Re: [drlvm] HARMONY-1363 - status update
2006/9/18, Geir Magnusson Jr. [EMAIL PROTECTED]: Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Actually, the same problem was already fixed for bootclasspath.properties, and now there is shared properties parser in luni/common. Also look at HARMONY-1376 - I believe these issues will be fixed together - just go for a patch if you don't want to wait for mine ;) Anyone else using launcher out there? :) Yes. We all will be. :) geir Pavel. On 9/14/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On Monday 18 September 2006 18:14, Geir Magnusson Jr. wrote: Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Wouldn't it be better (and safer) to fix the parser? Normally a properties file can contain any kind of line separators and still be parsed correctly, which is a Good Thing IMHO. E.g. according to the spec for java.util.Properties.load(), A natural line of input is terminated either by a set of line terminator characters (\n or \r or \r\n) or by the end of the file. Chris -- Chris Gray/k/ Embedded Java Solutions BE0503765045 Embedded Mobile Java, OSGihttp://www.k-embedded-java.com/ [EMAIL PROTECTED] +32 3 216 0369 - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On 9/19/06, Chris Gray [EMAIL PROTECTED] wrote: On Monday 18 September 2006 18:14, Geir Magnusson Jr. wrote: Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Wouldn't it be better (and safer) to fix the parser? +1. Indeed this is the good thing for doing to avoid the above-mentioned problem in future. Thanks, Vladimir. Normally a properties file can contain any kind of line separators and still be parsed correctly, which is a Good Thing IMHO. E.g. according to the spec for java.util.Properties.load(), A natural line of input is terminated either by a set of line terminator characters (\n or \r or \r\n) or by the end of the file. Chris -- Chris Gray/k/ Embedded Java Solutions BE0503765045 Embedded Mobile Java, OSGihttp://www.k-embedded-java.com/ [EMAIL PROTECTED] +32 3 216 0369 - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/19, Vladimir Gorr [EMAIL PROTECTED]: On 9/19/06, Chris Gray [EMAIL PROTECTED] wrote: On Monday 18 September 2006 18:14, Geir Magnusson Jr. wrote: Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Wouldn't it be better (and safer) to fix the parser? +1. Indeed this is the good thing for doing to avoid the above-mentioned problem in future. This is exactly the feature shared parser has, see HARMONY-1243 and HARMONY-1376. Thanks, Vladimir. Normally a properties file can contain any kind of line separators and still be parsed correctly, which is a Good Thing IMHO. E.g. according to the spec for java.util.Properties.load(), A natural line of input is terminated either by a set of line terminator characters (\n or \r or \r\n) or by the end of the file. Chris -- Chris Gray/k/ Embedded Java Solutions BE0503765045 Embedded Mobile Java, OSGihttp://www.k-embedded-java.com/ [EMAIL PROTECTED] +32 3 216 0369 - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On Sep 19, 2006, at 3:01 AM, Chris Gray wrote: On Monday 18 September 2006 18:14, Geir Magnusson Jr. wrote: Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Wouldn't it be better (and safer) to fix the parser? Normally a properties file can contain any kind of line separators and still be parsed correctly, which is a Good Thing IMHO. E.g. according to the spec for java.util.Properties.load(), A natural line of input is terminated either by a set of line terminator characters (\n or \r or \r\n) or by the end of the file. Yes - and I think there is a patch somewhere. But we should keep little nits like this under control. geir Chris -- Chris Gray/k/ Embedded Java Solutions BE0503765045 Embedded Mobile Java, OSGihttp://www.k-embedded-java.com/ [EMAIL PROTECTED] +32 3 216 0369 - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. trying... ant from enhanced/trunk failed on build.native, could not find lcms.h and icc34.h looks exactly as this failure: http://mail-archives.apache.org/mod_mbox/incubator-harmony-commits/200607.mbox/[EMAIL PROTECTED] If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. trying... ant from enhanced/trunk failed on build.native, could not find lcms.h and icc34.h sounds like a dependency problem... we need to document those. I'll do that in a fresh install of Ubuntu sometime. looks exactly as this failure: http://mail-archives.apache.org/mod_mbox/incubator-harmony-commits/200607.mbox/[EMAIL PROTECTED] If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. trying... ant from enhanced/trunk failed on build.native, could not find lcms.h and icc34.h sounds like a dependency