README
¶
Autotest results parser
[TOC]
This package defines the parse sub-command of phosphorus, used to extract
test results from the test result summary file (status.log) and the exit code
file (.autoserv_execute) created by the autotest harness inside the results
directory. The parsing is a simpler version of the one done by tko/parse.
status.log format
The status.log file consists of test event lines and miscellaneous stderr
output. The test event lines are nested according to the subtest hierarchy.
Each event line
-
starts with a number of tab characters represeting the nesting depth
-
consists of strings separated by the tab character
-
has the following format:
status testDirectory testName key1=value1 key2=value2 ... comments and/or failure reason
Status can be "START", "INFO", test result outcome ("GOOD", "FAIL", "WARN" etc) or "END " + test result outcome ("END GOOD", etc).
The default (undefined) value for testDirectory and testName is "----".
Each subtest block starts with a "START" event line and ends with an "END ..." event line.
A typical status.log file
START test1 test1 timestamp=1561420800 localtime=Jun 26 00:00:00
START test1_subtest test1_subtest timestamp=1561420801 localtime=Jun 26 00:00:01
START ---- ---- timestamp=1561420802 localtime=Jun 26 00:00:02
GOOD ---- sysinfo.before timestamp=1561420803 localtime=Jun 26 00:00:03
END GOOD ---- ---- timestamp=1561420804 localtime=Jun 26 00:00:04
FAIL test1_subtest test1_subtest timestamp=1561420805 localtime=Jun 26 00:00:05 The test failed because reasons.
END FAIL test1_subtest test1_subtest timestamp=1561420805 localtime=Jun 26 00:00:05
FAIL test1 test1 timestamp=1561420806 localtime=Jun 26 00:00:06 test1_subtest failed.
END FAIL test1 test1 timestamp=1561420806 localtime=Jun 26 00:00:06
START test2 ---- timestamp=1561420800 localtime=Jun 26 00:00:07
INFO test2 ---- timestamp=1561420805 localtime=Jun 26 00:00:08 This test has no name.
WARN test2 ---- timestamp=1561420806 localtime=Jun 26 00:00:09 Something suspicious happened.
END WARN test2 ---- timestamp=1561420806 localtime=Jun 26 00:00:09
START ---- test3 timestamp=1561420800 localtime=Jun 26 00:00:10
INFO ---- test3 timestamp=1561420810 localtime=Jun 26 00:00:10 This test has no directory.
ERROR ---- test3 timestamp=1561420811 localtime=Jun 26 00:00:11 Unhandled AutoservRunError: command execution error
* Command:
...
Exit status: 1
Duration: 0.00010000000
stderr:
...
[0626/000011.000000:ERROR:update_engine_client.cc(519)] Error checking for update.
Traceback (most recent call last):
...
File "/usr/local/autotest/server/hosts/ssh_host.py", line 268, in _run
raise error.AutoservRunError("command execution error", result)
AutoservRunError: command execution error
END ERROR ---- test3 timestamp=1561420811 localtime=Jun 26 00:00:11
Parsing status.log
The goal of this parser is to report a flat list of (sub)test cases together with a simple pass/fail verdict and comment on what went wrong. In the current implementation the verdict is obtained from the "END ..." test event lines and the human readable summary is obtained from "FAIL"/"ERROR"/"WARN"/etc test event lines. "INFO" and "GOOD" lines are ignored. Initial tab characters are ignored - the nesting is determined via keeping track of "START" and "END ..." event lines.
Directories