy $w = length (exists $dirabbrevs{$d} ? $dirabbrevs{$d} : $dirnames{$d});\n+\t$colwidth[$i] = $w if $w > $colwidth[$i];\n+}\n+for my $t (@subtests) {\n+\tmy $firstr;\n+\tfor my $i (0..$#dirs) {\n+\t\tmy $d = $dirs[$i];\n+\t\t$times{$prefixes{$d}.$t} = [get_times(\"test-results/$prefixes{$d}$t.times\")];\n+\t\tmy ($r,$u,$s) = @{$times{$prefixes{$d}.$t}};\n+\t\tmy $w = length format_times($r,$u,$s,$firstr);\n+\t\t$colwidth[$i] = $w if $w > $colwidth[$i];\n+\t\t$firstr = $r unless defined $firstr;\n+\t}\n+}\n+my $totalwidth = 3*@dirs+$descrlen;\n+$totalwidth += $_ for (@colwidth);\n+\n+printf \"%-${descrlen}s\", \"Test\";\n+for my $i (0..$#dirs) {\n+\tmy $d = $dirs[$i];\n+\tprintf \" %-$colwidth[$i]s\", (exists $dirabbrevs{$d} ? $dirabbrevs{$d} : $dirnames{$d});\n+}\n+print \"\\n\";\n+print \"-\"x$totalwidth, \"\\n\";\n+for my $t (@subtests) {\n+\tprintf \"%-${descrlen}s\", $descrs{$t};\n+\tmy $firstr;\n+\tfor my $i (0..$#dirs) {\n+\t\tmy $d = $dirs[$i];\n+\t\tmy ($r,$u,$s) = @{$times{$prefixes{$d}.$t}};\n+\t\tprintf \" %-$colwidth[$i]s\", format_times($r,$u,$s,$firstr);\n+\t\t$firstr = $r unless defined $firstr;\n+\t}\n+\tprint \"\\n\";\n+}\ndiff --git a/t/perf/min_time.perl b/t/perf/min_time.perl\nnew file mode 100755\nindex 0000000..c1a2717\n--- /dev/null\n+++ b/t/perf/min_time.perl\n@@ -0,0 +1,21 @@\n+#!/usr/bin/perl\n+\n+my $minrt = 1e100;\n+my $min;\n+\n+while (<>) {\n+\t# [h:]m:s.xx U.xx S.xx\n+\t/^(?:(\\d+):)?(\\d+):(\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?) (\\d+(?:\\.\\d+)?)$/\n+\t\tor die \"bad input line: $_\";\n+\tmy $rt = ((defined $1 ? $1 : 0.0)*60+$2)*60+$3;\n+\tif ($rt < $minrt) {\n+\t\t$min = $_;\n+\t\t$minrt = $rt;\n+\t}\n+}\n+\n+if (!defined $min) {\n+\tdie \"no input found\";\n+}\n+\n+print $min;\ndiff --git a/t/perf/p0000-perf-lib-sanity.sh b/t/perf/p0000-perf-lib-sanity.sh\nnew file mode 100755\nindex 0000000..2ca4aac\n--- /dev/null\n+++ b/t/perf/p0000-perf-lib-sanity.sh\n@@ -0,0 +1,41 @@\n+#!/bin/sh\n+\n+test_description='Tests whether perf-lib facilities work'\n+. ./perf-lib.sh\n+\n+test_perf_default_repo\n+\n+test_perf 'test_perf_default_repo works' '\n+\tfoo=$(git rev-parse HEAD) &&\n+\ttest_export foo\n+'\n+\n+test_checkout_worktree\n+\n+test_perf 'test_checkout_worktree works' '\n+\twt=$(find . | wc -l) &&\n+\tidx=$(git ls-files | wc -l) &&\n+\ttest $wt -gt $idx\n+'\n+\n+baz=baz\n+test_export baz\n+\n+test_expect_success 'test_export works' '\n+\techo \"$foo\" &&\n+\ttest \"$foo\" = \"$(git rev-parse HEAD)\" &&\n+\techo \"$baz\" &&\n+\ttest \"$baz\" = baz\n+'\n+\n+test_perf 'export a weird var' '\n+\tbar=\"weird # variable\" &&\n+\ttest_export bar\n+'\n+\n+test_expect_success 'test_export works with weird vars' '\n+\techo \"$bar\" &&\n+\ttest \"$bar\" = \"weird # variable\"\n+'\n+\n+test_done\ndiff --git a/t/perf/p0001-rev-list.sh b/t/perf/p0001-rev-list.sh\nnew file mode 100755\nindex 0000000..4f71a63\n--- /dev/null\n+++ b/t/perf/p0001-rev-list.sh\n@@ -0,0 +1,17 @@\n+#!/bin/sh\n+\n+test_description=\"Tests history walking performance\"\n+\n+. ./perf-lib.sh\n+\n+test_perf_default_repo\n+\n+test_perf 'rev-list --all' '\n+\tgit rev-list --all >/dev/null\n+'\n+\n+test_perf 'rev-list --all --objects' '\n+\tgit rev-list --all --objects >/dev/null\n+'\n+\n+test_done\ndiff --git a/t/perf/perf-lib.sh b/t/perf/perf-lib.sh\nnew file mode 100644\nindex 0000000..07e9b09\n--- /dev/null\n+++ b/t/perf/perf-lib.sh\n@@ -0,0 +1,198 @@\n+#!/bin/bash\n+#\n+# Copyright (c) 2011 Thomas Rast\n+#\n+# This program is free software: you can redistribute it and/or modify\n+# it under the terms of the GNU General Public License as published by\n+# the Free Software Foundation, either version 2 of the License, or\n+# (at your option) any later version.\n+#\n+# This program is distributed in the hope that it will be useful,\n+# but WITHOUT ANY WARRANTY; without even the implied warranty of\n+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n+# GNU General Public License for more details.