Add version information to jvmti demos.

[Sean Chou]

2011/6/8 Daniel D. Daugherty <daniel.daugherty at oracle.com>

> Adding Zhengyu since he is the RE on 6988099 and I'm not sure
> if he is on the serviceability-dev alias...
>
> Zhengyu, I don't know if you have a fix already in process for
> this issue or not.
>
> Comments on the Makefile change below.
>
>
> Dan
>
>
> On 6/6/2011 8:59 PM, David Holmes wrote:
>
>> Sean,
>>
>> Because this is a JVMTI demo it falls under serviceability, so I've cc'ed
>> that list.
>>
>> David
>>
>> Sean Chou said the following on 06/07/11 12:18:
>>
>>>
>>> Is there any comment ?
>>>
>>> 2011/5/26 Sean Chou <zhouyx at linux.vnet.ibm.com <mailto:
>>> zhouyx at linux.vnet.ibm.com>>
>>>
>>>    Hi all,
>>>
>>>       I made a simple patch for bug
>>>    http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6988099     (jvmti
>>> demos missing Publisher (COMPANY resource) in dll and exe
>>>    files on windows) .        Will anyone have a look at my patch ?
>>> Thanks.
>>>
>>>
>>>    diff --git make/common/Demo.gmk make/common/Demo.gmk
>>>    index a3788ae..8c43cf9 100644
>>>    --- make/common/Demo.gmk
>>>    +++ make/common/Demo.gmk
>>>    @@ -158,6 +158,7 @@ ifneq ($(strip $(DEMO_ALL_NATIVE_SOURCES)),)
>>>         #       bit between them.
>>>         LINK.demo   = $(LINK.c)
>>>         LDLIBS.demo = $(EXTRA_LIBS) $(LFLAGS_$(COMPILER_VERSION))
>>>    +    DEMO_VERSION_INFO = $(OBJDIR)/$(LIBRARY).res
>>>       else
>>>         ifneq ($(DEMO_NEEDS_CPP),)
>>>           LINK.demo   = $(LINK.cpp)
>>>    @@ -288,6 +289,13 @@ ifndef DEMO_SKIP_SRCZIP
>>>         $(install-file)
>>>     endif
>>>
>>
> The change below is not bracketed by "ifeq ($(PLATFORM),windows)"
> but it certainly looks Windows specific.
>
>
Yes, it is Windows specific and only used by RC.exe. I copied this piece of
 flag from Library.gmk, it is not bracketed by "ifeq ($(PLATFORM),windows)"
 in that file.



>
>
>          +# JDK name required here
>>>    +RC_FLAGS += /D "JDK_FNAME=$(LIBRARY).dll" \
>>>    +            /D "JDK_INTERNAL_NAME=$(LIBRARY)" \
>>>    +            /D "JDK_FTYPE=0x2L"
>>>    +
>>>    +
>>>    +
>>>     # Native library building
>>>     ifdef DEMO_LIBRARY
>>>         @@ -308,8 +316,11 @@ $(OBJDIR)/%.$(OBJECT_SUFFIX):
>>> $(DEMO_BUILD_SRCDIR)/%.cpp
>>>       # Actual creation of the native shared library (C++ and C are
>>> different)
>>>     $(DEMO_LIBRARY): $(DEMO_FULL_OBJECTS)
>>>         @$(prep-target)
>>>    +  ifeq ($(PLATFORM),windows)
>>>    +    $(RC) $(RC_FLAGS) $(CC_OBJECT_OUTPUT_FLAG)$(DEMO_VERSION_INFO)
>>> $(VERSIONINFO_RESOURCE)
>>>    +  endif
>>>         $(LINK.demo) $(SHARED_LIBRARY_FLAG) $(CC_PROGRAM_OUTPUT_FLAG)$@ \
>>>    -        $(DEMO_FULL_OBJECTS) $(LDLIBS.demo)
>>>    +        $(DEMO_FULL_OBJECTS) $(LDLIBS.demo) $(DEMO_VERSION_INFO)
>>>
>>
> The above change uses "$(DEMO_VERSION_INFO)" but that variable is
> only set under "ifeq ($(PLATFORM),windows)" above. That won't break
> make, but I wanted to be sure that was intentional.
>
> "$(DEMO_VERSION_INFO)" should be left blank on platforms other than
windows. And
I found it would be blank if it was not set, so I removed the code to set it
blank.



>
>
>          @$(call binary_file_verification,$@)
>>>            # Generation of any javah include file, make sure objects are
>>> dependent on it
>>>
>>>
>>>
>>>    --     Best Regards,
>>>    Sean Chou
>>>
>>>
>>>
>>>
>>> --
>>> Best Regards,
>>> Sean Chou
>>>
>>>


-- 
Best Regards,
Sean Chou
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/build-dev/attachments/20110608/f305833d/attachment.htm>