We are unable to run a test suite (e.g., 100 files) using PhantomJS on Mac OS X successfully.
The test suite runs for a while but at a certain point we get the following the console:
error: PhantomJS unable to load test URL: http://127.0.0.1:2013/venus-core/41 (timeout)
error: PhantomJS unable to load test URL: http://127.0.0.1:2013/venus-core/42 (timeout)
error: PhantomJS unable to load test URL: http://127.0.0.1:2013/venus-core/43 (timeout)
I believe the issue is due to to Mac OS X limiting the number of processes to run.
I tried setting ulimit -n 1024 before running the test suite but that did not work
ulimit -n 1024
The issue is not specific just to PhantomJS.
When I try running the test suite in Chrome (e.g., venus run -t mytestsuite/), some of the tests are not able to run at all. For example, if I were to click on the 40th test out of a 100 file test suite, it does not run and throws errors. However, if I run that same test by itself it works as expected.
It definitely seems like a Max OS X specific issue.
I realize this may be a bit of work, but could you do a PR with a new branch containing a large set of simple tests, which show this issue repro'ing? That would help me find the root cause faster. If not, I will try to create a repro sometime this weekend. Thanks :)
sounds good. will do.
Submitted PR with a new branch container a large test suite: #310
Hi Roel, I'm having a hard time reproducing this. I checked out your code and loaded all tests in the test/load directory. I ran the tests both in the browser (all at once, by going to the index page), and through phantom. All 100 tests executed.
Here's the error I'm getting when trying to run the test suite in test/load:
Here are my computer settings:
Mac OS X
Processor: 2.13 GHz Intel Core 2 Duo
Memory: 4 GB 1067 MHz DDR3
@roelrz , I think the issue you are having is with the template/config. Do me a favor and try this out:
rm -rf .venus/
I was able to reproduce your issue, and resolve it with the steps above.
@rbakr i tried that but still running into the same issue
@roelrz Since I was able to reproduce this then fix it based on using a working config, I think that is a good place to begin looking. I don't currently have bandwidth but if I get time tonight I'll try to track down the bug in the config or its dependencies. Also, if you'd like to see an example of what I'm doing feel free to swing by.
@rbakr gotcha. no worries. i'll take a deeper look into the config. we can definitely sync up tomorrow during our meeting.
@sethmcl what version of venus are you using?
Was there a resolution to this issue? We're still seeing it.