[SPAM]Re: JDK-7082220 : Visual Studio projects broken after change: backport to OpenJDK7/hotspot

Tue Sep 24 05:10:05 UTC 2013

Sean

I found a bypass for run a full build. here the explanation

the Z:\DEV\OpenJDK7u\hotspot\build\windows\windows_i486_compiler1\local.make 
contains
# Generated file
Variant=compiler1
WorkSpace=Z:\DEV\OpenJDK7u\hotspot
BootStrapDir=C:\Progra~1\Java\jdk1.6.0_35
BuildUser=FrancisANDRE
HS_VER=24,0,0,56
HS_DOTVER=24.0.0.56
HS_COMPANY=
HS_FILEDESC=OpenJDK Client VM
HOTSPOT_VM_DISTRO=OpenJDK
*OPENJDK=Z:\DEV\OpenJDK7 *
HS_COPYRIGHT=Copyright 2013
HS_NAME=OpenJDK 7.0
HS_BUILD_VER=24.0-b56-internal
BUILD_WIN_SA=1
SA_BUILD_VERSION=24.0-b56-internal
SA_INCLUDE=
SA_LIB=
JDK_VER=7,0,0,0
JDK_DOTVER=7.0.0.0
JRE_RELEASE_VER=1.7.0
BUILDARCH=i486
Platform_arch=x86
Platform_arch_model=x86_32
CXX=cl
LD=
MT=
RC=rc
MSC_VER=1600
LD_VER=1000
LD_VER_RAW=10.00.30319.01
ENABLE_FULL_DEBUG_SYMBOLS=1
ZIP_DEBUGINFO_FILES=1
RM=rm -f
ZIPEXE=zip

while the 
Z:\DEV\OpenJDK7u\hotspot\build\windows\windows_i486_compiler1\generated\local.make 
is empty
# Empty

By adding the line OPENJDK=true in the latest local.make which is now
# Empty
OPENJDK=true

the build succeeds for the compiler1, as well for compiler2 with the equivalent 
bypass.

So, one should find why the OPENJDK  is set to Z:\DEV\OpenJDK7 upper directory 
local.make instead of true.

the
Le 23/09/2013 19:46, Seán Coffey a écrit :
> I don't see the failure either. Ran an OpenJDK test job through JPRT.  I 
> wonder if make version and/or shell env could have an impact ? Can you share 
> more details Francis ?
>
> I see an OPENJDK condition expression in trace.make and wonder if that's being 
> parsed incorrectly for this issue ?
>
> http://hg.openjdk.java.net/jdk7u/jdk7u-dev/hotspot/file/tip/make/windows/makefiles/trace.make
>
> regards,
> Sean.
>
> On 23/09/13 17:54, Erik Joelsson wrote:
>> It's certainly not supposed to fail like that. I'm not able to reproduce this 
>> problem cloning from http://hg.openjdk.java.net/jdk7u/jdk7u-dev. From where 
>> did you get your sources?
>>
>> /Erik
>>
>> On 2013-09-23 17:51, Francis ANDRE wrote:
>>> Sean
>>>
>>> The point is that OpenJDK7u is not open source while OpenJDK7 is... 
>>> Following is a snippet of the make log joined...
>>> That's the same story for the HSX project. Thus either the OpenJDK7u is an 
>>> open source project and then the makefiles are wrong, either it is not and 
>>> then I need the backport on the OpenJDK7 project.
>>>
>>> Regards
>>>
>>> Francis
>>>
>>>
>>>         mv  ad_x86_32.cpp  ad_x86_32.hpp  ad_x86_32_clone.cpp 
>>> ad_x86_32_expand.cpp  ad_x86_32_forma
>>> t.cpp  ad_x86_32_gen.cpp  ad_x86_32_misc.cpp ad_x86_32_peephole.cpp  
>>> ad_x86_32_pipeline.cpp  adGlob
>>> als_x86_32.hpp  dfa_x86_32.cpp adfiles/
>>>         "C:\Progra~1\Java\jdk1.6.0_35\bin\javac" -g -encoding ascii -source 
>>> 6 -target 6 -d jvmtifile
>>> s Z:\DEV\OpenJDK7u\hotspot/src/share/vm/prims/jvmtiGen.java
>>> Generating jvmtifiles/jvmtiEnv.hpp
>>> Generating jvmtifiles/jvmtiEnter.cpp
>>> Generating jvmtifiles/jvmtiEnterTrace.cpp
>>> Generating jvmtifiles/jvmtiEnvRecommended.cpp
>>> Generating jvmtifiles/bytecodeInterpreterWithChecks.cpp
>>> Generating jvmtifiles/jvmti.h
>>> *NMAKE : fatal error U1073: incapable d'obtenir 
>>> 'Z:\DEV\OpenJDK7u\hotspot/src/closed/share/vm/trace/t**
>>> **raceEventClasses.xsl'*
>>> Stop.
>>> NMAKE : fatal error U1077: 'cd' : code retour '0x2'
>>> Stop.
>>> NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio 
>>> 10.0\VC\BIN\nmake.EXE"' : code
>>>  retour '0x2'
>>> Stop.
>>> Makefile:191: recipe for target `generic_build2' failed
>>> make[1]: *** [generic_build2] Error 2
>>> make[1] : on quitte le répertoire « /cygdrive/z/DEV/OpenJDK7u/hotspot/make »
>>> Makefile:151: recipe for target `jvmg' failed
>>> make: *** [jvmg] Error 2
>>>
>>> FrancisANDRE at idefix /cygdrive/z/DEV/OpenJDK7u/hotspot/make
>>> Le 22/09/2013 20:40, Seán Coffey a écrit :
>>>> Francis,
>>>>
>>>> the OpenJDK 7 repository corresponds to the JDK 7 GA release made a few 
>>>> years ago. The OpenJDK 7u repository is used to gather fixes for the JDK 7 
>>>> update releases.
>>>>
>>>> Isn't this patch already in the updates ? (7u2)
>>>>
>>>> http://hg.openjdk.java.net/jdk7u/jdk7u/hotspot/rev/2f27ed2a98fa
>>>> https://bugs.openjdk.java.net/browse/JDK-2213981
>>>>
>>>> regards,
>>>> Sean.
>>>>
>>>> On 22/09/2013 09:37, Francis ANDRE wrote:
>>>>> HI
>>>>>
>>>>> the patch "JDK-7082220 : Visual Studio projects broken after change 
>>>>> 7016797: Hotspot: securely/restrictive load dlls"  is missing in the 
>>>>> OpenJDK7/hotspot/src/share/tools/ProjectCreator/WinGammaPlatformVC10.java. 
>>>>> I applied it localy but some other OpenJDK developpers on WXP/VS2010 would 
>>>>> benefit of this patch if it was applied to the OpenJDK7 forest. So I am 
>>>>> wondering how to request a backport of this patch to the OpenJDK7/hotspot 
>>>>> repository?
>>>>>
>>>>> Regards
>>>>>
>>>>> FA
>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>