From mark at klomp.org Sun Jun 1 15:37:05 2008 From: mark at klomp.org (Mark Wielaard) Date: Mon, 02 Jun 2008 00:37:05 +0200 Subject: OpenJDK 6 b10 source posted In-Reply-To: <4840AC46.8030409@sun.com> References: <4840AC46.8030409@sun.com> Message-ID: <1212359825.3133.84.camel@dijkstra.wildebeest.org> Hi Joe, On Fri, 2008-05-30 at 18:39 -0700, Joseph D. Darcy wrote: > The OpenJDK 6 b10 source is now available at > > http://download.java.net/openjdk/jdk6/ Thanks for releasing this. I have updated icedtea6 to include it. Are you planning to support a mercurial repo for openjdk6? Although your release notes below helped a lot, it is still hard work to review all the changes for integration if all you got is just the diff between two full versions. For those interested I posted the b09-b10 diff here: http://icedtea.classpath.org/~mjw/b09-b10.diff.gz (Excluding the new/deleted (gervill) files) Posting individual changes and/or having a public repo might also help understanding smaller things that you don't explicitly mention in your release notes like the changes to java.c, MetaData.java or ChunkedInputStream/test.txt (also removed from the diff since it is essentially a binary blob). > A few fixes in the 2D area: > > 6604044: java crashes talking to second X screen > 6636469: Java Fullscreen Exclusive Mode not working with Xorg server > 1.3.0 and above Removed icedtea-xinerama.patch from icedtea6 which did the same thing. > 6523403: OSS CMM: Need to provide lcms library with PYCC and > LINEAR_RGB OS ICC profiles > > The last fix was contributed by Andrew Haley and Keith Seitz of Red Hat > and fixes color management related regression and JCK failures. Thanks, I removed the icedtea-color-profiles.patch and the overlay. > The encumbered sound engine has been replaced with Gervill: > > 6702956 OpenJDK: replace encumbered code (software synthesizer) Very nice. We had good experience with this code in IcedTea. It wasn't clear precisely which upstream version of Gervill you are including. It does look like you are missing some fixes Karl and I made since it got integrated into icedtea. For example you miss the sample rate conversion support. But it is somewhat hard to see because all the files got completely reformatted from the upstream sources. You should be able to find all the changes we made on the mailinglist (sound-dev, distro-pkg-dev and audio-engine-dev) and of course in Gervill CVS https://gervill.dev.java.net/source/browse/gervill/ (or icedtea6 mercurial under overlays). Could you make the changes you made to the upstream sources available separate from the changes you made to the indention of the sources for easy comparison? For now I kept the current upstream/icedtea version. > Some progress has been made on addressing the copyright/licensing > concerns raised by Debian > (http://mail.openjdk.java.net/pipermail/jdk6-dev/2008-April/000044.html): > > 6695553: Cleanup GPLv2+SPL legal notices in hat sources > 6703931 Remove jaxws doc-files from the jdk > > Work will continue in this area until all the issues are addressed. Very nice. I removed the patches/icedtea-hat-spl-gpl.patch. And this also closes: - http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=142 And I saw soundbank.gm was also removed, which closes: - http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=149 The general tracking bug for hopefully all these issues is: http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=138 > Mark Wielaard recently proposed publishing regression test results for > OpenJDK 6 > (http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2008-May/001921.html). > > > I'll publish results for b10 later. That would be cool. I am also happy to note that Lillian and Tom added jtregcheck to the Fedora rpm builds in addition to running the Mauve tests. The test results output (just the jtreg -v1 output for now, not yet the full JTreport and JTwork output) are then included in the src subpackage rpm and in the build logs. Which is great because it also allows us to easily monitor the (zero) ports that Fedora ships like ppc and ppc64. For an example see: http://koji.fedoraproject.org/koji/buildinfo?buildID=51083 Click on the build logs (build.log) and search for "JTREG TESTING". The results still have a bit too many failures because it doesn't currently provide a Xvfb wrapper like it does for