I actually don't know. Obviously it runs lztest because that broke
and caused things to fail. It may not run this test because of the
browser involvement which may not be available on all the build
machines?
On 2009-05-26, at 17:06EDT, Henry Minsky wrote:
OK I can update the test, but it worries me that this failed, and
the build
doesn't fail. Which ant target is being used for the build? I ran "ant
runlzunit", which fires up these tests in a browser, for swf8, and
dhtml:
<target name="runlzunit" depends="init" description="Run lzunit
test apps"
unless="skip.runlzunit">
<ant target="lztest" />
<ant target="runbrowserlzunit-swf8" />
<ant target="runbrowserlztest-swf8" />
<ant target="runbrowserlzunit-dhtml" />
<ant target="runbrowserlztest-dhtml" />
</target>
On Tue, May 26, 2009 at 4:55 PM, P T Withington
<[email protected]>wrote:
I'm pretty sure the test is conditionalized in smokecheck to work
around
these bugs. Basically in swf8 if you ask
NaN === NaN
it returns true. ECMA calls for it to return false, because NaN is
not
supposed to be comparable to anything else including itself.
My guess: Ben copied the tests from smokecheck, smokecheck has
since been
updated, but these 'lztest' copies have not.
On 2009-05-26, at 14:05EDT, Henry Minsky wrote:
I was testing a change and decided to run the "ant runlzunit"
suite, and
it
returned a failure in the test
test/lztest/lztest-selftest.lzx
I ran it manually with the debugger, and got this
http://localhost:8080/trunk6/test/lztest/lztest-selftest.lzx?lzr=swf8&debug=true
ERROR: In suite JavaScript Semantics, test 9, failed assertion:
NotSame
undefined expected anything but NaN, got
ERROR: In suite Tautologies, test 9, failed assertion: NotSame
undefined
expected anything but NaN, got
ERROR: Failed 2 suites, JavaScript Semantics, Tautologies
Any idea if this is a real bug ? Seem odd it wasn't caught by a
nightly
run
of the test suite
--
Henry Minsky
Software Architect
[email protected]
--
Henry Minsky
Software Architect
[email protected]
