As suggested, replaced relevant uses of my_fopen with fopen (and close)
Tested on HPUX where it was reproducable with test innodb_bug30919

Problems occur after killing threads on Windows
Get rid of the timeout threads, implement simple timer in wait_any_timeout()

Add =1 to --abort-on-error argument

mtr patch ported to v2

For some reason it usually picks up mysqltest
Wait .1s and then see if a server has died
Change from first commit: label on separate line

Wrong argument order, fixed

Searches for my_safe_process binary too early
Put this into a sub() and call it after examining options (incl. --help)

Seen in a few tests after 43418
Add code in extract_warning_lines() to skip this part

Checking is done before server is terminated
Adds post processing of server logs if --valgrind
Also had to remove --quiet option to valgrind

Actually, fails on 'perl' in any while
Fixed essentially the same way as for append_file

Some output is written, some is not
Finally concluded it's a Perl bug: after running with parallel threads
  for a while, print suddenly ignores all but the first argument.
Workaround: concatenate all the arguments into one, except in output that
  only comes before we start running tests

Prepend "Completed", "Timeout" or "Too many failed" to summary line

There's no need for --verbose for the mysqltest in check-warnings, remove it

Knowledge of no SSL support is not used
Skip tests the same way e.g. innodb tests are
Does not refer to have_ssl_communication.inc, 
     will add this when merging to 6.0-codebase

Don't print entire log, but use extract_server_log() introduced by 46007

Test batches may be terminated too early
Avoid counting exp-fail tests

Use "#ifdef", not plain "#if".