Sebastian, is Firebug is completely disabled? Firebug tends to cause some odd behaviors. It sounds like you already have disabled it (via safe mode?), but I want to be sure.
-Damien On Jul 9, 2010, at 4:09 PM, Sebastian Cohnen wrote: > On 09.07.2010, at 21:55, J Chris Anderson wrote: > >> >> On Jul 9, 2010, at 12:44 PM, Sebastian Cohnen wrote: >> >>> OS X 10.6.4, Erlang R13B04, Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; >>> en-US; rv:1.9.2.6) Gecko/20100625 Firefox/3.6.6 >>> >> >> Have you ensured that you've cleared the browser cache? > > I run Firefox in safemode, when running the test suite. but running it in > normal mode (with firebug & co) I always clear the cache. so yes, I ensured > that the cache was empty. (I begin to dislike a browser-based integration > test suite more and more). > >> Also, with Couch running on 5984, what do you get if you run this from the >> source directory: >> >> test/javascript/run (I get some failures like this, no biggie, but changes >> passes) > > "ok 11 changes" looks okay. but after some tests (e.g. "ok 26 http") it > simply freezes. I've waited a few minutes, killed both couch and the test and > tried again. after the 10th run, the test suite runs through (with "not ok 16 > cookie_auth TypeError: ddoc is null" and "not ok 43 replication > ReferenceError: $ is not defined"). > > > Sebastian > >> Chris >> >>> signature, make check, all okay. test suite: all okay, but the changes test >>> however always fails (Exception raised: "timeout: bar-only") when I run all >>> test, I've tried about 10 times now. when I run this test in isolation it >>> is successful most of the time, but not always 3 fails out of 10. I have >>> not yet pinpoint the responsible part of the test, that causes the failure, >>> but I notice, that for long periods of time the test simply does nothing (I >>> see "[info] [<0.8991.0>] 127.0.0.1 - - 'GET' >>> /test_suite_db/_changes?feed=continuous&timeout=500 200" and then nothing >>> for about a minute). >>> >>> -1 (unless someone can tell me, what I could have done wrong) > > *snip* >
