Re: [Webtest] java.lang.StackOverflowError on infinite redirects - Marc Guillemot - com.canoo.lists.webtest

3 messages in com.canoo.lists.webtestRe: [Webtest] java.lang.StackOverflow...

From	Sent On	Attachments
Andrew Ryan	18 Nov 2007 16:13
Marc Guillemot	21 Nov 2007 06:14
Andrew Ryan	21 Nov 2007 07:04

Subject:	Re: [Webtest] java.lang.StackOverflowError on infinite redirects
From:	Marc Guillemot (mgui...@yahoo.fr)
Date:	11/21/2007 06:14:42 AM
List:	com.canoo.lists.webtest

Hi Andrew,

which WebTest build do you use? If I correctly remember, this has been fixed in HtmlUnit for some time and it should now accept the same number of redirections as FF does per default (I think 20) before to throw an exception.

Marc. -- Blog: http://mguillem.wordpress.com

Andrew Ryan wrote:

A bug in the app I'm testing created an infinite redirect (302's) loop between 2 pages in one test we've written.

In FF, this creates the error "Firefox has detected that the server is redirecting the request for this address in a way that will never complete." That's friendly and informative. But in webtest (R_1638), it seems that webtest keeps trying until it blows up with a StackOverflowError: (100's of) [invoke] INFO (org.apache.commons.httpclient.HttpMethodDirector) - Redirect requested but followRedirects is disabled [invoke] INFO (org.apache.commons.httpclient.HttpMethodDirector) - Redirect requested but followRedirects is disabled [invoke] INFO (org.apache.commons.httpclient.HttpMethodDirector) - Redirect requested but followRedirects is disabled INFO (com.canoo.webtest.engine.WebClientContext) - Responses restored

BUILD FAILED /home/andrewr/tests/functional/webtest/webtest.xml:163: The following error occurred while executing this line: /home/andrewr/tests/functional/webtest/myproj/build.xml:63: The following error occurred while executing this line: /home/andrewr/tests/functional/webtest/myproj/AllUseCasesTestSuite.xml:1 9: The following error occurred while executing this line: java.lang.StackOverflowError

It would seem that max redirects should be something we can control, and that "too many redirects" should mean a test failure, not the blowing up of an entire run.

--andrew