\n+#\n+# You should have received a copy of the GNU General Public License\n+# along with this program. If not, see http://www.gnu.org/licenses/ .\n+\n+# do the --tee work early; it otherwise confuses our careful\n+# GIT_BUILD_DIR mangling\n+case \"$GIT_TEST_TEE_STARTED, $* \" in\n+done,*)\n+\t# do not redirect again\n+\t;;\n+*' --tee '*|*' --va'*)\n+\tmkdir -p test-results\n+\tBASE=test-results/$(basename \"$0\" .sh)\n+\t(GIT_TEST_TEE_STARTED=done ${SHELL-sh} \"$0\" \"$@\" 2>&1;\n+\t echo $? > $BASE.exit) | tee $BASE.out\n+\ttest \"$(cat $BASE.exit)\" = 0\n+\texit\n+\t;;\n+esac\n+\n+TEST_DIRECTORY=$(pwd)/..\n+TEST_OUTPUT_DIRECTORY=$(pwd)\n+if test -z \"$GIT_TEST_INSTALLED\"; then\n+\tperf_results_prefix=\n+else\n+\tperf_results_prefix=$(printf \"%s\" \"${GIT_TEST_INSTALLED%/bin-wrappers}\" | tr -c \"[a-zA-Z0-9]\" \"[_*]\")\".\"\n+\t# make the tested dir absolute\n+\tGIT_TEST_INSTALLED=$(cd \"$GIT_TEST_INSTALLED\" && pwd)\n+fi\n+\n+TEST_NO_CREATE_REPO=t\n+\n+. ../test-lib.sh\n+\n+perf_results_dir=$TEST_OUTPUT_DIRECTORY/test-results\n+mkdir -p \"$perf_results_dir\"\n+rm -f \"$perf_results_dir\"/$(basename \"$0\" .sh).subtests\n+\n+if test -z \"$GIT_PERF_REPEAT_COUNT\"; then\n+\tGIT_PERF_REPEAT_COUNT=3\n+fi\n+die_if_build_dir_not_repo () {\n+\tif ! ( cd \"$TEST_DIRECTORY/..\" &&\n+\t\t git rev-parse --build-dir >/dev/null 2>&1 ); then\n+\t\terror \"No $1 defined, and your build directory is not a repo\"\n+\tfi\n+}\n+\n+if test -z \"$GIT_PERF_REPO\"; then\n+\tdie_if_build_dir_not_repo '$GIT_PERF_REPO'\n+\tGIT_PERF_REPO=$TEST_DIRECTORY/..\n+fi\n+if test -z \"$GIT_PERF_LARGE_REPO\"; then\n+\tdie_if_build_dir_not_repo '$GIT_PERF_LARGE_REPO'\n+\tGIT_PERF_LARGE_REPO=$TEST_DIRECTORY/..\n+fi\n+\n+test_perf_create_repo_from () {\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 parameters to test-create-repo\"\n+\trepo=\"$1\"\n+\tsource=\"$2\"\n+\tsource_git=$source/$(cd \"$source\" && git rev-parse --git-dir)\n+\tmkdir -p \"$repo/.git\"\n+\t(\n+\t\tcd \"$repo/.git\" &&\n+\t\t{ cp -Rl \"$source_git/objects\" . 2>/dev/null ||\n+\t\t\tcp -R \"$source_git/objects\" .; } &&\n+\t\tfor stuff in \"$source_git\"/*; do\n+\t\t\tcase \"$stuff\" in\n+\t\t\t\t*/objects|*/hooks|*/config)\n+\t\t\t\t\t;;\n+\t\t\t\t*)\n+\t\t\t\t\tcp -R \"$stuff\" . || break\n+\t\t\t\t\t;;\n+\t\t\tesac\n+\t\tdone &&\n+\t\tcd .. &&\n+\t\tgit init -q &&\n+\t\tmv .git/hooks .git/hooks-disabled 2>/dev/null\n+\t) || error \"failed to copy repository '$source' to '$repo'\"\n+}\n+\n+# call at least one of these to establish an appropriately-sized repository\n+test_perf_default_repo () {\n+\ttest_perf_create_repo_from \"${1:-$TRASH_DIRECTORY}\" \"$GIT_PERF_REPO\"\n+}\n+test_perf_large_repo () {\n+\tif test \"$GIT_PERF_LARGE_REPO\" = \"$GIT_BUILD_DIR\"; then\n+\t\techo \"warning: \\$GIT_PERF_LARGE_REPO is \\$GIT_BUILD_DIR.\" >&2\n+\t\techo \"warning: This will work, but may not be a sufficiently large repo\" >&2\n+\t\techo \"warning: for representative measurements.\" >&2\n+\tfi\n+\ttest_perf_create_repo_from \"${1:-$TRASH_DIRECTORY}\" \"$GIT_PERF_LARGE_REPO\"\n+}\n+test_checkout_worktree () {\n+\tgit checkout-index -u -a ||\n+\terror \"git checkout-index failed\"\n+}\n+\n+# Performance tests should never fail. If they do, stop immediately\n+immediate=t\n+\n+test_run_perf_ () {\n+\ttest_cleanup=:\n+\ttest_export_=\"test_cleanup\"\n+\texport test_cleanup test_export_\n+\t/usr/bin/time -f \"%E %U %S\" -o test_time.$i \"$SHELL\" -c '\n+. '\"$TEST_DIRECTORY\"/../test-lib-functions.sh'\n+test_export () {\n+\t[ $# != 0 ] || return 0\n+\ttest_export_=\"$test_export_\\\\|$1\"\n+\tshift\n+\ttest_export \"$@\"\n+}\n+'\"$1\"'\n+ret=$?\n+set | sed -n \"s'\"/'/'\\\\\\\\''/g\"';s/^\\\\($test_export_\\\\)/export '\"'&'\"'/p\" >test_vars\n+exit $ret' >&3 2>&4\n+\teval_ret=$?