From Kelly.Ohair at Sun.COM Mon Oct 26 11:21:22 2009 From: Kelly.Ohair at Sun.COM (Kelly O'Hair) Date: Mon, 26 Oct 2009 11:21:22 -0700 Subject: Boot cycle builds Message-ID: <4AE5E8A2.6050900@sun.com> Heads up... As soon as we can do a jdk7 boot cycle build again, I plan on changing the default build to include a boot cycle build (where the jdk7 image is built, then used as the boot jdk to build it again). I consider it critical that boot cycle builds work, and I see no other course than requiring it as part of a default jdk7 build. It can be turned off with SKIP_BOOT_CYCLE=true, but I hope the integrators and everyone will be careful when it is turned off. Depending on your build situation, this could add 20-30 minutes (maybe more) to your jdk7 build, which is painful I know, but so is finding out that the jdk you just built doesn't work. Comments are welcome. -kto From mr at sun.com Thu Oct 29 10:08:00 2009 From: mr at sun.com (Mark Reinhold) Date: Thu, 29 Oct 2009 10:08:00 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: kelly.ohair@sun.com; Thu, 29 Oct 2009 10:00:57 PDT; <4AE9CA49.9060807@sun.com> Message-ID: <20091029170800.EB811384@eggemoggin.niobe.net> (moving to open lists) > Date: Thu, 29 Oct 2009 10:00:57 -0700 > From: kelly.ohair at sun.com > This hotspot change broke the jdk makefiles: > > 6722084: JPRT make file doesn't create required symbolic link to libjvm.so > http://hg.openjdk.java.net/jdk7/jdk7/hotspot/rev/23862fc517bb Thanks for the diagnosis. > In the meantime, I will need to try and fix this in the jdk makefiles. > Stay tuned ... Why fix this in the jdk makefiles if it's a HotSpot problem? - Mark From Kelly.Ohair at Sun.COM Thu Oct 29 10:31:29 2009 From: Kelly.Ohair at Sun.COM (Kelly O'Hair) Date: Thu, 29 Oct 2009 10:31:29 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <20091029170800.EB811384@eggemoggin.niobe.net> References: <20091029170800.EB811384@eggemoggin.niobe.net> Message-ID: <4AE9D171.40400@sun.com> Mark Reinhold wrote: > (moving to open lists) > >> Date: Thu, 29 Oct 2009 10:00:57 -0700 >> From: kelly.ohair at sun.com > >> This hotspot change broke the jdk makefiles: >> >> 6722084: JPRT make file doesn't create required symbolic link to libjvm.so >> http://hg.openjdk.java.net/jdk7/jdk7/hotspot/rev/23862fc517bb > > Thanks for the diagnosis. > >> In the meantime, I will need to try and fix this in the jdk makefiles. >> Stay tuned ... > > Why fix this in the jdk makefiles if it's a HotSpot problem? > Quickest path to getting us past this... File is jdk/make/java/redist/Makefile, maybe a one line change: 214c214 < $(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) --- > $(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/lib/$(LIBARCH)/$(LIBJSIG_NAME) I'm testing this out now... Undoing all the hotspot changes seemed a bit drastic... -kto > - Mark From mr at sun.com Thu Oct 29 10:34:34 2009 From: mr at sun.com (Mark Reinhold) Date: Thu, 29 Oct 2009 10:34:34 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: kelly.ohair@sun.com; Thu, 29 Oct 2009 10:31:29 PDT; <4AE9D171.40400@sun.com> Message-ID: <20091029173434.C615D384@eggemoggin.niobe.net> > Date: Thu, 29 Oct 2009 10:31:29 -0700 > From: kelly.ohair at sun.com > Quickest path to getting us past this... > > File is jdk/make/java/redist/Makefile, maybe a one line change: > > 214c214 > < $(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) > --- >> $(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/lib/$(LIBARCH)/$(LIBJSIG_NAME) > > I'm testing this out now... Okay, thanks. > Undoing all the hotspot changes seemed a bit drastic... Agreed; I wouldn't have suggested that. - Mark From Kelly.Ohair at Sun.COM Thu Oct 29 10:57:05 2009 From: Kelly.Ohair at Sun.COM (Kelly O'Hair) Date: Thu, 29 Oct 2009 10:57:05 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <20091029173434.C615D384@eggemoggin.niobe.net> References: <20091029173434.C615D384@eggemoggin.niobe.net> Message-ID: <4AE9D771.4010308@sun.com> I think this is the fix... it is taking some time to completely verify it. Can someone review this change? And do I have approval to push directly into master? ------- diff --git a/make/java/redist/Makefile b/make/java/redist/Makefile --- a/make/java/redist/Makefile +++ b/make/java/redist/Makefile @@ -216,7 +216,7 @@ endif # PLATFORM $(LIB_LOCATION)/$(KERNEL_LOCATION)/$(JVM_NAME): $(HOTSPOT_KERNEL_PATH)/$(JVM_NAME) $(install-file) -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/jre/lib/$(LIBARCH)/$(LIBJSIG_NAME) $(install-import-file) $(LIB_LOCATION)/$(CLIENT_LOCATION)/$(LIBJSIG_NAME) \ ------- A little background, for the record (while I wait for this build to finish ): The files exported by the hotspot makefiles for import to the jdk come in the shape of a sparse jdk install image, and the jdk Makefile jdk/make/java/redist/Makefile is responsible for copying or importing these files from this hotspot exported area, into the final jdk image. This hotspot area is specified via the ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH variables, or could be the ALT_JDK_IMPORT_PATH setting for partial builds. If the hotspot team changes what it exports, this jdk makefile may need to change, and if the hotspot team wants to impact the shape or layout of the final jdk image with regards to their files, they need to address this in the jdk/make/java/redist/Makefile. -kto From mr at sun.com Thu Oct 29 11:02:16 2009 From: mr at sun.com (Mark Reinhold) Date: Thu, 29 Oct 2009 11:02:16 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: kelly.ohair@sun.com; Thu, 29 Oct 2009 10:57:05 PDT; <4AE9D771.4010308@sun.com> Message-ID: <20091029180216.83F67384@eggemoggin.niobe.net> > Date: Thu, 29 Oct 2009 10:57:05 -0700 > From: kelly.ohair at sun.com > I think this is the fix... it is taking some time to completely > verify it. > > Can someone review this change? > > And do I have approval to push directly into master? Yes. Michael -- Could you please re-start the RE builds after Kelly pushes his fix? Thanks. - Mark From Tim.Bell at Sun.COM Thu Oct 29 11:04:23 2009 From: Tim.Bell at Sun.COM (Tim Bell) Date: Thu, 29 Oct 2009 11:04:23 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9D771.4010308@sun.com> References: <20091029173434.C615D384@eggemoggin.niobe.net> <4AE9D771.4010308@sun.com> Message-ID: <4AE9D927.30200@sun.com> Kelly O'Hair wrote: > I think this is the fix... it is taking some time to completely > verify it. > > Can someone review this change? See below > And do I have approval to push directly into master? > > ------- > > diff --git a/make/java/redist/Makefile b/make/java/redist/Makefile > --- a/make/java/redist/Makefile > +++ b/make/java/redist/Makefile > @@ -216,7 +216,7 @@ endif # PLATFORM > $(LIB_LOCATION)/$(KERNEL_LOCATION)/$(JVM_NAME): $(HOTSPOT_KERNEL_PATH)/$(JVM_NAME) > $(install-file) > > -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) > +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/jre/lib/$(LIBARCH)/$(LIBJSIG_NAME) > $(install-import-file) How about: -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) --- +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/$(ARCH_VM_SUBDIR)/$(LIBJSIG_NAME) This got me past the problem in my test build. I am re-running a full build from scratch now. Either way you go, feel free to use me as a reviewer. Tim > $(LIB_LOCATION)/$(CLIENT_LOCATION)/$(LIBJSIG_NAME) \ > > > ------- > > A little background, for the record (while I wait for this build to > finish ): > > The files exported by the hotspot makefiles for import to the jdk > come in the shape of a sparse jdk install image, and the jdk Makefile > jdk/make/java/redist/Makefile is responsible for copying or importing > these files from this hotspot exported area, into the final jdk image. > This hotspot area is specified via the > ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH > variables, or could be the ALT_JDK_IMPORT_PATH setting for partial builds. > > If the hotspot team changes what it exports, this jdk makefile may need > to change, and if the hotspot team wants to impact the shape or layout > of the final jdk image with regards to their files, they need to address > this in the jdk/make/java/redist/Makefile. > > -kto From Kelly.Ohair at Sun.COM Thu Oct 29 11:11:25 2009 From: Kelly.Ohair at Sun.COM (Kelly O'Hair) Date: Thu, 29 Oct 2009 11:11:25 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9D927.30200@sun.com> References: <20091029173434.C615D384@eggemoggin.niobe.net> <4AE9D771.4010308@sun.com> <4AE9D927.30200@sun.com> Message-ID: <4AE9DACD.4090703@sun.com> Thanks Tim, both changes do the same thing, I'll use yours. It's on it's way.... -kto Tim Bell wrote: > Kelly O'Hair wrote: >> I think this is the fix... it is taking some time to completely >> verify it. >> >> Can someone review this change? > > See below > >> And do I have approval to push directly into master? >> >> ------- >> >> diff --git a/make/java/redist/Makefile b/make/java/redist/Makefile >> --- a/make/java/redist/Makefile >> +++ b/make/java/redist/Makefile >> @@ -216,7 +216,7 @@ endif # PLATFORM >> $(LIB_LOCATION)/$(KERNEL_LOCATION)/$(JVM_NAME): $(HOTSPOT_KERNEL_PATH)/$(JVM_NAME) >> $(install-file) >> >> -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) >> +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/jre/lib/$(LIBARCH)/$(LIBJSIG_NAME) >> $(install-import-file) > > How about: > > -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) > --- > +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/$(ARCH_VM_SUBDIR)/$(LIBJSIG_NAME) > > > This got me past the problem in my test build. I am re-running a full build from > scratch now. > > Either way you go, feel free to use me as a reviewer. > > Tim > > >> $(LIB_LOCATION)/$(CLIENT_LOCATION)/$(LIBJSIG_NAME) \ >> >> >> ------- >> >> A little background, for the record (while I wait for this build to >> finish ): >> >> The files exported by the hotspot makefiles for import to the jdk >> come in the shape of a sparse jdk install image, and the jdk Makefile >> jdk/make/java/redist/Makefile is responsible for copying or importing >> these files from this hotspot exported area, into the final jdk image. >> This hotspot area is specified via the >> ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH >> variables, or could be the ALT_JDK_IMPORT_PATH setting for partial builds. >> >> If the hotspot team changes what it exports, this jdk makefile may need >> to change, and if the hotspot team wants to impact the shape or layout >> of the final jdk image with regards to their files, they need to address >> this in the jdk/make/java/redist/Makefile. >> >> -kto > From Kelly.Ohair at Sun.COM Thu Oct 29 11:19:21 2009 From: Kelly.Ohair at Sun.COM (Kelly O'Hair) Date: Thu, 29 Oct 2009 11:19:21 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9DB66.7030704@sun.com> References: <20091029180216.83F67384@eggemoggin.niobe.net> <4AE9DB66.7030704@sun.com> Message-ID: <4AE9DCA9.3050204@sun.com> The deed is done. -kto Michael Wilkerson wrote: > Mark Reinhold wrote: >>> Date: Thu, 29 Oct 2009 10:57:05 -0700 >>> From: kelly.ohair at sun.com >>> >> >> >>> I think this is the fix... it is taking some time to completely >>> verify it. >>> >>> Can