Community
Participate
Working Groups
Sample Error Log !ENTRY org.eclipse.ui.browser 4 0 2013-04-18 08:44:05.419 !MESSAGE Launching C:\Program Files (x86)\Mozilla Firefox\firefox.exe has failed. !STACK 0 java.io.IOException: Cannot run program "C:\Program": CreateProcess error=2, Das System kann die angegebene Datei nicht finden at java.lang.ProcessBuilder.start(ProcessBuilder.java:1042) at java.lang.Runtime.exec(Runtime.java:615) at java.lang.Runtime.exec(Runtime.java:448) at java.lang.Runtime.exec(Runtime.java:345) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.openBrowser(MozillaBrowser.java:94) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.run(MozillaBrowser.java:165) Caused by: java.io.IOException: CreateProcess error=2, Das System kann die angegebene Datei nicht finden at java.lang.ProcessImpl.create(Native Method) at java.lang.ProcessImpl.<init>(ProcessImpl.java:288) at java.lang.ProcessImpl.start(ProcessImpl.java:133) at java.lang.ProcessBuilder.start(ProcessBuilder.java:1023) ... 5 more
What's the new behaviour (as opposed to what happened before)? PW
http://www.oracle.com/technetwork/java/javase/7u21-relnotes-1932873.html#jruntime
Created attachment 231340 [details] Log about this error It happens when I try to use Firefox as external browser. It doesn't happen when I use Google Chrome or Internet Explorer. This bug is caused in both Windows 7 and Windows 8 64-bit and Firefox is 32-bit.
!ENTRY org.eclipse.wst.common.project.facet.core 4 0 2013-05-21 23:04:24.554 !MESSAGE Project facet jst.ejb has not been defined. It is used in plugin org.eclipse.jst.server.core. !SESSION 2013-05-22 16:16:09.142 ----------------------------------------------- eclipse.buildId=M20130204-1200 java.version=1.7.0_21 java.vendor=Oracle Corporation BootLoader constants: OS=win32, ARCH=x86_64, WS=win32, NL=es_CO Framework arguments: -product org.eclipse.epp.package.jee.product Command-line arguments: -os win32 -ws win32 -arch x86_64 -product org.eclipse.epp.package.jee.product !ENTRY org.eclipse.egit.ui 2 0 2013-05-22 16:17:11.393 !MESSAGE Warning: EGit couldn't detect the installation path "gitPrefix" of native Git. Hence EGit can't respect system level Git settings which might be configured in ${gitPrefix}/etc/gitconfig under the native Git installation directory. The most important of these settings is core.autocrlf. Git for Windows by default sets this parameter to true in this system level configuration. The Git installation location can be configured on the Team > Git > Configuration preference page's 'System Settings' tab. This warning can be switched off on the Team > Git > Confirmations and Warnings preference page. !ENTRY org.eclipse.egit.ui 2 0 2013-05-22 16:17:11.408 !MESSAGE Warning: The environment variable HOME is not set. The following directory will be used to store the Git user global configuration and to define the default location to store repositories: 'C:\Users\Freddy'. If this is not correct please set the HOME environment variable and restart Eclipse. Otherwise Git for Windows and EGit might behave differently since they see different configuration options. This warning can be switched off on the Team > Git > Confirmations and Warnings preference page. !ENTRY org.eclipse.wst.server.core 4 0 2013-05-22 16:18:28.522 !MESSAGE Could not publish to the server. !STACK 0 java.lang.NullPointerException at org.eclipse.wst.server.http.core.internal.HttpServerBehaviour.publishModule(HttpServerBehaviour.java:95) at org.eclipse.wst.server.core.model.ServerBehaviourDelegate.publishModule(ServerBehaviourDelegate.java:1091) at org.eclipse.wst.server.core.model.ServerBehaviourDelegate.publishModules(ServerBehaviourDelegate.java:1183) at org.eclipse.wst.server.core.model.ServerBehaviourDelegate.publish(ServerBehaviourDelegate.java:987) at org.eclipse.wst.server.core.model.ServerBehaviourDelegate.publish(ServerBehaviourDelegate.java:774) at org.eclipse.wst.server.core.internal.Server.publishImpl(Server.java:3153) at org.eclipse.wst.server.core.internal.Server$PublishJob.run(Server.java:345) at org.eclipse.core.internal.jobs.Worker.run(Worker.java:53) !ENTRY org.eclipse.ui.browser 4 0 2013-05-22 16:19:18.585 !MESSAGE Launching C:\Program Files (x86)\Mozilla Firefox\firefox.exe has failed. !STACK 0 java.io.IOException: Cannot run program "C:\Program": CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessBuilder.start(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.openBrowser(MozillaBrowser.java:94) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.run(MozillaBrowser.java:165) Caused by: java.io.IOException: CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessImpl.create(Native Method) at java.lang.ProcessImpl.<init>(Unknown Source) at java.lang.ProcessImpl.start(Unknown Source) ... 6 more !ENTRY org.eclipse.ui.browser 4 0 2013-05-22 16:20:47.830 !MESSAGE Launching C:\Program Files (x86)\Mozilla Firefox\firefox.exe has failed. !STACK 0 java.io.IOException: Cannot run program "C:\Program": CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessBuilder.start(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.openBrowser(MozillaBrowser.java:94) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.run(MozillaBrowser.java:165) Caused by: java.io.IOException: CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessImpl.create(Native Method) at java.lang.ProcessImpl.<init>(Unknown Source) at java.lang.ProcessImpl.start(Unknown Source) ... 6 more !ENTRY org.eclipse.ui.browser 4 0 2013-05-22 16:22:40.087 !MESSAGE Launching C:\Program Files (x86)\Mozilla Firefox\firefox.exe has failed. !STACK 0 java.io.IOException: Cannot run program "C:\Program": CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessBuilder.start(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at java.lang.Runtime.exec(Unknown Source) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.openBrowser(MozillaBrowser.java:94) at org.eclipse.ui.internal.browser.browsers.MozillaBrowser$BrowserThread.run(MozillaBrowser.java:165) Caused by: java.io.IOException: CreateProcess error=2, El sistema no puede encontrar el archivo especificado at java.lang.ProcessImpl.create(Native Method) at java.lang.ProcessImpl.<init>(Unknown Source) at java.lang.ProcessImpl.start(Unknown Source) ... 6 more
We have to fix this for Kepler and not just for opening the Firefox browser: almost all programs on Windows are in "Program Files" or "Program Files (x86)" i.e. have a space in it. Any code that uses java.lang.Runtime.exec(String) to call a program from there, will fail on the latest Oracle JREs, both 32- and 64-bit. We have to make a pass over the whole code base.
