On 03/27/2013 02:05 PM, Wendall Cada wrote:
In 1.3.0, there is a new part of the test suite to run the javascript
tests from the command line. I'm running into various issues on
different hardware/OS configurations. Mostly, tests hanging or timing
out and failing. These are really hard to troubleshoot, as they all
pass just fine if run individually.
What I'm experimenting with today is rewriting how the tests are
implemented to be run one at a time from a loop in bash, versus a loop
in javascript. I think the failures I'm running into are improper
setup/teardown. There may be an issue with rapid delete and adding a
db, or rapidly starting and stopping couchdb, but I think this is not
what's happening in my failures.
The nature of spidermonkey doesn't allow for spawning threads, or
sandboxing, etc, so it's hard looking at the test suite to see how I
can improve running all tests. I think it's far better to have the
setup spawn a new interpreter for each test. Tear down will kill the
interpreter.
Wendall
An update here. I spent some time with this yesterday.
There are two problems here I'd like to address.
1. When running `make check`, the entire javascript test suite runs in a
single interpreter thread. This means that if there are any broken
tests, then entire suite bails, as it can crash the interpreter. As I
said, to address this, I simply want to refactor the test suite to spawn
a new interpreter thread for every test. This is working well for me
with my quick work yesterday, and allows me to put a small delay in
between tests to ensure proper tear down from the previous test.
2. Some of the tests are written in a way that can cause a race
condition in the test itself. Quite a few places exist in the current
javascript tests where infinite loops can get generated. I simply want
to re-factor several of these tests to work as expected and avoid the
race conditions entirely.
One additional thought I had was to mimic the output of the etap tests a
bit more. Show the entire path to the test file being run, with a status
at the end of the line. Change the individual test run to include the
full path to the test to be run. Additionally, the run scripts would
benefit from having some help output. I'm in the process of gathering
notes for the documentation as well.
My thinking is that if the test suites are similar, it will be helpful
to new developers in troubleshooting issues.
If nobody has any objections, I'd like to get started on this work. I
have a local branch in progress for this already, but it's mostly just a
scratch space for testing so far.
Wendall