\n+\n+\tif test $eval_ret = 0 || test -n \"$expecting_failure\"\n+\tthen\n+\t\ttest_eval_ \"$test_cleanup\"\n+\t\tsource ./test_vars || error \"failed to load updated environment\"\n+\tfi\n+\tif test \"$verbose\" = \"t\" && test -n \"$HARNESS_ACTIVE\"; then\n+\t\techo \"\"\n+\tfi\n+\treturn \"$eval_ret\"\n+}\n+\n+\n+test_perf () {\n+\ttest \"$#\" = 3 && { test_prereq=$1; shift; } || test_prereq=\n+\ttest \"$#\" = 2 ||\n+\terror \"bug in the test script: not 2 or 3 parameters to test-expect-success\"\n+\texport test_prereq\n+\tif ! test_skip \"$@\"\n+\tthen\n+\t\tbase=$(basename \"$0\" .sh)\n+\t\techo \"$test_count\" >>\"$perf_results_dir\"/$base.subtests\n+\t\techo \"$1\" >\"$perf_results_dir\"/$base.$test_count.descr\n+\t\tif test -z \"$verbose\"; then\n+\t\t\techo -n \"perf $test_count - $1:\"\n+\t\telse\n+\t\t\techo \"perf $test_count - $1:\"\n+\t\tfi\n+\t\tfor i in $(seq 1 $GIT_PERF_REPEAT_COUNT); do\n+\t\t\tsay >&3 \"running: $2\"\n+\t\t\tif test_run_perf_ \"$2\"\n+\t\t\tthen\n+\t\t\t\tif test -z \"$verbose\"; then\n+\t\t\t\t\techo -n \" $i\"\n+\t\t\t\telse\n+\t\t\t\t\techo \"* timing run $i/$GIT_PERF_REPEAT_COUNT:\"\n+\t\t\t\tfi\n+\t\t\telse\n+\t\t\t\ttest -z \"$verbose\" && echo\n+\t\t\t\ttest_failure_ \"$@\"\n+\t\t\t\tbreak\n+\t\t\tfi\n+\t\tdone\n+\t\tif test -z \"$verbose\"; then\n+\t\t\techo \" ok\"\n+\t\telse\n+\t\t\ttest_ok_ \"$1\"\n+\t\tfi\n+\t\tbase=\"$perf_results_dir\"/\"$perf_results_prefix$(basename \"$0\" .sh)\".\"$test_count\"\n+\t\t\"$TEST_DIRECTORY\"/perf/min_time.perl test_time.* >\"$base\".times\n+\tfi\n+\techo >&3 \"\"\n+}\n+\n+# We extend test_done to print timings at the end (./run disables this\n+# and does it after running everything)\n+test_at_end_hook_ () {\n+\tif test -z \"$GIT_PERF_AGGREGATING_LATER\"; then\n+\t\t( cd \"$TEST_DIRECTORY\"/perf && ./aggregate.perl $(basename \"$0\") )\n+\tfi\n+}\n+\n+test_export () {\n+\texport \"$@\"\n+}\ndiff --git a/t/perf/run b/t/perf/run\nnew file mode 100755\nindex 0000000..cfd7012\n--- /dev/null\n+++ b/t/perf/run\n@@ -0,0 +1,82 @@\n+#!/bin/sh\n+\n+case \"$1\" in\n+\t--help)\n+\t\techo \"usage: $0 [other_git_tree...] [--] [test_scripts]\"\n+\t\texit 0\n+\t\t;;\n+esac\n+\n+die () {\n+\techo >&2 \"error: $*\"\n+\texit 1\n+}\n+\n+run_one_dir () {\n+\tif test $# -eq 0; then\n+\t\tset -- p????-*.sh\n+\tfi\n+\techo \"=== Running $# tests in ${GIT_TEST_INSTALLED:-this tree} ===\"\n+\tfor t in \"$@\"; do\n+\t\t./$t $GIT_TEST_OPTS\n+\tdone\n+}\n+\n+unpack_git_rev () {\n+\trev=$1\n+\tmkdir -p build/$rev\n+\t(cd \"$(git rev-parse --show-cdup)\" && git archive --format=tar $rev) |\n+\t(cd build/$rev && tar x)\n+}\n+build_git_rev () {\n+\trev=$1\n+\tcp ../../config.mak build/$rev/config.mak\n+\t(cd build/$rev && make $GIT_PERF_MAKE_OPTS) ||\n+\tdie \"failed to build revision '$mydir'\"\n+}\n+\n+run_dirs_helper () {\n+\tmydir=${1%/}\n+\tshift\n+\twhile test $# -gt 0 -a \"$1\" != -- -a ! -f \"$1\"; do\n+\t\tshift\n+\tdone\n+\tif test $# -gt 0 -a \"$1\" = --; then\n+\t\tshift\n+\tfi\n+\tif [ ! -d \"$mydir\" ]; then\n+\t\trev=$(git rev-parse --verify \"$mydir\" 2>/dev/null) ||\n+\t\tdie \"'$mydir' is neither a directory nor a valid revision\"\n+\t\tif [ ! -d build/$rev ]; then\n+\t\t\tunpack_git_rev $rev\n+\t\tfi\n+\t\tbuild_git_rev $rev\n+\t\tmydir=build/$rev\n+\tfi\n+\tif test \"$mydir\" = .; then\n+\t\tunset GIT_TEST_INSTALLED\n+\telse\n+\t\tGIT_TEST_INSTALLED=\"$mydir/bin-wrappers\"\n+\t\texport GIT_TEST_INSTALLED\n+\tfi\n+\trun_one_dir \"$@\"\n+}\n+\n+run_dirs () {\n+\twhile test $# -gt 0 -a \"$1\" != -- -a ! -f \"$1\"; do\n+\t\trun_dirs_helper \"$@\"\n+\t\tshift\n+\tdone\n+}\n+\n+GIT_PERF_AGGREGATING_LATER=t\n+export GIT_PERF_AGGREGATING_LATER\n+\n+cd \"$(dirname $0)\"\n+. ../../GIT-BUILD-OPTIONS\n+\n+if test $# = 0 -o \"$1\" = -- -o -f \"$1\"; then\n+\tset -- . \"$@\"\n+fi\n+run_dirs \"$@\"\n+./aggregate.perl \"$@\"\ndiff --git a/t/test-lib.sh b/t/test-lib.sh\nindex 1da3f40..d75766a 100644\n--- a/t/test-lib.sh\n+++ b/t/test-lib.sh\n@@ -55,6 +55,7 @@ unset $(perl -e '\n \t\t.*_TEST\n \t\tPROVE\n \t\tVALGRIND\n+\t\tPERF_AGGREGATING_LATER\n \t));\n \tmy @vars = grep(/^GIT_/ && !/^GIT_($ok)/o, @env);\n \tprint join(\"\\n\", @vars);\n@@ -98,6 +99,8 @@ _z40=0000000000000000000000000000000000000000\n LF='\n '\n \n+export _x05 _x40 _z40 LF\n+\n # Each test should start with something like this, after copyright notices:\n #\n # test_description='Description of this test...