<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 4/30/15 7:52 AM, Leonid Mesnik
wrote:<br>
</div>
<blockquote cite="mid:55421795.2080401@oracle.com" type="cite">
<meta http-equiv="content-type" content="text/html; charset=utf-8">
Hi<br>
<br>
Please review this test fix:<a moz-do-not-send="true"
href="https://bugs.openjdk.java.net/browse/JDK-8079134"><br>
https://bugs.openjdk.java.net/browse/JDK-8079134</a><br>
<br>
Webrev:<br>
<a moz-do-not-send="true"
href="http://cr.openjdk.java.net/%7Elmesnik/8079134/webrev.00/">http://cr.openjdk.java.net/~lmesnik/8079134/webrev.00/</a><br>
<br>
Summary:<br>
The temporary workaround for GC conflicting options should be
removed. The jtreg tag '@requires' should be used to resolve GC
conflicts.<br>
</blockquote>
Hi Leonid,<br>
<br>
After talking with Dmitry Fazunenko and David Holmes, it sounds like
this isn't a good idea.<br>
<br>
The issue I'm trying to solve is that there are a set of GC
regression tests that should not be run unless the JVM under test
implements G1, or CMS, or ParallelGC. They shouldn't fail or error,
there should be no attempt to run them. For example, the Embedded
JVMs should not run some of the GC tests.<br>
<br>
My understanding is that @requires doesn't handle this case yet.<br>
<br>
A possible complication is that these tests tend to use
ProcessBuilder, not @run, so annotations don't contain everything.<br>
<br>
Thanks,<br>
<br>
- Derek<br>
</body>
</html>