someone review this change? >>> >>> And do I have approval to push directly into master? >>> >> >> Yes. >> >> Michael -- Could you please re-start the RE builds after Kelly >> pushes his fix? Thanks. >> >> - Mark >> > Yes, will do. > > thanks, > --Michael > From Paul.Hohensee at Sun.COM Thu Oct 29 11:04:47 2009 From: Paul.Hohensee at Sun.COM (Paul Hohensee) Date: Thu, 29 Oct 2009 14:04:47 -0400 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9D771.4010308@sun.com> References: <20091029173434.C615D384@eggemoggin.niobe.net> <4AE9D771.4010308@sun.com> Message-ID: <4AE9D93F.6020600@sun.com> Looks good to me. Kelly O'Hair wrote: > > I think this is the fix... it is taking some time to completely > verify it. > > Can someone review this change? > > And do I have approval to push directly into master? > > ------- > > diff --git a/make/java/redist/Makefile b/make/java/redist/Makefile > --- a/make/java/redist/Makefile > +++ b/make/java/redist/Makefile > @@ -216,7 +216,7 @@ endif # PLATFORM > $(LIB_LOCATION)/$(KERNEL_LOCATION)/$(JVM_NAME): > $(HOTSPOT_KERNEL_PATH)/$(JVM_NAME) > $(install-file) > > -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) > +$(LIB_LOCATION)/$(LIBJSIG_NAME): > $(HOTSPOT_IMPORT_PATH)/jre/lib/$(LIBARCH)/$(LIBJSIG_NAME) > $(install-import-file) > > $(LIB_LOCATION)/$(CLIENT_LOCATION)/$(LIBJSIG_NAME) \ > > > ------- > > A little background, for the record (while I wait for this build to > finish ): > > The files exported by the hotspot makefiles for import to the jdk > come in the shape of a sparse jdk install image, and the jdk Makefile > jdk/make/java/redist/Makefile is responsible for copying or importing > these files from this hotspot exported area, into the final jdk image. > This hotspot area is specified via the > ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH > variables, or could be the ALT_JDK_IMPORT_PATH setting for partial > builds. > > If the hotspot team changes what it exports, this jdk makefile may need > to change, and if the hotspot team wants to impact the shape or layout > of the final jdk image with regards to their files, they need to address > this in the jdk/make/java/redist/Makefile. > > -kto From Vladimir.Kozlov at Sun.COM Thu Oct 29 11:07:28 2009 From: Vladimir.Kozlov at Sun.COM (Vladimir Kozlov) Date: Thu, 29 Oct 2009 11:07:28 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9D927.30200@sun.com> References: <20091029173434.C615D384@eggemoggin.niobe.net> <4AE9D771.4010308@sun.com> <4AE9D927.30200@sun.com> Message-ID: <4AE9D9E0.1030201@sun.com> Kelly, I like Tim's fix. Vladimir Tim Bell wrote: > Kelly O'Hair wrote: >> I think this is the fix... it is taking some time to completely >> verify it. >> >> Can someone review this change? > > See below > >> And do I have approval to push directly into master? >> >> ------- >> >> diff --git a/make/java/redist/Makefile b/make/java/redist/Makefile >> --- a/make/java/redist/Makefile >> +++ b/make/java/redist/Makefile >> @@ -216,7 +216,7 @@ endif # PLATFORM >> $(LIB_LOCATION)/$(KERNEL_LOCATION)/$(JVM_NAME): $(HOTSPOT_KERNEL_PATH)/$(JVM_NAME) >> $(install-file) >> >> -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) >> +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/jre/lib/$(LIBARCH)/$(LIBJSIG_NAME) >> $(install-import-file) > > How about: > > -$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_SERVER_PATH)/$(LIBJSIG_NAME) > --- > +$(LIB_LOCATION)/$(LIBJSIG_NAME): $(HOTSPOT_IMPORT_PATH)/$(ARCH_VM_SUBDIR)/$(LIBJSIG_NAME) > > > This got me past the problem in my test build. I am re-running a full build from > scratch now. > > Either way you go, feel free to use me as a reviewer. > > Tim > > >> $(LIB_LOCATION)/$(CLIENT_LOCATION)/$(LIBJSIG_NAME) \ >> >> >> ------- >> >> A