\n@@ -313,11 +316,16 @@ test_skip () {\n \tesac\n }\n \n+# stub; perf-lib overrides it\n+test_at_end_hook_ () {\n+\t:\n+}\n+\n test_done () {\n \tGIT_EXIT_OK=t\n \n \tif test -z \"$HARNESS_ACTIVE\"; then\n-\t\ttest_results_dir=\"$TEST_DIRECTORY/test-results\"\n+\t\ttest_results_dir=\"$TEST_OUTPUT_DIRECTORY/test-results\"\n \t\tmkdir -p \"$test_results_dir\"\n \t\ttest_results_path=\"$test_results_dir/${0%.sh}-$$.counts\"\n \n@@ -356,6 +364,8 @@ test_done () {\n \t\tcd \"$(dirname \"$remove_trash\")\" &&\n \t\trm -rf \"$(basename \"$remove_trash\")\"\n \n+\t\ttest_at_end_hook_\n+\n \t\texit 0 ;;\n \n \t*)\n@@ -378,6 +388,12 @@ then\n \t# itself.\n \tTEST_DIRECTORY=$(pwd)\n fi\n+if test -z \"$TEST_OUTPUT_DIRECTORY\"\n+then\n+\t# Similarly, override this to store the test-results subdir\n+\t# elsewhere\n+\tTEST_OUTPUT_DIRECTORY=$TEST_DIRECTORY\n+fi\n GIT_BUILD_DIR=\"$TEST_DIRECTORY\"/..\n \n if test -n \"$valgrind\"\n@@ -513,7 +529,7 @@ test=\"trash directory.$(basename \"$0\" .sh)\"\n test -n \"$root\" && test=\"$root/$test\"\n case \"$test\" in\n /*) TRASH_DIRECTORY=\"$test\" ;;\n- *) TRASH_DIRECTORY=\"$TEST_DIRECTORY/$test\" ;;\n+ *) TRASH_DIRECTORY=\"$TEST_OUTPUT_DIRECTORY/$test\" ;;\n esac\n test ! -z \"$debug\" || remove_trash=$TRASH_DIRECTORY\n rm -fr \"$test\" || {\n@@ -525,7 +541,11 @@ rm -fr \"$test\" || {\n HOME=\"$TRASH_DIRECTORY\"\n export HOME\n \n-test_create_repo \"$test\"\n+if test -z \"$TEST_NO_CREATE_REPO\"; then\n+\ttest_create_repo \"$test\"\n+else\n+\tmkdir -p \"$test\"\n+fi\n # Use -P to resolve symlinks in our working directory so that the cwd\n # in subprocesses like git equals our $PWD (for pathname comparisons).\n cd -P \"$test\" || exit 1\n-- \n1.7.9.1.365.ge223f\n"},{"id":"184873","messageId":"95250e73e5610f8415fa0c6a3527b067ec0f75f2.1329472405.git.trast@student.ethz.ch","threadId":"29643","inReplyTo":"cover.1329472405.git.trast@student.ethz.ch","subject":"[PATCH v3 3/3] Add a performance test for git-grep","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2012-02-17T10:25:10Z","receivedAt":"2012-02-17T10:25:10Z","isPatch":true,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"The only catch is that we don't really know what our repo contains, so\nwe have to ignore any possible \"not found\" status from git-grep.\n\nSigned-off-by: Thomas Rast \n---\n t/perf/p7810-grep.sh | 23 +++++++++++++++++++++++\n 1 file changed, 23 insertions(+)\n create mode 100755 t/perf/p7810-grep.sh\n\ndiff --git a/t/perf/p7810-grep.sh b/t/perf/p7810-grep.sh\nnew file mode 100755\nindex 0000000..9f4ade6\n--- /dev/null\n+++ b/t/perf/p7810-grep.sh\n@@ -0,0 +1,23 @@\n+#!/bin/sh\n+\n+test_description=\"git-grep performance in various modes\"\n+\n+. ./perf-lib.sh\n+\n+test_perf_large_repo\n+test_checkout_worktree\n+\n+test_perf 'grep worktree, cheap regex' '\n+\tgit grep some_nonexistent_string || :\n+'\n+test_perf 'grep worktree, expensive regex' '\n+\tgit grep \"^.* *some_nonexistent_string$\" || :\n+'\n+test_perf 'grep --cached, cheap regex' '\n+\tgit grep --cached some_nonexistent_string || :\n+'\n+test_perf 'grep --cached, expensive regex' '\n+\tgit grep --cached \"^.* *some_nonexistent_string$\" || :\n+'\n+\n+test_done\n-- \n1.7.9.1.365.ge223f\n"},{"id":"184889","messageId":"7v62f5v1d1.fsf@alter.siamese.dyndns.org","threadId":"29643","inReplyTo":"cover.1329472405.git.trast@student.ethz.ch","subject":"Re: [PATCH v3 0/3]","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2012-02-17T17:03:22Z","receivedAt":"2012-02-17T17:03:22Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Thomas Rast writes:\n\n> There were actually more mistakes lurking :-( so I am resending the\n> whole series.\n\nOk, will requeue. The diff you attached to this cover letter looked at\nleast halfway sane, compared to the previous round ;-), though it is not\nexactly clear what goes to lib-test-functions and what goes to lib-test\n(for example, you moved test_expect_success to 'test-functions', but it\ncalls test_ok_ that is in 'test-lib', and test_ok_ is directly used by\ntest_perf in the new 't/perf/perf-lib.sh'), making it harder for people to\ndecide where to put their additions to the test infrastructure from now\non. There needs a bit of description in the first patch to guide them.