[prev in list] [next in list] [prev in thread] [next in thread] 

List:       cactus-dev
Subject:    [jira] Commented: (CACTUS-256) ServletTestRunner should build test
From:       "Keith D Gregory (JIRA)" <cactus-dev () jakarta ! apache ! org>
Date:       2008-08-16 14:05:44
Message-ID: 1720476022.1218895544444.JavaMail.jira () brutus
[Download RAW message or body]


    [ https://issues.apache.org/jira/browse/CACTUS-256?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12623124#action_12623124 \
] 

Keith D Gregory commented on CACTUS-256:
----------------------------------------

Work and Real Life(tm) has kept me from this, but look for something in the \
September/October timeframe.

> ServletTestRunner should build test redirector URL from incoming request, not use \
>                 CACTUS_CONTEXT_URL_PROPERTY unless instructed
> -------------------------------------------------------------------------------------------------------------------------------
>  
> Key: CACTUS-256
> URL: https://issues.apache.org/jira/browse/CACTUS-256
> Project: Cactus
> Issue Type: Improvement
> Components: Framework
> Affects Versions: 1.7.2, 1.8
> Reporter: Keith D Gregory
> 
> The first time ServletTestRunner gets invoked, it builds the context URL for the \
> actual test requests, and saves it in the system property "cactus.contextURL" \
> (BaseConfiguration.CACTUS_CONTEXT_URL_PROPERTY). All subsequent requests then use \
> that property value, ignoring the actual request URL. This is bad on several \
> levels. First, because it's impolite to set system-wide properties in a shared \
> execution environment.  Second, and more important, because it prevents running \
> tests in an environment that uses virtual homes and extracts information from the \
> request URL. If you want to have particular behavior for requests to \
> "foo.example.com", and have already run tests for "bar.example.com", you have to \
> restart your container. Or spend a couple hours with a debugger and the Cactus \
> source code, trying to figure out why your tests are failing (no, I'm not bitter \
> :-)). And third (just throwing this in for good measure), because it breaks the \
> JUnit premise that tests execute in isolation, and are not dependent on execution \
> order. Personally, I can't see a reason for caching this value. It's not terribly \
> expensive to create it anew for each run. For those people who do want to have \
> their first test influence all subsequent tests, it could be a request parameter \
> ("rememberThisHost=yes").

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: cactus-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: cactus-dev-help@jakarta.apache.org


[prev in list] [next in list] [prev in thread] [next in thread]