RFR: 7134690: remove legacy jnilib support from ClassLoader and System [macosx]
Can I get the following change reviewed please? The change is to remove some mac specific code from: src/share/classes/java/lang/System.java and src/share/classes/java/lang/ClassLoader.java which added support for non-standard native library suffixes on Mac OS (ie. .jnilib as well as the standard .dylib). We would like to remove this code, so there will be no changes to those sources when the Mac changes get pushed to jdk7u-dev I have submitted a hotspot CR to track providing the same functionality in Mac specific code in hotspot. Though we can probably live without it in the short-term. http://cr.openjdk.java.net/~michaelm/7134690/webrev.1/ Thanks Michael
Re: RFR: 7134690: remove legacy jnilib support from ClassLoader and System [macosx]
On 27/01/2012 11:41, Michael McMahon wrote: Can I get the following change reviewed please? The change is to remove some mac specific code from: src/share/classes/java/lang/System.java and src/share/classes/java/lang/ClassLoader.java which added support for non-standard native library suffixes on Mac OS (ie. .jnilib as well as the standard .dylib). We would like to remove this code, so there will be no changes to those sources when the Mac changes get pushed to jdk7u-dev I have submitted a hotspot CR to track providing the same functionality in Mac specific code in hotspot. Though we can probably live without it in the short-term. http://cr.openjdk.java.net/~michaelm/7134690/webrev.1/ This looks fine to me as this wasn't the right place to support this. I don't know how common .jnilib was to know if the VM changes will be required in the short term. -Alan
Please review: 7127906: (launcher) convert the launcher regression tests to java
Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
Hi Kumar, I have a question/suggestion on TestHelper.isEnglishLocale field. - Is there any chance that the test case changes the default locale by Locale.setDefault() during the test execution? If that's the case, static initialization of the field may not work. - If the default locale does not change during the test execution, then I'd suggest to initialize it with something like: isEnglishLocale = Locale.getDefault().getLanguage().equals(en); Otherwise it won't work with other English locales such as en_GB. Other than that, it looks good to me. Naoto On 1/27/12 10:30 AM, Joe Darcy wrote: Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
On 1/27/2012 11:05 AM, Naoto Sato wrote: Hi Kumar, I have a question/suggestion on TestHelper.isEnglishLocale field. - Is there any chance that the test case changes the default locale by Locale.setDefault() during the test execution? If that's the case, static initialization of the field may not work. Yes, the tests that do change the default locale are UnicodeTest.java and I18NJarTest.java, however they execute in their own child Process/VM. I think I should also the save default to a static, such that we can revert back the locale to the original default. - If the default locale does not change during the test execution, then I'd suggest to initialize it with something like: isEnglishLocale = Locale.getDefault().getLanguage().equals(en); What do you suggest ? Thanks Kumar Otherwise it won't work with other English locales such as en_GB. Other than that, it looks good to me. Naoto On 1/27/12 10:30 AM, Joe Darcy wrote: Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
On 1/27/12 11:36 AM, Kumar Srinivasan wrote: On 1/27/2012 11:05 AM, Naoto Sato wrote: Hi Kumar, I have a question/suggestion on TestHelper.isEnglishLocale field. - Is there any chance that the test case changes the default locale by Locale.setDefault() during the test execution? If that's the case, static initialization of the field may not work. Yes, the tests that do change the default locale are UnicodeTest.java and I18NJarTest.java, however they execute in their own child Process/VM. Then I'd prefer to have isEnglishLocale as a simple static method, rather than a field. I think I should also the save default to a static, such that we can revert back the locale to the original default. Good point. - If the default locale does not change during the test execution, then I'd suggest to initialize it with something like: isEnglishLocale = Locale.getDefault().getLanguage().equals(en); What do you suggest ? To replace the current isEnglishLocale logic with the above, which should work for all English language locales. Naoto Thanks Kumar Otherwise it won't work with other English locales such as en_GB. Other than that, it looks good to me. Naoto On 1/27/12 10:30 AM, Joe Darcy wrote: Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Updated unit test for BigInteger patch (#4837946)
On 01/11/2012 11:59 AM, Joe Darcy wrote: Thanks Tim. I'll add this to my review queue after Alan Eliasen's work (finally) gets in. Is there a timeframe for this yet?