\n\nI seem to be getting intermittent test failures, and every time the\nfailing tests are different, when these three are queued to 'pu'. I didn't\nlook for what goes wrong and how.\n\nThanks.\n"},{"id":"184912","messageId":"7vk43lqbt8.fsf@alter.siamese.dyndns.org","threadId":"29643","inReplyTo":"7v62f5v1d1.fsf@alter.siamese.dyndns.org","subject":"Re: [PATCH v3 0/3]","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2012-02-17T23:28:51Z","receivedAt":"2012-02-17T23:28:51Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Junio C Hamano writes:\n\n> Thomas Rast writes:\n> ...\n> I seem to be getting intermittent test failures, and every time the\n> failing tests are different, when these three are queued to 'pu'. I didn't\n> look for what goes wrong and how.\n\nFalse alarm. I suspect that it is jb/required-filter topic that is causing\nintermittent failures from convert.c depending on the timing of how fast\nfilter subprocess dies vs how fast we consume its result or something.\n\nRepeatedly running t0021 like this:\n\n $ cd t\n $ while sh t0021-conversion.sh ; do :; done\n\nunder load seems to make it fail every once in a while.\n\ntest_must_fail: died by signal: git add test.fc\n\nAre we dying on SIGPIPE or something?\n"},{"id":"184916","messageId":"20120218005148.GA1940@sigill.intra.peff.net","threadId":"29643","inReplyTo":"7vk43lqbt8.fsf@alter.siamese.dyndns.org","subject":"Re: [PATCH v3 0/3]","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2012-02-18T00:51:48Z","receivedAt":"2012-02-18T00:51:48Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Fri, Feb 17, 2012 at 03:28:51PM -0800, Junio C Hamano wrote:\n\n> Junio C Hamano writes:\n> \n> > Thomas Rast writes:\n> > ...\n> > I seem to be getting intermittent test failures, and every time the\n> > failing tests are different, when these three are queued to 'pu'. I didn't\n> > look for what goes wrong and how.\n> \n> False alarm. I suspect that it is jb/required-filter topic that is causing\n> intermittent failures from convert.c depending on the timing of how fast\n> filter subprocess dies vs how fast we consume its result or something.\n> \n> Repeatedly running t0021 like this:\n> \n> $ cd t\n> $ while sh t0021-conversion.sh ; do :; done\n> \n> under load seems to make it fail every once in a while.\n> \n> test_must_fail: died by signal: git add test.fc\n> \n> Are we dying on SIGPIPE or something?\n\nI would be unsurprised if that is the case. Joey Hess mentioned similar\nissues with hooks a month or two ago. And I have been seeing\nintermittent failures of t5541 under load that I traced back to SIGPIPE.\nI've been meaning to dig further and come up with a good solution.\nHere's some previous discussion:\n\n http://article.gmane.org/gmane.comp.version-control.git/186291\n\nI'd be happy if we just ignored SIGPIPE everywhere, but turned it on for\nthe log family.\n\n-Peff\n"},{"id":"184922","messageId":"7v8vk0r481.fsf@alter.siamese.dyndns.org","threadId":"29643","inReplyTo":"20120218005148.GA1940@sigill.intra.peff.net","subject":"Re: [PATCH v3 0/3]","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2012-02-18T07:27:26Z","receivedAt":"2012-02-18T07:27:26Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jeff King writes:\n\n> I'd be happy if we just ignored SIGPIPE everywhere, but turned it on for\n> the log family.\n\nHmmmm, now you confused me... What is special about the log family? Do\nyou mean \"when we use pager\"? But then we are writing into the pager,\nwhich the user can make it exit, which in turn causes us to write into the\npipe, so I would expect that we would want to ignore SIGPIPE --- ah, then\nwe explicitly catch error in xwrite() and say die() which we do not want.\n\nSo you want to let SIGPIPE silently kill us when the pager is in use; is\nthat it?