problem... we need to document those. I'll do that in a fresh install of Ubuntu sometime. yeah, but isn't it a kind of false dependence? I wonder, why we need liblcms.. need to investigate this.. maybe, unset HAVE_LCMS_H ... looks exactly as this failure: http://mail-archives.apache.org/mod_mbox/incubator-harmony-commits/200607.mbox/[EMAIL PROTECTED] If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib maybe, put it in SVN directly? are there any caveats? and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. trying... ant from enhanced/trunk failed on build.native, could not find lcms.h and icc34.h sounds like a dependency problem... we need to document those. I'll do that in a fresh install of Ubuntu sometime. yeah, but isn't it a kind of false dependence? I wonder, why we need liblcms.. need to investigate this.. maybe, unset HAVE_LCMS_H ... looks exactly as this failure: http://mail-archives.apache.org/mod_mbox/incubator-harmony-commits/200607.mbox/[EMAIL PROTECTED] If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib Doh! of course :) maybe, put it in SVN directly? are there any caveats? Well, I don't think so - the only problem I can imagine is what happens when you do the federated build from enhanced/trunk, but now that I think of it, I've been working this way for months now, and it hasn't gotten in the way. I'll check it in. geir and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Egor Pasko wrote: If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib maybe, put it in SVN directly? are there any caveats? and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. interesting... if I build working_classlib and working_drlvm separately, there is no issue with lcms.h, and the build is successful (i.e. let's get rid of this lcms.h dependency later on) And yet, I just cannot run it... bash-boom-bang$ LD_LIBRARY_PATH=`pwd`:`pwd`/default ./java -showversion -cp ... Hello Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software Foundation or its licensors, as applicable. java version 1.5.0 pre-alpha : not complete or compatible svn = r446534, (Sep 18 2006), Linux/ia32/gcc 3.3.3, debug build http://incubator.apache.org/harmony free(): invalid pointer 0xbfffd0b8! SIGSEGV in VM code. Stack trace: 1: free (??:-1) 2: ?? (??:-1) 3: hymem_free_memory (??:-1) 4: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 5: properties_free (??:-1) 6: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 7: ?? (??:-1) 8: readClassPathFromPropertiesFile (??:-1) 9: ?? (??:-1) 10: ?? (??:-1) 11: ?? (??:-1) 12: ?? (??:-1) 13: ?? (??:-1) 14: _dl_runtime_resolve (??:-1) 15: ?? (??:-1) 16: JNI_OnLoad (??:-1) 17: ?? (??:-1) 18: ?? (??:-1) end of stack trace Segmentation fault Am I running the Harmony Launcher? Should I run it somewhat differently? -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Hm. What platform are you running on? I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? using the launcher is the default now : $ cd working_vm $ cd build $ sh build.sh $ cd deploy/jre/bin $ ./java -version $ ./java Foo works... geir Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib maybe, put it in SVN directly? are there any caveats? and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. interesting... if I build working_classlib and working_drlvm separately, there is no issue with lcms.h, and the build is successful (i.e. let's get rid of this lcms.h dependency later on) And yet, I just cannot run it... bash-boom-bang$ LD_LIBRARY_PATH=`pwd`:`pwd`/default ./java -showversion -cp ... Hello Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software Foundation or its licensors, as applicable. java version 1.5.0 pre-alpha : not complete or compatible svn = r446534, (Sep 18 2006), Linux/ia32/gcc 3.3.3, debug build http://incubator.apache.org/harmony free(): invalid pointer 0xbfffd0b8! SIGSEGV in VM code. Stack trace: 1: free (??:-1) 2: ?? (??:-1) 3: hymem_free_memory (??:-1) 4: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 5: properties_free (??:-1) 6: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 7: ?? (??:-1) 8: readClassPathFromPropertiesFile (??:-1) 9: ?? (??:-1) 10: ?? (??:-1) 11: ?? (??:-1) 12: ?? (??:-1) 13: ?? (??:-1) 14: _dl_runtime_resolve (??:-1) 15: ?? (??:-1) 16: JNI_OnLoad (??:-1) 17: ?? (??:-1) 18: ?? (??:-1) end of stack trace Segmentation fault Am I running the Harmony Launcher? Should I run it somewhat differently? - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Hm. What platform are you running on? SUSE 9 I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? many tests pass, hm.. but it runs on the RI from my path with