HEAD tests

Stanislav Malyshev wrote:

I was recently working on fixing some things in HEAD in
ext/standard/strings.c and when I tried to run tests to check if my
fixes are OK I discovered that huge amount of strings tests now fail -
out of ~800 tests we have 420 fails! Most of them seem to be some
trivial matters like string/unicode mismatch or different error message
syntax but this thing makes the tests totally useless, as it is
impossible to verify fixes against them.
Is someone going to fix it?

I have actually been fixing tests, so this surprised me. I checked with
Stas on irc and it turned out to be a case of user error on his part :)
We don't have 420 failed string tests in HEAD. We still have 22 though,
plus a number of borked tests due to charset issues, so please pitch in
and work on these.

-Rasmus