[
https://issues.jboss.org/browse/ARQGRA-461?page=com.atlassian.jira.plugin...
]
Lukáš Fryč commented on ARQGRA-461:
-----------------------------------
Hey Alex, this behavior is expectable as {{parallelStream()}} causes the collection to
evaluate outside of parental thread.
You will end up with same issues for lot of Arquillian extensions.
----
In fact I think you won't get much gain with {{parallelStream()}} since the Selenium
itself is likely serializing requests.
*Idea:* In order to speed up bulk operations, you can use JavaScript-based retrieval logic
(JavaScript can return JSON of table structure that you can parse on server-side).
Running Graphene tests and use parallelStream method makes test fail
with java.lang.IllegalStateException: The Graphene runtime isn't initialized.
Exception.
-------------------------------------------------------------------------------------------------------------------------------------------------------------
Key: ARQGRA-461
URL:
https://issues.jboss.org/browse/ARQGRA-461
Project: Arquillian Graphene
Issue Type: Bug
Reporter: Alex Soto
Priority: Minor
Labels: java8,, threads
Running Graphene tests and use parallelStream method makes test fail with
java.lang.IllegalStateException: The Graphene runtime isn't initialized. Exception.
Basically the problem is when you want to get information about tables. You can iterate
over elements in a foreach fashion, but you can also use new Java 8 stream API.
This works:
ession.findElements(xpath("//table/tbody/tr/td/span[@class='title']")).stream().map(WebElement::getText).collect(Collectors.toList());
But if table is big you will try to improve parsing speed by adding parallelSteam
(session.findElements(xpath("//table/tbody/tr/td/span[@class='title']")).parallelStream().map(WebElement::getText).collect(Collectors.toList());)
And in this case an exception is thrown.
Internally parallelStream uses a pool of threads so probably Graphene instance is tight
to thread of test case, so the threads used by parallelStream does not contains a valid
instance of Graphene.
Of course this is not blocking nor a big problem in the sense that you can use stream()
without any problem, and moreover Java 8 is not spread yet, but it may be a problem in the
future.
--
This message was sent by Atlassian JIRA
(v6.3.1#6329)