org.apache.tools.ant.launch.Launcher using the launcher is the default now : $ cd working_vm $ cd build $ sh build.sh yep, sometimes, things like this: sh build.sh update $ cd deploy/jre/bin $ ./java -version $ ./java Foo works... well, the ./java of mine catches no runtime path for libhysig.so, there is only LD_LIBRARY_PATH for the last try... Something is definitely wrong. And not many ideas I have how to repair geir Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib maybe, put it in SVN directly? are there any caveats? and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. interesting... if I build working_classlib and working_drlvm separately, there is no issue with lcms.h, and the build is successful (i.e. let's get rid of this lcms.h dependency later on) And yet, I just cannot run it... bash-boom-bang$ LD_LIBRARY_PATH=`pwd`:`pwd`/default ./java -showversion -cp ... Hello Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software Foundation or its licensors, as applicable. java version 1.5.0 pre-alpha : not complete or compatible svn = r446534, (Sep 18 2006), Linux/ia32/gcc 3.3.3, debug build http://incubator.apache.org/harmony free(): invalid pointer 0xbfffd0b8! SIGSEGV in VM code. Stack trace: 1: free (??:-1) 2: ?? (??:-1) 3: hymem_free_memory (??:-1) 4: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 5: properties_free (??:-1) 6: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 7: ?? (??:-1) 8: readClassPathFromPropertiesFile (??:-1) 9: ?? (??:-1) 10: ?? (??:-1) 11: ?? (??:-1) 12: ?? (??:-1) 13: ?? (??:-1) 14: _dl_runtime_resolve (??:-1) 15: ?? (??:-1) 16: JNI_OnLoad (??:-1) 17: ?? (??:-1) 18: ?? (??:-1) end of stack trace Segmentation fault Am I running the Harmony Launcher? Should I run it somewhat differently? - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Hm. What platform are you running on? SUSE 9 I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? many tests pass, hm.. but it runs on the RI from my path with org.apache.tools.ant.launch.Launcher waiting a little bit gives what's expected: test failures: ./lnx_ia32_gcc_debug/deploy/jre/bin/java: error while loading shared libraries: libhysig.so: cannot open shared object file: No such file or directory -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Hm. What platform are you running on? SUSE 9 I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? many tests pass, hm.. but it runs on the RI from my path with org.apache.tools.ant.launch.Launcher waiting a little bit gives what's expected: test failures: ./lnx_ia32_gcc_debug/deploy/jre/bin/java: error while loading shared libraries: libhysig.so: cannot open shared object file: No such file or directory Hm. You shouldn't have to set any LD_LIBRARY_PATH - the harmony launcher is doing that. In build/default/jre/bin, is there a libhysig.so anywhere? geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Hm. What platform are you running on? SUSE 9 I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? many tests pass, hm.. but it runs on the RI from my path with org.apache.tools.ant.launch.Launcher waiting a little bit gives what's expected: test failures: ./lnx_ia32_gcc_debug/deploy/jre/bin/java: error while loading shared libraries: libhysig.so: cannot open shared object file: No such file or directory Hm. You shouldn't have to set any LD_LIBRARY_PATH - the harmony launcher is doing that. What is the launcher? deploy/jre/bin/java? this binary cannot find libhysig.so: $ ldd deploy/jre/bin/java | grep hysig libhysig.so = not found In build/default/jre/bin, is there a libhysig.so anywhere? yes, in deploy/jre/bin/ -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Geir Magnusson, Jr. wrote: Hm. What platform are you running on? SUSE 9 I can build and run on Ubuntu 6, 5 and WinXP in debug and release. Do your tests pass? many tests pass, hm.. but it runs on the RI from my path with org.apache.tools.ant.launch.Launcher waiting a little bit gives what's expected: test failures: ./lnx_ia32_gcc_debug/deploy/jre/bin/java: error while loading shared libraries: libhysig.so: cannot open shared object file: No such file or directory Hm. You shouldn't have to set any LD_LIBRARY_PATH - the harmony launcher is doing that. What is the launcher? deploy/jre/bin/java? yes this binary cannot find libhysig.so: $ ldd deploy/jre/bin/java | grep hysig libhysig.so = not found In build/default/jre/bin, is there a libhysig.so anywhere? yes, in deploy/jre/bin/ That's really weird. I can do the same, and it resolves on Ubuntu 6 w/ no LD_LIBRARY_PATH set. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Geir Magnusson Jr. wrote: Egor Pasko wrote: SNIP this binary cannot find libhysig.so: $ ldd deploy/jre/bin/java | grep hysig libhysig.so = not found In build/default/jre/bin, is there a libhysig.so anywhere? yes, in deploy/jre/bin/ That's really weird. I can do the same, and it resolves on Ubuntu 6 w/ no LD_LIBRARY_PATH set. I get the same thing on my SLES9 machine. If I dont manually set LD_LIBRARY_PATH, I get an error message: ./java: error while loading shared libraries: libhysig.so: cannot open shared object file: No such file or directory ldd ./java gives me the following: linux-gate.so.1 = (0xe000) libhyprt.so = /harmony/svn-checkouts/test/deploy/jdk/jre/bin/libhyprt.so (0x40018000) libhythr.so = /harmony/svn-checkouts/test/deploy/jdk/jre/bin/libhythr.so (0x40035000) libm.so.6 = /lib/tls/libm.so.6 (0x4004c000) libpthread.so.0 = /lib/tls/libpthread.so.0 (0x4006f000) libc.so.6 = /lib/tls/libc.so.6 (0x4007f000) libdl.so.2 = /lib/libdl.so.2 (0x40199000) /lib/ld-linux.so.2 (0x4000) libhysig.so = not found so oddly it appears that libhyprt.so and libhythr.so are found, but libhysig.so is not, even though they are in the same directory! I have taken a look at the Linux makefile for launcher, and I notice that libhysig.so is not listed in MDLLIBFILES, but libhyport.so and libhythr.so are. Adding $(DLLPATH)libhysig.so to this line of the makefile solves the LD_LIBRARY_PATH problem for me. Do you see the same thing Egor? Regards, Oliver geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Oliver Deakin IBM United Kingdom Limited - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
I'm trying on Ubuntu 5 and I'm getting problems as well. Problem is that it's in a Parallels VM on OS X and I can't figure out how to cut and paste between them... It's a really useless stacktrace w/ addr2line... Egor Pasko wrote: On the 0x1E8 day of Apache Harmony Egor Pasko wrote: If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. this one does not allow me to build.sh update on downloading XALAN. This is a kind of problem that appeared when classlib was not built, but now it looks like built successfully. trying more... Simple. You need to tell the DRLVM build where to find the built classlib. Right now it assumes a structure like : enhanced/classlib/trunk enhanced/drlvm/trunk but with the federated build structure of enhanced/trunk/ working_vm/ working_classlib/ I just put a file called drlvm.properties into enhanced/trunk/working_vm/build that contains external.dep.CLASSLIB.loc=../../../working_vm gotcha! :) external.dep.CLASSLIB.loc=../../../working_classlib maybe, put it in SVN directly? are there any caveats? and then I can go into working_classlib and refresh and build, and working_vm just picks that up on build and local assembly of jre. building further, drinking coffee.. interesting... if I build working_classlib and working_drlvm separately, there is no issue with lcms.h, and the build is successful (i.e. let's get rid of this lcms.h dependency later on) And yet, I just cannot run it... bash-boom-bang$ LD_LIBRARY_PATH=`pwd`:`pwd`/default ./java -showversion -cp ... Hello Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software Foundation or its licensors, as applicable. java version 1.5.0 pre-alpha : not complete or compatible svn = r446534, (Sep 18 2006), Linux/ia32/gcc 3.3.3, debug build http://incubator.apache.org/harmony free(): invalid pointer 0xbfffd0b8! SIGSEGV in VM code. Stack trace: 1: free (??:-1) 2: ?? (??:-1) 3: hymem_free_memory (??:-1) 4: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 5: properties_free (??:-1) 6: find_call_JNI_OnLoad (/export/users/evpasko/svn/harmony/enhanced/trunk/working_vm/vm/vmcore/src/util/natives_support.cpp:117) 7: ?? (??:-1) 8: readClassPathFromPropertiesFile (??:-1) 9: ?? (??:-1) 10: ?? (??:-1) 11: ?? (??:-1) 12: ?? (??:-1) 13: ?? (??:-1) 14: _dl_runtime_resolve (??:-1) 15: ?? (??:-1) 16: JNI_OnLoad (??:-1) 17: ?? (??:-1) 18: ?? (??:-1) end of stack trace Segmentation fault Am I running the Harmony Launcher? Should I run it somewhat differently? - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E8 day of Apache Harmony Oliver Deakin wrote: I have taken a look at the Linux makefile for launcher, and I notice that libhysig.so is not listed in MDLLIBFILES, but libhyport.so and libhythr.so are. Adding $(DLLPATH)libhysig.so to this line of the makefile solves the LD_LIBRARY_PATH problem for me. Do you see the same thing Egor? I do. Thanks, Oliver! Here is the patch: --- working_classlib/modules/luni/src/main/native/launcher/linux/makefile (revision 447251) +++ working_classlib/modules/luni/src/main/native/launcher/linux/makefile (working copy) @@ -21,7 +21,7 @@ BUILDFILES = $(SHAREDSUB)main.o $(SHAREDSUB)cmain.o \ $(SHAREDSUB)launcher_copyright.o $(SHAREDSUB)strbuf.o \ $(SHAREDSUB)libhlp.o -MDLLIBFILES = $(DLLPATH)libhyprt.so $(DLLPATH)libhythr.so +MDLLIBFILES = $(DLLPATH)libhyprt.so $(DLLPATH)libhythr.so $(DLLPATH)libhysig.so EXENAME = $(EXEPATH)java include $(HY_HDK)/build/make/rules.mk './java -showversion Hello' fails as previously. ...under investigation... -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize. I workarounded this by running unix2dos on harmonyvm.properties. Anyone else using launcher out there? :) Pavel. On 9/14/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Pavel Pervov wrote: Tried it on Windows and found the problem which, as it looks like, have never been caught before. As I discovered, launcher uses platform specific line separators to parse harmonyvm.properties on a specific platform. But haromynvm.properties, which is copied into deploy, has unix line endings and is skipped. As the result, EM can't initialize.' A ha! I workarounded this by running unix2dos on harmonyvm.properties. Ok - we should get that in as eol-native Anyone else using launcher out there? :) Yes. We all will be. :) geir Pavel. On 9/14/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Evgueni Brevnov [EMAIL PROTECTED]: On 9/15/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: Evgueni Brevnov wrote: I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? Because this code (call to port_filepath_canonical()) is commented out with the last changes :-) Yes - to get anything to work w/ the launcher, I had to remove it. I do believe you are talking about classloader.cpp, right? Right! I understand why removed it. I just give a point to others (somebody asked) why there was no such failure before :-) It appears, in fact, that assertion was added very recently, along with lib names canonicalization :) And indeed some lack of consideration shows up behind this quick-fix... geir I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Alexey Varlamov wrote: 2006/9/15, Gregory Shimansky [EMAIL PROTECTED]: 2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? Because that was utterly insane to assume that all resources were in the same directory as the running executable, so I took that out at least for one thing and I'm going back now to solve. And I thought that method name was misleading, because it didn't return a canonical path for something, it returned the concatenation of the running executable's full path and whatever it was passed. Like the name of my cat. Hmm, this is either exaggeration or bug: looking at implementation, it only appends current dir if an original path is relative. And supposed that port_filepath_canonical() is targeted for ordinary files, then it's behavior does not look insane, rather models general file access. But applying this method for dynamic library name indeed would produce queer result, which can drive you mad ;) Which brings me to a good question I'll ask in a separate thread re APR string pools. geir I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Good progress! I just now downloaded drlvm from apache repository to my Linux 2.6.xx machine. DRLVM builds and runs. Below is a (rough) list of tests that for some reason had problems on my test server. Does anyone else see the same thing? exception/exceptions.c threaGroupTest.xml logManager.cpp stack_iterator_ia32.cpp (problem with interpreter configuration??) - Weldon On 9/14/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: HARMONY-1363 has been committed - I've done all the work under linux, and now doing on windows based on the checkin. Once I'm done there, I'll report back what I think we need to do first. There are clear problems with stack overflow (I had to disable StackTest and an assert in signals_ia32.cpp just to get it to run - the code is broken somewhere...) and from the app perspective, there appears to be some reflection problem with Tomcat. It's a NPE under release build, and more info under debug. Anyway, moving forward geir Geir Magnusson Jr. wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Weldon Washburn Intel Middleware Products Division
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. After it disabled Eclipse3.1.1 runs OK! On 15 Sep 2006 16:05:25 +0700, Egor Pasko [EMAIL PROTECTED] wrote: On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Mikhail Fursov
Re: [drlvm] HARMONY-1363 - status update
Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) There should be no need for that if you are using the Harmony launcher, as it sets things up properly. And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. I should document this... geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? After it disabled Eclipse3.1.1 runs OK! That means you too are using some kind of custom build and not using the harmony launcher. Please - why aren't we all using the launcher? geir On 15 Sep 2006 16:05:25 +0700, Egor Pasko [EMAIL PROTECTED] wrote: On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I think this assertion failure needs investigation before we decide it should be removed. -- Gregory Shimansky, Intel Middleware Products Division
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) There should be no need for that if you are using the Harmony launcher, as it sets things up properly. I have the same problem as Egor. And, kernel.tests fail due to this - i..e when I run build.sh test, smoke tests pass and then kernel tests fail. BTW, I recall this is actual also for classlib tests on j9 - simple ant test does not work. And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? ? I build only w/ build.sh Here's how I work these days : I have checked out enhanced/trunk and run ant in that directory, which will do a svn switch on working_classlib and working_vm, checking out the head of classlib/trunk and drlvm/trunk into those two directories. If you let that run, it will build classlib and drlvm, and then package in to hdk and jre, but I don't use this in my day to day life. I tend to go into either working_vm and build there via build/build.sh (note you need to either set -Dextern.dep.CLASSLIB.loc correctly, or just have a drlvm.properties file in working_vm/build or when I need a classlib update, go up into working_classlib, do an svn update, and then ant. I should document this... geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
15 Sep 2006 16:05:25 +0700, Egor Pasko [EMAIL PROTECTED]: On the 0x1E5 day of Apache Harmony Egor Pasko wrote: On the 0x1E5 day of Apache Harmony Geir Magnusson, Jr. wrote: Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. I am looking into the slow startup problem. It'll take some time to download everything... hmm.. I had to export LD_LIBRARY_PATH to dynamically link to libhysig.so (I remember there was a pretty java.exec some day, now it is gone) And got a failure as follows: $ ./java Hello free(): invalid pointer 0xbfffd148! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6da7c! free(): invalid pointer 0xbfffd148! SIGSEGV in VM code. Stack trace: after that VM hang not responding to ^C, killed similar things on debug and release configs... Ooops. Got it exactly the same way, but even after killing console is frozen: console $ LD_LIBRARY_PATH=`pwd` ./java Hello free(): invalid pointer 0xbfffd988! free(): invalid pointer 0x40b7960e! free(): invalid pointer 0x40d6e93c! free(): invalid pointer 0xbfffd988! SIGSEGV in VM code. Stack trace: /console And one more crash: console $ LD_LIBRARY_PATH=`pwd` ./java Harmony Java launcher Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software Foundation or its licensors, as applicable. java [-vm:vmdll -vmdir:dir -D... [-X...]] [args] java: /nfs/ins/proj/drl/coreapi/avarlamo/harmony/linux.ia32/svn-repo/drlvm/vm/thread/src/thread_native_fat_monitor.c:183: monitor_wait_impl: Assertion `saved_recursion1' failed. Aborted /console It is curious that build.sh test passes smoke tests... SUSE 9 Am I doing something wrong? do we build with 'build.sh' yet? -- Egor Pasko, Intel Managed Runtime Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Gregory Shimansky [EMAIL PROTECTED]: 2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On 9/15/06, Alexey Varlamov [EMAIL PROTECTED] wrote: 2006/9/15, Gregory Shimansky [EMAIL PROTECTED]: 2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? Because this code (call to port_filepath_canonical()) is commented out with the last changes :-) I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
2006/9/15, Alexey Varlamov [EMAIL PROTECTED]: 2006/9/15, Gregory Shimansky [EMAIL PROTECTED]: I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? If you experience this assertion failure only when running eclipse it may be after VM startup. It may be ecliipse SWT natives library. I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division
Re: [drlvm] HARMONY-1363 - status update
