On Mar 4, 2013 5:58 PM, "Frédéric THOMAS" <webdoubl...@hotmail.com> wrote: > > Hi guys, > > Well, If I understand well, Om’s got all the tests passed and if it doesn’t pass constantly for me, it’s machine dependent, right ? why not but I wonder, if it is like that, why another test wouldn’t failed for the same reason for Om or someone else ? > > Another point, even though I update the FP version I use to execute the tests, I’ve got other failures but the bubble charts one still remains... that’s a constant, do you still think it is machine dependent ? it could maybe but I’m not sure at the end, so to clarify the things, I’m going to setup another machine and execute the tests again let’s see. > > -Fred >
Just to be clear, the tests pass on my local machine but they still fail on the VM. Not sure how to proceed. I will probably connect to the debugger and see. Thanks, Om > > From: Alex Harui > Sent: Monday, March 04, 2013 11:51 PM > To: Om > Cc: dev@flex.apache.org > Subject: Re: Mustella CI server > > Are you using the same version of player? Frederic Thomas has one failure on Windows in BubbleChart. In your cases and his, the differences look like slight differences in the floating point math. I wonder if there are machine-dependent issues that cause floating point math to do slightly different things and if we can control them. > > > On 3/4/13 2:32 PM, "Om" <wlmailhtml:bigosma...@gmail.com> wrote: > >> On my local machine, I tried this: >> >> $ ./mini_run.sh tests/components/Charts/Area >> >> All 19 tests pass. >> >> Thanks, >> Om >> >> On Mon, Mar 4, 2013 at 12:11 PM, Alex Harui <wlmailhtml:aha...@adobe.com> wrote: >>> >>> Om, do these three bitmap compares pass for you locally? >>> >>> >>> >>> On 3/4/13 11:48 AM, "Om" <wlmailhtml:bigosma...@gmail.com < http://bigosma...@gmail.com> > wrote: >>> >>>> On Mon, Mar 4, 2013 at 3:33 PM, Alex Harui <wlmailhtml:aha...@adobe.com< http://aha...@adobe.com> > wrote: >>>>> >>>>> >>>>> >>>>> >>>>> > I see the swf files under the assets folder >>>>> > (ex. C:\apacheflex\mustella\tests\MarshallPlan\SWFs\assets\UntrustedApp.swf) >>>>> > but it is not loadable via >>>>> > http://localhost/MarshallPlan/SWFs/assets/UntrustedApp.swf >>>>> > >>>>> > Are there any settings I need to change in httpd? >>>>> > >>>>> > >>>>> No idea. Can you load it if you run that test locally? >>>>> >>>> >>>> I had a bad entry in the httpd conf. I fixed that and it fixed all the timeout issues. >>>> >>>>> >>>>> >> Most of the timeout errors will go away when we get the sub-swfs >>>>> >> to load. The focusIn errors might be related to the bitmap compare >>>>> >> failures, so let's worry about them last. If there is some font or layout >>>>> >> issue then a focusIn will get missed if a mouse hit was supposed to focus a >>>>> >> control but now misses the control. >>>>> >> >>>>> >> >>>>> > I hope so. >>>>> > >>>>> > So, let's take a look at some bad.png files from the bitmap failures. >>>>> >> >>>>> >> >>>>> > Will it be easier if you took a peek into the windows box up on Azure where >>>>> > I ran these tests? I can PM the login details if you would like. >>>>> > >>>>> It would be easier for me if you took a look at the differences and sent two >>>>> or three bad.png files that represent patterns you see in the diffs. >>>>> >>>> >>>> I have attached results of 3 tests. The first two dont have any visible differences. In the third one (AreaSeries_sameAxis_diffRenderers.png.bad.png) there is a pretty serious error. An extra axis renderer shows up on the top of the chart. >>>> >>>> >>>> >>>> >>>> Thanks, >>>> Om >>>> > > -- > Alex Harui > Flex SDK Team > Adobe Systems, Inc. > http://blogs.adobe.com/aharui
