JDK-8131321 : 8u60 Windows 64-bit packager - install succeeds but application fails to start
  • Type: Bug
  • Component: deploy
  • Sub-Component: packager
  • Affected Version: 8u60
  • Priority: P2
  • Status: Closed
  • Resolution: Fixed
  • OS: windows_8
  • CPU: x86
  • Submitted: 2015-07-02
  • Updated: 2017-07-25
  • Resolved: 2015-07-21
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.
JDK 8 JDK 9
8u60Fixed 9Fixed
Related Reports
Relates :  
Relates :  
Description
FULL PRODUCT VERSION :
8u60 b21 early access - 64-bit Windows JDK

ADDITIONAL OS VERSION INFORMATION :
Windows 8 (6.2.8400)

A DESCRIPTION OF THE PROBLEM :
We use the JavaFX native packager to produce native install bundles for Windows, Mac, and Linux versions of our product. We invoke the packager from an ant build script via the packager ant tasks. Our ant script and overall build process has not changed and continues to produce working bundles on 8u45 and earlier JDK releases. We create all of our Windows bundles from a 64-bit Windows 7 build machine. We use this Windows 7 build machine to produce both 64-bit and 32-bit Windows bundles (using a 64-bit JDK for the former and a 32-bit JDK for the latter, of course). While 8u60 continues to produce working 64-bit bundles for Windows 7, when the 8u60-generated native installer is run on Windows 8, the installer runs to completion, but the application fails to start, with the following stack trace (would attach a screenshot of the popup error dialog, but this form doesn't allow attachments):

Problem signature:
  Problem Event Name:	APPCRASH
  Application Name:	Siterise.exe
  Application Version:	0.0.0.0
  Application Timestamp:	5589aad6
  Fault Module Name:	packager.dll
  Fault Module Version:	0.0.0.0
  Fault Module Timestamp:	5589aae7
  Exception Code:	c0000005
  Exception Offset:	0000000000021582
  OS Version:	6.2.8400.2.0.0.256.48
  Locale ID:	1033
  Additional Information 1:	bbbb
  Additional Information 2:	bbbb0b3246c50470cc13a9e412ede897
  Additional Information 3:	cc16
  Additional Information 4:	cc162f5b589719b068380f674f5b753a

REGRESSION.  Last worked in version 8u45

ADDITIONAL REGRESSION INFORMATION: 
In addition to the above problem, 8u60 also breaks 32-bit Windows XP support. Yes, I know that XP is no longer officially supported by Oracle, but it was working for our product just fine up to and including 8u45. I mention this in the hopes that both issues are somehow related and that a fix will allow us to continue to support our Windows XP customers, even if Oracle no longer supports that platform. 

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
I can't easily bundle up our entire application and build but I will work with whoever looks at this bug to help him or her reproduce it.


REPRODUCIBILITY :
This bug can be reproduced always.

CUSTOMER SUBMITTED WORKAROUND :
An unrelated rant: I used to have JavaFX JIRA access and the ability to submit bugs there. That has been productive for me and I believe for the JavaFX team, but that access has now been revoked apparently. I am now presented with an OpenJDK login screen, which doesn't take my credentials, and when I try the "forgot password" link it says an email has been sent yet the email never arrives. 


Comments
Verification: Packaged in Windows 7 64 bit using JDK 9 178 build. Deployed on Windows 7 64bit, Windows 8 64bit, Windows 8.1 64bit, Windows 10 64bit. Applications runs successfully on deployment. fix verified in java version "9" Java(TM) SE Runtime Environment (build 9+178) Java HotSpot(TM) 64-Bit Server VM (build 9+178, mixed mode)
25-07-2017

Don't know and yes we should. I'm not sure how this ever worked, so I suspect it was a recent change or just never found. I don't really care enough to go into version control and use the blame feature.
15-07-2015

Why are we shipping a debug compiled version? Shouldn't we be shipping a release version?
15-07-2015

I found that the compile options for launcher.exe are wrong in buildSrc/win.gradle: WIN.launcher.ccFlags = ["/nologo", "/W3", "/EHsc", "/D_WINDOWS", "/DUNICODE", "/D_UNICODE", "/DWIN32", "/D_LITTLE_ENDIAN", "/DWIN32_LEAN_AND_MEAN", "/I$JDK_HOME/include", "/I$JDK_HOME/include/win32", ccDebugFlags].flatten(); This is missing: "/MT���, So we are linking to the debug Microsoft runtime and msvcp120d.dll isn���t found. However, this isn't shown as an error and there still appears to be issues with the Enesable installer on Windows 8.1 and the launcher still doesn���t work when I fix the issue. Still investigating...
15-07-2015