\n"},{"id":"184923","messageId":"20120218085221.GA13922@sigill.intra.peff.net","threadId":"29643","inReplyTo":"7v8vk0r481.fsf@alter.siamese.dyndns.org","subject":"Re: [PATCH v3 0/3]","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2012-02-18T08:52:21Z","receivedAt":"2012-02-18T08:52:21Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Fri, Feb 17, 2012 at 11:27:26PM -0800, Junio C Hamano wrote:\n\n> > I'd be happy if we just ignored SIGPIPE everywhere, but turned it on for\n> > the log family.\n> \n> Hmmmm, now you confused me... What is special about the log family? Do\n> you mean \"when we use pager\"? But then we are writing into the pager,\n> which the user can make it exit, which in turn causes us to write into the\n> pipe, so I would expect that we would want to ignore SIGPIPE --- ah, then\n> we explicitly catch error in xwrite() and say die() which we do not want.\n\nSort of. I mentioned the log family because those are the commands that\ntend to generate large amounts of input, and which are likely to hit\nSIGPIPE. But let me elaborate on my thinking a bit.\n\nThere are basically three types of writing callsites in git:\n\n 1. Careful calls to write() or fwrite().\n\n These are the majority of calls. In these, we check the return\n value of write() and die, return the error code, or whatever is\n appropriate for the callsite. SIGPIPE kills the program before the\n careful code gets the chance to make a decision about what to do.\n At best, this is a nuisance; even if the program is going to die(),\n it's likely that the custom code could produce a useful error\n message. At worst it causes bugs. For example, we may write to a\n subprocess that only needs part of our input to make a decision\n (e.g., some hooks and credential helpers). With SIGPIPE, we end up\n dying even though no error has occurred. Worse, these are often\n annoying heisenbugs that depend on the timing of write() and\n close() between the two processes.\n\n 2. Non-careful calls of small, fixed-size output.\n\n Things like \"git remote\" will write some output to stdout via\n printf and not bother checking the return code. As the main purpose\n of the program is to create output, it's OK to be killed by\n SIGPIPE if it happens due to a write to stdout. But respecting\n SIGPIPE doesn't buy as all that much. It might save us from making\n a few write() calls that will go to nowhere, but most of the work\n has already been done by the time we are outputting.\n\n And it has a cost to the other careful write calls in the same\n program, because respecting SIGPIPE is a per-process thing. So for\n something like \"git remote show foo\", while we would like SIGPIPE\n to kick in for output to stdout, we would not want it for a pipe we\n opened to talk to ls-remote.\n\n 3. Non-careful calls of large, streaming data.\n\n Commands like \"git log\" will non-carefully output to stdout, as\n well, but they will generate tons of data, consuming possibly\n minutes of CPU time (e.g., \"git log -p\"). If whoever is reading the\n output stops doing so, we really want to kill the program to avoid\n wasting CPU time. In this instance, SIGPIPE is a big win.\n\n It still has the downside that careful calls in the same program\n will be subject to using SIGPIPE. For \"log\" and friends, this is\n probably OK with the current code, as we don't make a lot of pipes.\n But that is somewhat of an implementation detail. E.g., \"git log\n -p\" with external diff or textconv writes to a tempfile, and then\n runs a subprocess with the tempfile as input. But we could just as\n easily have used pipes, and may choose to do so in the future. You\n may even be able to trigger a convert_to_git filter in the current\n code, which does use pipes.\n\nSo basically, I find SIGPIPE to be a simplistic too-heavy hammer that\nends up affecting all writes to pipes, when in reality we are only\ninterested in affecting ones to some \"main\" output (usually stdout).\nThat works OK for small Unix-y programs like \"head\", but is overly\nsimplistic for something as big as git.