Gregory Shimansky wrote: 2006/9/15, Geir Magnusson Jr. [EMAIL PROTECTED]: Mikhail Fursov wrote: Testing WindowsXP build. Have to disable this assersion: assert(is_name_lowercase(library_name)) in natives_support.cpp file to run Hello application. Yes, I was staring that that last night, trying to figure out why lowercase stuff is so important. Any clue? I am not sure why this assertion fails, but I think I know why it is important. On windows file paths may be different as strings but point to the same file. To detect duplications among native modules it is necessary to have unified form to compare names as strings. So it seems like all library names should be lowercase. I dunno why one of them doesn't satisfy this requirement. I think this assertion failure needs investigation before we decide it should be removed. Yes, I'm not removing it because I didn't understand it :) I'm looking into it now. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On 9/15/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: Evgueni Brevnov wrote: I recall there is port_filepath_canonical() method specifically to solve this, and at least classpath items were canonicalized during VM startup. Maybe this step is just missing for libraries... But why it started to fail only now? Because this code (call to port_filepath_canonical()) is commented out with the last changes :-) Yes - to get anything to work w/ the launcher, I had to remove it. I do believe you are talking about classloader.cpp, right? Right! I understand why removed it. I just give a point to others (somebody asked) why there was no such failure before :-) geir I think this assertion failure needs investigation before we decide it should be removed. Agree. -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
HARMONY-1363 has been committed - I've done all the work under linux, and now doing on windows based on the checkin. Once I'm done there, I'll report back what I think we need to do first. There are clear problems with stack overflow (I had to disable StackTest and an assert in signals_ia32.cpp just to get it to run - the code is broken somewhere...) and from the app perspective, there appears to be some reflection problem with Tomcat. It's a NPE under release build, and more info under debug. Anyway, moving forward geir Geir Magnusson Jr. wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On Friday 15 September 2006 02:46 Geir Magnusson Jr. wrote: HARMONY-1363 has been committed - I've done all the work under linux, and now doing on windows based on the checkin. Once I'm done there, I'll report back what I think we need to do first. There are clear problems with stack overflow (I had to disable StackTest and an assert in signals_ia32.cpp just to get it to run - the code is broken somewhere...) and from the app perspective, there appears to be some reflection problem with Tomcat. It's a NPE under release build, and more info under debug. Anyway, moving forward Yow! Yes, we're really moving forward. Thank you Geir! -- Gregory Shimansky, Intel Middleware Products Division - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Ok - I was testing with Tomcat 5.5.17 on Ubuntu and it seems to work wiht both release and debug builds. Strange problem is that something is really slow - server start time is something like 17 sec on my machine, but 1.7 w/ Sun. Clearly something is broken. geir Geir Magnusson Jr. wrote: HARMONY-1363 has been committed - I've done all the work under linux, and now doing on windows based on the checkin. Once I'm done there, I'll report back what I think we need to do first. There are clear problems with stack overflow (I had to disable StackTest and an assert in signals_ia32.cpp just to get it to run - the code is broken somewhere...) and from the app perspective, there appears to be some reflection problem with Tomcat. It's a NPE under release build, and more info under debug. Anyway, moving forward geir Geir Magnusson Jr. wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
On 9/15/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. This problem will disappear after applying the H-1431 patch. Thanks, Vladimir. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [drlvm] HARMONY-1363 - status update
Vladimir Gorr wrote: On 9/15/06, Geir Magnusson Jr. [EMAIL PROTECTED] wrote: I now have things building and running with the 1363 patch - more work for the launcher was needed. I'm now using the harmonyvm.properties file in the vmdir, and this is working well. I am now able to run tests, although I have one failure in StackTest, and interestingly, the new testsuite that came with the thread manager isn't running. This problem will disappear after applying the H-1431 patch. Nice, I read through the patch, and I like it. However, as I did spend some time trying to grok the smoke and cunit test frameworks, can you write a short paragraph explaining why it had to be changed? I don't understand the theory behind the test system, why it had parts spread in places like vm.xml... geir Thanks, Vladimir. So progress. geir - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]