Mauve. But that is in the works. Cheers, Mark 2008-06-01 Mark Wielaard * Makefile.am (OPENJDK_DATE): Updated to 30_may_2008. (OPENJDK_MD5SUM): Updated. (OPENJDK_VERSION): Updated to b10. (OPENJDK_SRC_ZIP): Adjusted to new underscore convention between version and date in b10. (ICEDTEA_PATCHES): Removed icedtea-xinerama.patch, icedtea-color-profiles.patch and icedtea-hat-spl-gpl.patch. * Makefile.in: Regenerated. * patches/icedtea-copy-plugs.patch: Adjusted to remove sound references in b10. * patches/icedtea-gervill.patch: Adjusted to new sound service files in b10. * overlays/openjdk/jdk/src/share/lib/cmm/lcms/GRAY.pf: Removed. * overlays/openjdk/jdk/src/share/lib/cmm/lcms/LINEAR_RGB.pf: Removed. * overlays/openjdk/jdk/src/share/lib/cmm/lcms/PYCC.pf: Removed. * patches/icedtea-color-profiles.patch: Removed. * patches/icedtea-hat-spl-gpl.patch: Removed. * patches/icedtea-xinerama.patch: Removed. -------------- next part -------------- A non-text attachment was scrubbed... Name: icedtea6-b10.patch Type: text/x-patch Size: 7653 bytes Desc: not available Url : http://mail.openjdk.java.net/pipermail/jdk6-dev/attachments/20080602/8b225606/attachment.bin From Joe.Darcy at Sun.COM Mon Jun 2 10:19:44 2008 From: Joe.Darcy at Sun.COM (Joseph D. Darcy) Date: Mon, 02 Jun 2008 10:19:44 -0700 Subject: OpenJDK 6 b10 source posted In-Reply-To: <1212359825.3133.84.camel@dijkstra.wildebeest.org> References: <4840AC46.8030409@sun.com> <1212359825.3133.84.camel@dijkstra.wildebeest.org> Message-ID: <48442BB0.90802@sun.com> Mark Wielaard wrote: > Hi Joe, > > On Fri, 2008-05-30 at 18:39 -0700, Joseph D. Darcy wrote: >> The OpenJDK 6 b10 source is now available at >> >> http://download.java.net/openjdk/jdk6/ > > Thanks for releasing this. I have updated icedtea6 to include it. > > Are you planning to support a mercurial repo for openjdk6? Yes, a mercurial repo for OpenJDK 6 should be up and running withing a month or two. > Although your > release notes below helped a lot, it is still hard work to review all > the changes for integration if all you got is just the diff between two > full versions. Agreed. > For those interested I posted the b09-b10 diff here: > http://icedtea.classpath.org/~mjw/b09-b10.diff.gz > (Excluding the new/deleted (gervill) files) > > Posting individual changes and/or having a public repo might also help > understanding smaller things that you don't explicitly mention in your > release notes like the changes to java.c, MetaData.java or > ChunkedInputStream/test.txt (also removed from the diff since it is > essentially a binary blob). > >> A few fixes in the 2D area: >> >> 6604044: java crashes talking to second X screen >> 6636469: Java Fullscreen Exclusive Mode not working with Xorg server >> 1.3.0 and above > > Removed icedtea-xinerama.patch from icedtea6 which did the same thing. > >> 6523403: OSS CMM: Need to provide lcms library with PYCC and >> LINEAR_RGB OS ICC profiles >> >> The last fix was contributed by Andrew Haley and Keith Seitz of Red Hat >> and fixes color management related regression and JCK failures. > > Thanks, I removed the icedtea-color-profiles.patch and the overlay. > >> The encumbered sound engine has been replaced with Gervill: >> >> 6702956 OpenJDK: replace encumbered code (software synthesizer) > > Very nice. We had good experience with this code in IcedTea. > > It wasn't clear precisely which upstream version of Gervill you are > including. It does look like you are missing some fixes Karl and I made > since it got integrated into icedtea. For example you miss the sample > rate conversion support. But it is somewhat hard to see because all the > files got completely reformatted from the upstream sources. > > You should be able to find all the changes we made on the mailinglist > (sound-dev, distro-pkg-dev and audio-engine-dev) and of course in > Gervill CVS https://gervill.dev.java.net/source/browse/gervill/ (or > icedtea6 mercurial under overlays). > > Could you make the changes you made to the upstream sources available > separate from the changes you made to the indention of the sources for > easy comparison? I'll forward this request to the engineer who did the Gervill integration. > For now I kept the current upstream/icedtea version. > >> Some progress has been made on addressing the copyright/licensing >> concerns raised by Debian >> (http://mail.openjdk.java.net/pipermail/jdk6-dev/2008-April/000044.html): >> >> 6695553: Cleanup GPLv2+SPL legal notices in hat sources >> 6703931 Remove jaxws doc-files from the jdk >> >> Work will continue in this area until all the issues are addressed. > > Very nice. I removed the patches/icedtea-hat-spl-gpl.patch. > And this also closes: > - http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=142 > > And I saw soundbank.gm was also removed, which closes: > - http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=149 > > The general tracking bug for hopefully all these issues is: > http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=138 > >> Mark Wielaard recently proposed publishing regression test results for >> OpenJDK 6 >> (http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2008-May/001921.html). >> Thanks for the update on incorporating the changes in b10. >> I'll publish results for b10 later. > > That would be cool. Yes; however, my own work on OpenJDK 6 will be delayed for a while since I am out of the office for a death in the family. Regards, -Joe > I am also happy to note that Lillian and Tom added jtregcheck to the > Fedora rpm builds in addition to running the Mauve tests. The test > results output (just the jtreg -v1 output for now, not yet the full > JTreport and JTwork output) are then included in the src subpackage rpm > and in the build logs. > > Which is great because it also allows us to easily monitor the (zero) > ports that Fedora ships like ppc and ppc64. For an example see: > http://koji.fedoraproject.org/koji/buildinfo?buildID=51083 > Click on the build logs (build.log) and search for "JTREG TESTING". > The results still have a bit too many failures because it doesn't > currently provide a Xvfb wrapper like it does for Mauve. But that is in > the works. > > Cheers, > > Mark > > > 2008-06-01 Mark Wielaard > > * Makefile.am (OPENJDK_DATE): Updated to 30_may_2008. > (OPENJDK_MD5SUM): Updated. > (OPENJDK_VERSION): Updated to b10. > (OPENJDK_SRC_ZIP): Adjusted to new underscore convention between > version and date in b10. > (ICEDTEA_PATCHES): Removed icedtea-xinerama.patch, > icedtea-color-profiles.patch and icedtea-hat-spl-gpl.patch. > * Makefile.in: Regenerated. > * patches/icedtea-copy-plugs.patch: Adjusted to remove sound > references in b10. > * patches/icedtea-gervill.patch: Adjusted to new sound service > files in b10. > * overlays/openjdk/jdk/src/share/lib/cmm/lcms/GRAY.pf: Removed. > * overlays/openjdk/jdk/src/share/lib/cmm/lcms/LINEAR_RGB.pf: Removed. > * overlays/openjdk/jdk/src/share/lib/cmm/lcms/PYCC.pf: Removed. > * patches/icedtea-color-profiles.patch: Removed. > * patches/icedtea-hat-spl-gpl.patch: Removed. > * patches/icedtea-xinerama.patch: Removed. > > From kalli at midverk.is Tue Jun 3 18:09:55 2008 From: kalli at midverk.is (Karl Helgason) Date: Wed, 4 Jun 2008 01:09:55 +0000 Subject: OpenJDK 6 b10 source posted Message-ID: <36EC82E93EB0AD40A4301DAD654323868CA4FB2AE1@mail.midverk.is> Mark Wielaard wrote: > It wasn't clear precisely which upstream version of Gervill you are > including. It does look like you are missing some fixes Karl and I made > since it got integrated into icedtea. For example you miss the sample > rate conversion support. But it is somewhat hard to see because all the > files got completely reformatted from the upstream sources. They are using Gervill 0.9 source dump with these fixes (which are also in current cvs): - Fix: Default tempo in SoftMidiFileReader was supposed to be 500000 mpq or 120 bpm - Fix: Change type of info variable in SoftSynthesize into MidiDevice.Info - Fix: Fix instanceof check of p1 variable in ModelInstrumentComparator - Fix: Remove redundant constant variable in ModelByteBufferWaveTable The code was also slightly reformatted: a) lines "* @version %I%, %E%" was removed (such lines was removed from all JDK sources); b) line indention (to make lines no longer than 80 symbols); c) tabs to spaces conversion; I will look into to reformat newest Gervill CVS code the same way. Regards, Karl From mark at klomp.org Wed Jun 4 02:25:24 2008 From: mark at klomp.org (Mark Wielaard) Date: Wed, 04 Jun 2008 11:25:24 +0200 Subject: OpenJDK 6 b10 source posted In-Reply-To: <36EC82E93EB0AD40A4301DAD654323868CA4FB2AE1@mail.midverk.is> References: <36EC82E93EB0AD40A4301DAD654323868CA4FB2AE1@mail.midverk.is> Message-ID: <1212571525.3172.11.camel@dijkstra.wildebeest.org> Hi Karl, On Wed, 2008-06-04 at 01:09 +0000, Karl Helgason wrote: > Mark Wielaard wrote: > > It wasn't clear precisely which upstream version of Gervill you are > > including. It does look like you are missing some fixes Karl and I made > > since it got integrated into icedtea. For example you miss the sample > > rate conversion support. But it is somewhat hard to see because all the > > files got completely reformatted from the upstream sources. > > They are using Gervill 0.9 source dump with these fixes (which are also in current cvs): > - Fix: Default tempo in SoftMidiFileReader was supposed to be 500000 mpq or 120 bpm > - Fix: Change type of info variable in SoftSynthesize into MidiDevice.Info > - Fix: Fix instanceof check of p1 variable in ModelInstrumentComparator > - Fix: Remove redundant constant variable in ModelByteBufferWaveTable Thanks for checking. These are all in the icedtea gervill overlay also. I'll make sure to upgrade to the latest CVS version in which I see you fixed some of the other issues I reported. > The code was also slightly reformatted: > a) lines "* @version %I%, %E%" was removed (such lines was removed from all JDK > sources); > b) line indention (to make lines no longer than 80 symbols); > c) tabs to spaces conversion; > > I will look into to reformat newest Gervill CVS code the same way. That would be great. Since 0.9 there have been lots of changes and fixes (which is why icedtea tracks your CVS tree and didn't use the old 0.9 code). Would it be a good idea to do a new version upgrade after syncing the sources and reformatting to celebrate the inclusion? Cheers, Mark From kalli at midverk.is Wed Jun 4 18:56:18 2008 From: kalli at midverk.is (Karl Helgason) Date: Thu, 5 Jun 2008 01:56:18 +0000 Subject: SV: OpenJDK 6 b10 source posted In-Reply-To: <1212571525.3172.11.camel@dijkstra.wildebeest.org> References: <36EC82E93EB0AD40A4301DAD654323868CA4FB2AE1@mail.midverk.is>, <1212571525.3172.11.camel@dijkstra.wildebeest.org> Message-ID: <36EC82E93EB0AD40A4301DAD654323868CA4FB2AE2@mail.midverk.is> Hi Mark, I have reformatted the Gervill CVS code. It should be possible now to make diffs between them easily. cheers, Karl ________________________________________ Fr?: Mark Wielaard [mark at klomp.org] Sent: 4. j?n? 2008 09:25 Vi?takandi: Karl Helgason Afrit: jdk6-dev at openjdk.java.net; distro-pkg-dev at openjdk.java.net Efni: Re: OpenJDK 6 b10 source posted Hi Karl, On Wed, 2008-06-04 at 01:09 +0000, Karl Helgason wrote: > Mark Wielaard wrote: > > It wasn't clear precisely which upstream version of Gervill you are > > including. It does look like you are missing some fixes Karl and I made > > since it got integrated into icedtea. For example you miss the sample > > rate conversion support. But it is somewhat hard to see because all the > > files got completely reformatted from the upstream sources. > > They are using Gervill 0.9 source dump with these fixes (which are also in current cvs): > - Fix: Default tempo in SoftMidiFileReader was supposed to be 500000 mpq or 120 bpm > - Fix: Change type of info variable in SoftSynthesize into MidiDevice.Info > - Fix: Fix instanceof check of p1 variable in ModelInstrumentComparator > - Fix: Remove redundant constant variable in ModelByteBufferWaveTable Thanks for checking. These are all in the icedtea gervill overlay also. I'll make sure to upgrade to the latest CVS version in which I see you fixed some of the other issues I reported. > The code was also slightly reformatted: > a) lines "* @version %I%, %E%" was removed (such lines was removed from all JDK > sources); > b) line indention (to make lines no longer than 80 symbols); > c) tabs to spaces conversion; > > I will look into to reformat newest Gervill CVS code the same way. That would be great. Since 0.9 there have been lots of changes and fixes (which is why icedtea tracks your CVS tree and didn't use the old 0.9 code). Would it be a good idea to do a new version upgrade after syncing the sources and reformatting to celebrate the inclusion? Cheers, Mark No virus found in this incoming message. Checked by AVG. Version: 7.5.524 / Virus Database: 269.24.6/1480 - Release Date: 3.6.2008 07:00 From Joe.Darcy at Sun.COM Tue Jun 10 20:45:40 2008 From: Joe.Darcy at Sun.COM (Joseph D. Darcy) Date: Tue, 10 Jun 2008 20:45:40 -0700 