little background, for the record (while I wait for this build to >> finish ): >> >> The files exported by the hotspot makefiles for import to the jdk >> come in the shape of a sparse jdk install image, and the jdk Makefile >> jdk/make/java/redist/Makefile is responsible for copying or importing >> these files from this hotspot exported area, into the final jdk image. >> This hotspot area is specified via the >> ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH >> variables, or could be the ALT_JDK_IMPORT_PATH setting for partial builds. >> >> If the hotspot team changes what it exports, this jdk makefile may need >> to change, and if the hotspot team wants to impact the shape or layout >> of the final jdk image with regards to their files, they need to address >> this in the jdk/make/java/redist/Makefile. >> >> -kto > From Michael.Wilkerson at Sun.COM Thu Oct 29 11:13:58 2009 From: Michael.Wilkerson at Sun.COM (Michael Wilkerson) Date: Thu, 29 Oct 2009 11:13:58 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <20091029180216.83F67384@eggemoggin.niobe.net> References: <20091029180216.83F67384@eggemoggin.niobe.net> Message-ID: <4AE9DB66.7030704@sun.com> Mark Reinhold wrote: >> Date: Thu, 29 Oct 2009 10:57:05 -0700 >> From: kelly.ohair at sun.com >> > > >> I think this is the fix... it is taking some time to completely >> verify it. >> >> Can someone review this change? >> >> And do I have approval to push directly into master? >> > > Yes. > > Michael -- Could you please re-start the RE builds after Kelly > pushes his fix? Thanks. > > - Mark > Yes, will do. thanks, --Michael From Michael.Wilkerson at Sun.COM Thu Oct 29 11:23:47 2009 From: Michael.Wilkerson at Sun.COM (Michael Wilkerson) Date: Thu, 29 Oct 2009 11:23:47 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9DCA9.3050204@sun.com> References: <20091029180216.83F67384@eggemoggin.niobe.net> <4AE9DB66.7030704@sun.com> <4AE9DCA9.3050204@sun.com> Message-ID: <4AE9DDB3.40706@sun.com> OK, ....RE cleanup scripts are running right now, as soon as that's done I'll restart... --Michael Kelly O'Hair wrote: > The deed is done. > > -kto > > Michael Wilkerson wrote: >> Mark Reinhold wrote: >>>> Date: Thu, 29 Oct 2009 10:57:05 -0700 >>>> From: kelly.ohair at sun.com >>>> >>> >>> >>>> I think this is the fix... it is taking some time to completely >>>> verify it. >>>> >>>> Can someone review this change? >>>> >>>> And do I have approval to push directly into master? >>>> >>> >>> Yes. >>> >>> Michael -- Could you please re-start the RE builds after Kelly >>> pushes his fix? Thanks. >>> >>> - Mark >>> >> Yes, will do. >> >> thanks, >> --Michael >> From David.Katleman at Sun.COM Thu Oct 29 13:34:12 2009 From: David.Katleman at Sun.COM (david katleman) Date: Thu, 29 Oct 2009 13:34:12 -0700 Subject: JDK 7 linux/solaris build failures In-Reply-To: <4AE9D771.4010308@sun.com> References: <20091029173434.C615D384@eggemoggin.niobe.net> <4AE9D771.4010308@sun.com> Message-ID: <4AE9FC44.3080609@sun.com> Kelly O'Hair wrote: > This hotspot area is specified via the > ALT_HOTSPOT_{IMPORT,SERVER,CLIENT}_PATH > variables, or could be the ALT_JDK_IMPORT_PATH setting for partial > builds. > > If the hotspot team changes what it exports, this jdk makefile may need > to change, and if the hotspot team wants to impact the shape or layout > of the final jdk image with regards to their files, they need to address > this in the jdk/make/java/redist/Makefile. And if hotspot does change what it exports, there really needs to be a build of jdk to confirm their modifications don't result in a broken build down the line. This shouldn't have gone back to the master without such testing, and especially not on the eve of a promotion. Dave