JDK-8141527 : AddRemoveTest tests fail as second request to jnlp not getting to server, but resource requests going through fine
  • Type: Bug
  • Component: deploy
  • Sub-Component: webstart
  • Affected Version: 7u95,8u72
  • Priority: P2
  • Status: Closed
  • Resolution: Fixed
  • Submitted: 2015-11-05
  • Updated: 2016-04-27
  • Resolved: 2015-11-19
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
7u101Fixed 8u101Fixed 9Fixed
Related Reports
Duplicate :  
Duplicate :  
Relates :  
Product(s) tested: 7u95 b06
OS/architecture: Windows x64_64
Error log file (if present):  http://aurora.ru.oracle.com/functional/faces/RunDetails.xhtml?names=1143706.STT-DEPLOYMENT-1
Reproducible: Always
Reproducible on machine: spb23094.ru.oracle.com

Is it a Regression: Yes
[If Regression] Regression introduced in release/build: 7u95 b01
Test result on the last GAed release for this train: Pass (7u91 b15)
Is it a platform specific issue: No
Exception/Error from Log:
Test run log location:
Steps to recreate: 
1. Install jre 7u95 on win7/x86. 
2. Add deployment.trace.level=all to deployment.properties 
3. Add http://stt-13.ru.oracle.com to ESL
4. Switch java console on. 
5. launch javaws -import http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/jnlp/CacheViewer.jnlp
The following two entries could be seen in the trace log at AppData/LocalLow/Sun/Java/Deployment/log 

	http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/jnlp/CacheViewer.jnlp : 200
	http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/classes/CacheViewerMFSandbox.jar : 200

The corresponding entries could be seen in Cache Viewer 'Resources' view as well.

6. launch javaws http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/jnlp/CacheViewer.jnlp
Expected entries in trace log:
	http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/jnlp/CacheViewer.jnlp : 304
	http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/classes/CacheViewerMFSandbox.jar : 304
Actual :
	http://stt-13.ru.oracle.com/newroot/testsuites/180_user_ws/vivek/AddRemove/classes/CacheViewerMFSandbox.jar : 304
	The entry for jnlp file is not seen

RULE AddRemoveTest::testAddRemove01 any any
RULE AddRemoveTest::testAddRemove02 any any
RULE AddRemoveTest::testAddRemove03 any any
RULE AddRemoveTest::testAddRemove04 any any
RULE JarDownloadTest::testEager1 any any
RULE JarDownloadTest::testEager2 any any
RULE JarDownloadTest::testEager3 any any
RULE JarDownloadTest::testEagerConcurrent any any
RULE JarDownloadTest::testEagerNoCodeBase any any
RULE JarDownloadTest::testIndexing1 any any
RULE JarDownloadTest::testIndexing2 any any
RULE JarDownloadTest::testIndexing3 any any
RULE JarDownloadTest::testIndexing4 any any
RULE JarDownloadTest::testIndexing5  any any
RULE JarDownloadTest::testLazy1 any any
RULE JarDownloadTest::testLazy2 any any
RULE JarDownloadTest::testLazy3  any any
RULE JarDownloadTest::testLazy4  any any
RULE JarDownloadTest::testLazyConcurrent any any

>Does it affect JDK 6? No, it doesn't affect jdk6.

Does it affect JDK 6?

Verified with http://jre.us.oracle.com/java/re/jdk/9/jdk9-client/b94_2015-11-24-0302_3911/bundles/windows-i586/

Let's take it 8u72 only (PSU16_01). It make sense to take it in 7u95 as far as JDK-8098826 has been pushed in.

which fix caused this regression ?

SQE OK to take this

This is a serious regression and as a result the request is being made for inclusion in the CPU for JDK8 and CPSU for JDK 7. See the justification above and the crucible review for further details.

Justification : this changeset fixes major bug, that prevents update of JNLP files Risk Analysis:��� low risk Webrev: https://java.se.oracle.com/code/cru/CR-JDK8UCPU-277 Testing (done/to-be-done) : I used a testcase described in JDK-8141527 Back ports (done/to-be-done) : it was backported to jdk7 and forwadported to jdk9 FX Impact : n/a Fix For Release : JDK 7, 8

Resources were marked as downloaded on very first request even if it was provided from cache, so if JNLP was requested with doDownload=false, it will be marked as downloaded and for all next requests it will be provided from cache, even if next calls have doDownload=true. To fix it, resources will be marked as downloaded only when they are really downloaded via network.