\n\nIn an ideal world, we could set a per-descriptor version of SIGPIPE, and\njust turn it on for stdout (or somehow find out which descriptor caused\nthe SIGPIPE after the fact). But that's not possible.\n\nSo our next best thing would be to actually check the results of\nthese non-careful writes. Unfortunately, this means either:\n\n a. wrapping every printf with a function that will appropriately die\n on error. This makes the code more cumbersome.\n\n or\n\n b. occasionally checking ferror(stdout) while doing long streams\n (e.g., checking it after each commit is written in git log, and\n aborting if we saw a write error). This is less cumbersome, but it\n does mean that errno may or may not still be accurate by the time\n we notice the error. So it's hard to reliably differentiate EPIPE\n from other errors. It would be nice, for example, to have git log\n exit silently on EPIPE, but properly print an error for something\n like ENOSPC. But perhaps that isn't a big deal, as I believe right\n now that we would silently ignore something like ENOSPC.\n\nLess robust than that is to just ignore SIGPIPE in most git programs\n(which don't benefit from it, and where it is only a liability), but\nthen manually enable it for the few that care (the log family, and\nperhaps diff. Maybe things like \"git tag -l\", though that output tends\nto be pretty small). But I think it would work OK in practice, because\nthose commands don't tend to make pipes other than the \"main\" output.\nAnd it's quite easy to implement.\n\n> So you want to let SIGPIPE silently kill us when the pager is in use; is\n> that it?\n\nThat's an OK heuristic, as the pager being in use is a good sign that we\nwill generate long, streaming output. It has two downsides, though. One\nis that it suffers from the too-heavy hammer of the preceding paragraph.\nThe other is that it only catches when _we_ create the pipe. You would\nalso want to catch something like:\n\n git rev-list | head\n\nand stop the traversal. So I think it is less about whether a pager is\nin use and more about whether we are creating long output that would\nbenefit from being cut off early. The pager is an indicator of that, but\nit's not a perfect one; I think we'd do better to mark those spots\nmanually (i.e., by re-enabling SIGPIPE in commands that we deem\nappropriate).\n\n-Peff\n"},{"id":"184924","messageId":"20120218100517.GA8998@burratino","threadId":"29643","inReplyTo":"20120218085221.GA13922@sigill.intra.peff.net","subject":"SIGPIPE handling (Re: [PATCH v3 0/3])","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2012-02-18T10:06:07Z","receivedAt":"2012-02-18T10:06:07Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Hi,\n\nJeff King wrote:\n\n> Less robust than that is to just ignore SIGPIPE in most git programs\n> (which don't benefit from it, and where it is only a liability), but\n> then manually enable it for the few that care\n\nThis seems backwards. Aren't the only places where it is just a\nliability places where git is writing to a pipe that git has created?\n\nWe could keep the benefits of SIGPIPE (including simpler error\nhandling and lack of distracting EPIPE message) in most code, and only\nswitch to SIGPIPE-ignored semantics where the signal has a chance to\ncause harm. Maybe run_command should automatically ignore SIGPIPE\nwhen creating a pipe for the launched command's standard input (with a\nflag to ask not to), as a rough heuristic.\n\nThere's a subtlety I'm glossing over here, which is that for commands\nthat produce a lot of output (think: \"git fetch --all\"), output may\nstill not the primary goal. I think even they should not block\nSIGPIPE, to follow the principle of least surprise in the following\ninteraction:\n\n\tgit fetch --all 2>&1 | less\n\t... one page later, get bored ...\n\tq (to quit)\n\nMost Unix programs would be killed by SIGPIPE after such a sequence,\nso I would expect git to be, too.