Code review comment for lp:~tsarev/percona-server/5.1_fix_bug_924872

The Jenkins failures are too many to be confident, although none look
like they are directly related to userstat. Please get a cleaner
Jenkins run.

What is the exact source of the patches, tests, etc? I cannot find them by
casually browsing http://code.google.com/p/google-mysql-tools/ .

No need to replace "// foo" comments by "/* foo */"

I am unable to review the whole MP as-is. Maybe other reviews would
be able to, but I'd have doubts in their review quality :)

In order for me to be able to review it, there should be several
smaller MPs. One suggested split would be (might be not fully correct
due to technical details I don't know):

1) Port the Google tests that pass with the current PS code;
2) userstats_bug602047 (incorrect behavior on --repeat run);
3) user_stats.test: 3.1) include/show_stats.inc refactoring, 3.2)
Concurrency fixes (remove sleep, use end_connection debug signal) 3.3)
backup of mysql.user;
4) Code refactorings that do not change behaviour: 4.1) remove
sent_row_count_2, 4.2) remove LOCK_stats, 4.3) STATS struct
refactoring 4.4) etc...
5) Code fixes with associated tests split as necessary.

Yes, such split suggests many MPs, but I don't see any other way for
me to review it.