Subject: OpenJDK 6 b10 source posted In-Reply-To: <4840AC46.8030409@sun.com> References: <4840AC46.8030409@sun.com> Message-ID: <484F4A64.6050002@sun.com> PS Another fix of note was 6707643: (tz) Support tzdata2008c -Joe Joseph D. Darcy wrote: > Hello. > > The OpenJDK 6 b10 source is now available at > > http://download.java.net/openjdk/jdk6/ > > In this source drop, the build instructions received some clarifications > and corrections: > > 6644659: Error in default target of make/javax/crypto in OpenJDK build > 6682167: Add cygwin faq to README-builds.html > 6563616: Clarify instructions for unpacking openjdk binary "plug" > 6611685: Incorrect link to CA certs info from build README > 6597857: JDK build instructions need to be updated > (BUILD_JDK_IMPORT_PATH, BUILD_BINARY_PLUGS_PATH) > > and some other work too: > > 6683078: Update JCE framework and provider builds to work on read-only > filesystems > 6644659: Error in default target of make/javax/crypto in OpenJDK build > > > A few fixes in the 2D area: > > 6604044: java crashes talking to second X screen > 6636469: Java Fullscreen Exclusive Mode not working with Xorg server > 1.3.0 and above > 6523403: OSS CMM: Need to provide lcms library with PYCC and > LINEAR_RGB OS ICC profiles > > The last fix was contributed by Andrew Haley and Keith Seitz of Red Hat > and fixes color management related regression and JCK failures. > > > The encumbered sound engine has been replaced with Gervill: > > 6702956 OpenJDK: replace encumbered code (software synthesizer) > > With this change, the only remaining binary plug in OpenJDK 6 is for SNMP. > > I applied Martin's recent fixes in collections: > > 6691185: (coll) TreeMap.navigableKeySet's descendingIterator method > starts at first instead of last entry > 6691215: (coll) IdentityHashMap.containsValue(null) returns true when > null value not present > > > Some progress has been made on addressing the copyright/licensing > concerns raised by Debian > (http://mail.openjdk.java.net/pipermail/jdk6-dev/2008-April/000044.html): > > 6695553: Cleanup GPLv2+SPL legal notices in hat sources > 6703931 Remove jaxws doc-files from the jdk > > Work will continue in this area until all the issues are addressed. > > > Mark Wielaard recently proposed publishing regression test results for > OpenJDK 6 > (http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2008-May/001921.html). > > > I'll publish results for b10 later. However, some of the > regression tests assume they are either running inside Sun's firewall or > checking aspects of Sun's production/commercial/proprietary JDK rather > than an OpenJDK build. The tests should be modified so they are > appropriate for OpenJDK out of the box too. Along those lines, some > regression tests were deleted or modified in b10: > > > 6673853: LegacyIntrospectorTest is testing an old deprecated com.sun API > not present in OpenJDK > 6602426: TEST_BUG: test/sun/misc/Version/VersionCheck.java fails > 6704655: Test test/java/lang/reflect/Generics/Probe.java fails under > OpenJDK > 6705893: javax.script tests should not require a js engine on OpenJDK > > Some regression tests, such as those in networking and javax.script, > should probably be modified to more easily accept configuration options > without having to modify the source of the tests. Discussion on how > best to do that has started > (http://mail.openjdk.java.net/pipermail/nio-dev/2008-May/000028.html). > > -Joe > > From Joe.Darcy at Sun.COM Fri Jun 20 15:21:41 2008 From: Joe.Darcy at Sun.COM (Joe Darcy) Date: Fri, 20 Jun 2008 15:21:41 -0700 Subject: duplicated sanity check warning In-Reply-To: <1ccfd1c10805310054r48dd5c95obf21123ecd9b7c93@mail.gmail.com> References: <1ccfd1c10805310054r48dd5c95obf21123ecd9b7c93@mail.gmail.com> Message-ID: <485C2D75.6070205@sun.com> Hello. This fix will be reflected in the next OpenJDK 6 source drop, b11. Thanks, -Joe Martin Buchholz wrote: > This is a bug report. > > I was trying to build openjdk6, and make sanity got duplicated > BUILD_JDK = true > lines in the output. > > Looking at control/make/make/sanity-rules.gmk, I see duplicate lines > > ifeq ($(JDK_SRC_AVAILABLE), true) > @$(ECHO) " BUILD_JDK = $(BUILD_JDK) " >> $(MESSAGE_FILE) > endif > ifeq ($(JDK_SRC_AVAILABLE), true) > @$(ECHO) " BUILD_JDK = $(BUILD_JDK) " >> $(MESSAGE_FILE) > endif > > Delete one of the above. > > Martin >