Re: [rvm-research] Researching the Jikes RVM

Hi,

On 05.03.2013 22:13, Kathiravelu Pradeeban wrote:
> Hi,
> Thanks for the pointers.

An additional hint: If you're trying to build code from the old GSoC
projects, it's probably easier to use the diffs that are contained in
the code samples submitted to Google (e.g. at
http://code.google.com/p/google-summer-of-code-2010-jikesrvm/ ) than to
check out something from the old repositories. There have been some
fixes for build errors since then.

> 
> I was trying to build jikes with GCspy [1] heap visualization
> framework, using the command,
> ant -Dconfig.name=BaseBaseSemiSpaceGCSpy
> -Dconfig.include.gcspy-client=1 -Dhost.name=x86_64-linux
> -Dtarget.name=x86_64-linux -Dcp.enable.gtk-peer="--disable-gtk-peer"

The development of the GCSpy framework has seriously lagged behind that
of the rest of the Jikes RVM. I don't know the current status of GCSpy
but I suspect that it's not good: older GSoC idea lists included an idea
for improvements to GCSpy and there's a related JIRA issue at
http://jira.codehaus.org/browse/RVM-388 .

> I fixed a minor issue - gcspy.dir was pointing to a wrong location in
> build/components/gcspy.xml (patch attached herewith, as gcspy.diff).

Please open an issue in our bugtracker (see
http://www.jikesrvm.org/Reporting+Bugs for details) and attach your
patch (or mercurial bundle). You'll also need to attach one of the
contributor statements from http://www.jikesrvm.org/Contributions (found
at the end of the page).

> However, now I am getting the below exception. Could you pls help me
> rectify this issue?
> 
> build:
>      [exec] rm -rf obj
>      [exec] rm -rf bin
> 
> .....
> 
>      [exec] gcc -shared -o lib/libgcspy.so  obj/gcspy_array_input.o
> obj/gcspy_array_output.o  obj/gcspy_buffered_input.o
> obj/gcspy_buffered_output.o  obj/gcspy_color_db.o
> obj/gcspy_command_stream.o  obj/gcspy_comm.o  obj/gcspy_d_utils.o
> obj/gcspy_gc_driver.o  obj/gcspy_gc_stream.o  obj/gcspy_interpreter.o
> obj/gcspy_main_server.o  obj/gcspy_shape_stream.o  obj/gcspy_timer.o
> obj/gcspy_utils.o
>      [exec] /usr/bin/ld: obj/gcspy_buffered_input.o: relocation
> R_X86_64_32 against `.rodata' can not be used when making a shared
> object; recompile with -fPIC
>      [exec] obj/gcspy_buffered_input.o: could not read symbols: Bad value
>      [exec] collect2: ld returned 1 exit status
>      [exec] make: *** [lib/libgcspy.so] Error 1
> BUILD FAILED
> /home/pradeeban/jikesrvm/build.xml:301: The following error occurred
> while executing this line:
> /home/pradeeban/jikesrvm/build.xml:311: The following error occurred
> while executing this line:
> /home/pradeeban/jikesrvm/build/components/gcspy.xml:122: The following
> error occurred while executing this line:
> /home/pradeeban/jikesrvm/build/components/base.xml:70: The following
> error occurred while executing this line:
> /home/pradeeban/jikesrvm/build/components/base.xml:76: The following
> error occurred while executing this line:
> /home/pradeeban/jikesrvm/build/components/gcspy.xml:101: exec returned: 2
> 
> Full log is attached as log.txt, for the interested.
> 
> [1] http://jikesrvm.org/Using+GCSpy

The GCSpy page claims that it's only supported on ia32-linux (but I
don't know if that's still true). The x86_64 configurations are 32-bit
so it might be the case that some of the libraries are getting compiled
for the wrong address size.


Kind regards,

Erik Brangs

------------------------------------------------------------------------------
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev