New test failures on the trunk...

http.test
Running httpd in thread tid0x7f37456fb700
Running httpd in thread tid0x7f3744efa700
==== Test with ThreadLevel 0 ====
Running httpd in thread tid0x7f3744efa700
==== Test with ThreadLevel 1 ====


==== http-4.15.1 http::Event FAILED
==== Contents of test case:

    # This test may fail if you use a proxy server. That is to be
    # expected and is not a problem with Tcl.
    # With http::config -threadlevel 1 or 2, the script enters the event loop
    # during the DNS lookup, and has the opportunity to time out.
    # Increase -timeout from 3000 to 10000 to prevent this.
    set token [http::geturl //not_a_host.tcl.tk -timeout 10000 -command \#]
    http::wait $token
    set result "[http::status $token] -- [lindex [http::error $token] 0]"
    # error codes vary among platforms.

---- Result was:
timeout -- 
---- Result should have been (glob matching):
error -- couldn't open socket*
==== http-4.15.1 FAILED

Running httpd in thread tid0x7f3744efa700
==== Test with ThreadLevel 2 ====


==== http-4.15.2 http::Event FAILED
==== Contents of test case:

    # This test may fail if you use a proxy server. That is to be
    # expected and is not a problem with Tcl.
    # With http::config -threadlevel 1 or 2, the script enters the event loop
    # during the DNS lookup, and has the opportunity to time out.
    # Increase -timeout from 3000 to 10000 to prevent this.
    set token [http::geturl //not_a_host.tcl.tk -timeout 10000 -command \#]
    http::wait $token
    set result "[http::status $token] -- [lindex [http::error $token] 0]"
    # error codes vary among platforms.

---- Result was:
timeout -- 
---- Result should have been (glob matching):
error -- couldn't open socket*
==== http-4.15.2 FAILED


The comments suggest these test failures might be a false alarm, but
if that's the case we should construct a better way to deal with that.
We really want a `make test` to be a useful tool for everyone, and we
really want zero failing tests to be the expected result.

On the other hand, if the new failures indicate a real problem, we 
should fix it.

Fixed by Keith, thanks !

Here is the commit to core-8-branch: [7759882996].

Thank you all, Harald