Runtime.exec(String) specifies: * <p>More precisely, the <code>command</code> string is broken * into tokens using a {@link StringTokenizer} created by the call * <code>new {@link StringTokenizer}(command)</code> with no * further modification of the character categories. The tokens * produced by the tokenizer are then placed in the new string * array <code>cmdarray</code>, in the same order. So far, the implementation for Windows tried to be smart and was able to pass the right string to the native API. In the latest updates it now only treats the first element in the command array as the executable. The general fix is to use the Runtime.exec(String[] cmdArray) API where cmdArray[0] = the command cmdArray[1..] = the parameters split with the StringTokenizer In cases where we only have one parameter we can also use this pattern: Runtime.exec(new String[] {command, parameter}). If we assume that the implementations on all platforms use the StringTokenizer approach as specified by the Javadoc, then this fix works on all platforms and we don't have to make the fix Windows specific.
Created attachment 231380 [details] Fix for org.eclipse.ui.browser
This fix looks good to me. I will do some more testing this afternoon. We should able to use String.split instead of StringTokenizer to simplify the code a bit, but either approach works.
(In reply to comment #8) > This fix looks good to me. I will do some more testing this afternoon. > > We should able to use String.split instead of StringTokenizer to simplify > the code a bit, but either approach works. I know, but wanted to be as close to the specified behavior, which uses StringTokenizer. Once we have a reviewed fix, we can open bugs for the other components. I will take care of Platform UI.
+1, the fix looks good and works with new and older JREs. The StringTokenizer is explicitly specified in the crazy java.lang.Runtime#exec(String[,..]) APIs, so that's the safest fix at this time. I checked a few Linux and Mac VMs and they all use the StringTokenizer. Unfortunately, there's no way to be fully correct on all platforms with a single API, see bug 387504 comment 13. Correct parsing on most platforms would use org.eclipse.debug.core.DebugPlugin#parseArguments(String), which also supports spaces/quotes/escapes in the arguments as specified by sh. On Windows, the native CreateProcess API takes a complete "commandLine" string argument. Since Java doesn't offer a direct API for that, everything we do means we have to reverse-engineer the JRE's transformations and bugs (see bug 387504 comment 0).
There's another change for this in the pipeline (on top of 7u21 etc.): http://bugs.sun.com/view_bug.do?bug_id=8012453
(In reply to comment #10) > On Windows, the native CreateProcess API takes a complete "commandLine" > string argument. Since Java doesn't offer a direct API for that, everything > we do means we have to reverse-engineer the JRE's transformations and bugs Oracle recommends using ProcessBuilder which is available since Java 5, would that be an option ? http://www.oracle.com/technetwork/java/javase/7u21-relnotes-1932873.html#jruntime
All Runtime.exec(..) implementations already end up using a ProcessBuilder, so it doesn't matter which entry point we use (but Runtime.exec(..) is more convenient). All implementation changes were in ProcessBuilder/ProcessImpl. The advice to use ProcessBuilder is a general advice, but it doesn't change anything for anyone affected by this change. The advice doesn't make much sense in these release notes (which on the other hand fail to tell that the behavior of the Runtime.exec(String[], ..) and ProcessBuilder APIs have also been changed).
(In reply to comment #9) > Once we have a reviewed fix, we can open bugs for the other components. I > will take care of Platform UI. The following cases: MozillaBrowserAdapter.BrowserThread.openBrowser(String) MozillaFactory.isAvailable() in 'org.eclipse.help.base' are fine, because the factory is for Linux based Mozilla browsers only. The usages in 'org.eclipse.ui.workbench': DefaultWebBrowser.openURL(URL) DefaultWebBrowser.openWebBrowser(String) are also OK, because for Windows, we use Program.launch(String).
+1, the patch looks good. Tested using Java 1.7.0_21 and it works fine.
Fixed with http://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=67208447235112204ac5e190e2a9fc0358dbb1db
Verified in I20130528-1200.
*** Bug 419808 has been marked as a duplicate of this bug. ***