Review request for 7133220: Additional patches to JAXP 1.4.5 update 1 for 7u4
Hi, With this additional patch, I'm fixing a SpecJvm2008 failure against 7u4 b07 (7098746), along with a couple of other fixes and accumulated Xalan update. The changes are listed below and at http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7133220. The webrev is at: http://cr.openjdk.java.net/~joehw/7u4-7133220/webrev/ Additional patches to JAXP 1.4.5 update 1 Nr Category ID Synopsis/Description 1 JAXP/OTHER 7098746 SpecJvm2008 xml.transform subbenchmark fails validation 2 JAXP/OTHER 7131589 NPE IN PARSERCONFIGURATIONSETTINGS.ADDRECOGNIZEDFEATURES 3 JAXP/OTHER 7133058 FINDBUGS WARNINGS IN COM.SUN.XML.INTERNAL.* Apache Xalan update Nr Key Synopsis/Description 1 XALANJ-1243 java.lang.StackOverflowError in XString.equals() 2 XALANJ-1991 StackOverflowException comparing two strings 3 XALANJ-2001 normalize-space gives StackOverflowError 4 XALANJ-1434 org.apache.xpath.axes.AxesWalker getLastPos: duplicate predicate testing (one line change) 5 XALANJ-1497 xsl:copy adds a newline to processing instructions 6 XALANJ-1706 DocumentFragment returned by extension element causes multiple SAX endDocument() events 7 XALANJ-2218 XML/HTML serializers should have default m_escapeSetting = true 8 XALANJ-2316 WriterToUTF8Buffered.write(String s) fail with big strings 9 XALANJ-2336 Xalan-J should stop using java.util.Vector in some cases 10 XALANJ-2218 major 11/Mar/06 XML/HTML serializers should have default m_escapeSetting = true 11 XALANJ-2236 trivial 30/Oct/06 [PATCH] clean up static calls thru instance references 12 XALANJ-2271 major08/Mar/06 XML 1.1 Serialization, char in attribute value not escaped Thanks, Joe
hg: jdk8/tl/jdk: 7136538: typo in test/Makefile under the jdk_security3 target
Changeset: ff24779c147f Author:valeriep Date: 2012-01-27 15:25 -0800 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/ff24779c147f 7136538: typo in test/Makefile under the jdk_security3 target Summary: Fixed the typo of secrity. Reviewed-by: wetmore ! test/Makefile
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
Hi Naoto, Here is the incremental webrev since last revision: http://cr.openjdk.java.net/~ksrini/7127906/webrev.1/webrev.delta/index.html Here is the full webrev for reference: http://cr.openjdk.java.net/~ksrini/7127906/webrev.1/index.html Thanks Kumar On 1/27/12 11:36 AM, Kumar Srinivasan wrote: On 1/27/2012 11:05 AM, Naoto Sato wrote: Hi Kumar, I have a question/suggestion on TestHelper.isEnglishLocale field. - Is there any chance that the test case changes the default locale by Locale.setDefault() during the test execution? If that's the case, static initialization of the field may not work. Yes, the tests that do change the default locale are UnicodeTest.java and I18NJarTest.java, however they execute in their own child Process/VM. Then I'd prefer to have isEnglishLocale as a simple static method, rather than a field. I think I should also the save default to a static, such that we can revert back the locale to the original default. Good point. - If the default locale does not change during the test execution, then I'd suggest to initialize it with something like: isEnglishLocale = Locale.getDefault().getLanguage().equals(en); What do you suggest ? To replace the current isEnglishLocale logic with the above, which should work for all English language locales. Naoto Thanks Kumar Otherwise it won't work with other English locales such as en_GB. Other than that, it looks good to me. Naoto On 1/27/12 10:30 AM, Joe Darcy wrote: Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar
Re: Please review: 7127906: (launcher) convert the launcher regression tests to java
Looks good to me. Thanks. Naoto On 1/27/12 3:31 PM, Kumar Srinivasan wrote: Hi Naoto, Here is the incremental webrev since last revision: http://cr.openjdk.java.net/~ksrini/7127906/webrev.1/webrev.delta/index.html Here is the full webrev for reference: http://cr.openjdk.java.net/~ksrini/7127906/webrev.1/index.html Thanks Kumar On 1/27/12 11:36 AM, Kumar Srinivasan wrote: On 1/27/2012 11:05 AM, Naoto Sato wrote: Hi Kumar, I have a question/suggestion on TestHelper.isEnglishLocale field. - Is there any chance that the test case changes the default locale by Locale.setDefault() during the test execution? If that's the case, static initialization of the field may not work. Yes, the tests that do change the default locale are UnicodeTest.java and I18NJarTest.java, however they execute in their own child Process/VM. Then I'd prefer to have isEnglishLocale as a simple static method, rather than a field. I think I should also the save default to a static, such that we can revert back the locale to the original default. Good point. - If the default locale does not change during the test execution, then I'd suggest to initialize it with something like: isEnglishLocale = Locale.getDefault().getLanguage().equals(en); What do you suggest ? To replace the current isEnglishLocale logic with the above, which should work for all English language locales. Naoto Thanks Kumar Otherwise it won't work with other English locales such as en_GB. Other than that, it looks good to me. Naoto On 1/27/12 10:30 AM, Joe Darcy wrote: Hi Kumar, Looks fine; good to see the shell code go away! -Joe On 01/27/2012 10:03 AM, Kumar Srinivasan wrote: Hi Joe, Naoto, Sherman, Here are the highlights of these changes: 1. The TestHelper.java is now extensible, this simplifies the usage of static variables and methods, added a couple of convenience methods and some minor refactoring. 2. Eliminated some shell tests, (few more remaining, saved for a future project). a. UnicodeTest.sh, UnicodeCleanup.java merged into UnicodeTest.java b. i18nTest.sh, deleteI18n.sh, CreatePlatformFile.java merged into I18NTest.java c. ChangeDataModel.sh converted to ChangeDataModel.java d. unresolvedExceptions.sh converted and merged into UnresolvedExceptions.java 3. Arrrghs.java when testing on Solaris with ja_JP.UTF-8 locale, some of the tests fail as the launcher messages in java are localized, therefore these tests are skipped, under such conditions. Tests: Windows English/Japanese, Solaris C/ja_JP.UTF-8 + usual jprt systems (undergoing). webrev: http://cr.openjdk.java.net/~ksrini/7127906/webrev.0/ Thanks Kumar