[Findit] Plain objects to represent and parse stack trace.

tools/findit/README

Index: tools/findit/README
diff --git a/tools/findit/README b/tools/findit/README
new file mode 100644
index 0000000000000000000000000000000000000000..1f76e62a6a3de4baa1d76ae88ec9352fb3df4288
--- /dev/null
+++ b/tools/findit/README
@@ -0,0 +1,2 @@
+Usage: python findit.py key, where key is a testcase key in cluster-fuzz. In order to use the script, a config file containing API key from cluster-fuzz must be in the same folder as the script. A new API key can be retrieved from https://cluster-fuzz.appspot.com/#help. 

[stgao, 2014/07/31 01:52:23]

How about leaving this file in the fourth patch?
The findit.py is not in this patch yet.

[jeun, 2014/07/31 18:41:14]

Done.

+Currently, the script only works for job type linux_asan_chrome_mp (query = Job: linux_asan_chrome_mp). The script works by extracting file and line information from the stacktrace in failed test case, and matching it with cls regression range. If either the regression range or the stacktrace information is not present in the testcase, the script will print out an error message. In a (very) near future, it will also look at git/svn blame for error line and possibly function names in the cls. Also, it would be able to parse and normalize stacktrace from all sources, so that the script would work for all types of job.