\n\nJust my two cents,\nJonathan\n"},{"id":"184925","messageId":"20120218101020.GA9054@burratino","threadId":"29643","inReplyTo":"20120218100517.GA8998@burratino","subject":"Re: SIGPIPE handling (Re: [PATCH v3 0/3])","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2012-02-18T10:10:20Z","receivedAt":"2012-02-18T10:10:20Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Jonathan Nieder wrote:\n\n> There's a subtlety I'm glossing over here, which is that for commands\n> that produce a lot of output (think: \"git fetch --all\"), output may\n> still not the primary goal.\n\nGah. The output that goes to the terminal may not be the primary\ngoal, I mean (missing \"be\").\n\nSorry for the noise.\nJonathan\n"},{"id":"184926","messageId":"20120218102439.GA17263@sigill.intra.peff.net","threadId":"29643","inReplyTo":"20120218100517.GA8998@burratino","subject":"Re: SIGPIPE handling (Re: [PATCH v3 0/3])","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2012-02-18T10:24:40Z","receivedAt":"2012-02-18T10:24:40Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Sat, Feb 18, 2012 at 04:06:07AM -0600, Jonathan Nieder wrote:\n\n> Jeff King wrote:\n> \n> > Less robust than that is to just ignore SIGPIPE in most git programs\n> > (which don't benefit from it, and where it is only a liability), but\n> > then manually enable it for the few that care\n> \n> This seems backwards. Aren't the only places where it is just a\n> liability places where git is writing to a pipe that git has created?\n\nYes. But the problem is that those spots are buried deep within library\ncode, and are an implementation detail that the caller shouldn't need to\nknow about.\n\nBut more importantly, I see SIGPIPE as an optimization. The program on\nthe generating side of a pipe _could_ keep making output and dumping it\nnowhere. So the optimization is about telling it early that it can stop\nbothering. But that optimization is affecting correctness in some cases.\nAnd in cases of correctness versus optimization, it's nice if we can be\ncorrect by default and then optimize in places where it matters most and\nwhere we've verified that correctness isn't harmed.\n\nI also think it's simply easier to list the places that benefit from\nSIGPIPE than those that are harmed by it. But if we wanted to go the\nother way (allow by default and turn it off in a few programs), at the\nvery least all of the network programs (receive-pack. upload-pack, etc)\nshould start ignoring it completely.\n\n> We could keep the benefits of SIGPIPE (including simpler error\n> handling and lack of distracting EPIPE message) in most code, and only\n> switch to SIGPIPE-ignored semantics where the signal has a chance to\n> cause harm. Maybe run_command should automatically ignore SIGPIPE\n> when creating a pipe for the launched command's standard input (with a\n> flag to ask not to), as a rough heuristic.\n\nThat's a reasonable heuristic, but not foolproof. If I have a\nlong-running subprocess with a pipe, then SIGPIPE will be off for a long\ntime, meaning the code you thought was protected by it is not. In\npractice, I think it would work OK with our current code-base, as we\ntend to have short-lived subprocesses. You'd have to special case the\nstarting of the pager, of course, but that's not too hard.\n\n> There's a subtlety I'm glossing over here, which is that for commands\n> that produce a lot of output (think: \"git fetch --all\"), output may\n> still not the primary goal. I think even they should not block\n> SIGPIPE, to follow the principle of least surprise in the following\n> interaction:\n> \n> \tgit fetch --all 2>&1 | less\n> \t... one page later, get bored ...\n> \tq (to quit)\n> \n> Most Unix programs would be killed by SIGPIPE after such a sequence,\n> so I would expect git to be, too.\n\nBut it's quite non-deterministic whether or when git-fetch will be\nkilled. It may have written all data to the pipe. You may quit, but\nfetch still runs for a while before producing output. If you want it\nkilled, you are much better off to do so by sending it SIGINT.\n\n-Peff\n"}]}