[FLASH-USERS] buld problem, Makefile generation

Fri Jun 14 14:53:40 EDT 2013

Hi Anthony,

Okay, I guess it must have something to do with my python setup, though I
don't quite see how.  I am now using python 2.7 that my system folks
installed whereas before I was using anaconda python 2.7.  I'm attaching
the outputs you asked for, but they're almost identical and not too
illuminating.  The key would be to find what code writes the line in the
Makefile:
LIB := $(LIB_HDF5) $(LIB_MPI) $(LIB)
I guess another thing that would be helpful is to know if others are able
to compile successfully with that line with LIB rather than LIB_OPT in
their Makefiles.  In that case then the problem is with the definition of
LIB at that point in the Makefile rather than the line itself.

Jon

________________________________________________________
Jonathan D. Slavin                 Harvard-Smithsonian CfA
jslavin at cfa.harvard.edu       60 Garden Street, MS 83
phone: (617) 496-7981       Cambridge, MA 02138-1516
fax: (617) 496-7577            USA
________________________________________________________

On Fri, Jun 14, 2013 at 2:04 PM, Anthony Scopatz <scopatz at flash.uchicago.edu
> wrote:

> Hi Jon,
>
> flmake doesn't change the way that the setup script works at all.  Their
> implementations are separate and flmake doesn't add anything to the flash
> source dir, it only pulls things out.  If you are seeing a change with
> setup now vs before, this is likely because some part of your environment
> has changed.  Since it is affecting Makefiles my guess is that the site
> that is being selected is different and/or your hostname has changed.
>
> Can you please post the output of ./setup Sedov -auto and flmake setup?
> Thanks.
>
> Be Well
> Anthony
>
>
> On Fri, Jun 14, 2013 at 12:53 PM, Slavin, Jonathan <
> jslavin at cfa.harvard.edu> wrote:
>
>> Hi Anthony,
>>
>> So I did leave something out of my message.  That is that when, today, I
>> did
>> ./setup Sedov -auto
>> (in the main FLASH directory) and then
>> cd object
>> make
>> I had the same problem as with using flmake.  But when I followed the
>> same steps a few days ago, I didn't have this problem.  The only thing that
>> has changed in the meantime is that I used flmake.  So it seems to me that
>> flmake must have generated some file that has changed the way setup creates
>> the Makefile.
>> The flmake commands were (in a sub-directory temp under main dir):
>> flmake setup Sedov -auto
>> flmake build
>>
>> Which routine actually writes the Makefile?  It seems to involve a bunch
>> of routines in FLASH4.0.1/bin, at least using setup.
>>
>> Jon
>>
>> ________________________________________________________
>> Jonathan D. Slavin                 Harvard-Smithsonian CfA
>> jslavin at cfa.harvard.edu       60 Garden Street, MS 83
>> phone: (617) 496-7981       Cambridge, MA 02138-1516
>> fax: (617) 496-7577            USA
>> ________________________________________________________
>>
>>
>>
>> On Fri, Jun 14, 2013 at 1:17 PM, Anthony Scopatz <
>> scopatz at flash.uchicago.edu> wrote:
>>
>>> Hi Jon,
>>>
>>> I am not sure what would cause this difference unless it is flmake is
>>> picking up a different Makefile than what the setup script does.  What
>>> setup script line did you use?  What were the flmake commands that you
>>> issued to get this, and what are the contents of your flashrc.py if you
>>> have one?
>>>
>>> Thanks.
>>>
>>> Be Well
>>> Anthony
>>>
>>>
>>> On Fri, Jun 14, 2013 at 11:57 AM, Slavin, Jonathan <
>>> jslavin at cfa.harvard.edu> wrote:
>>>
>>>> Hi all,
>>>>
>>>> I've run into a completely different flmake related problem.  Earlier,
>>>> following the Quickstart instructions (and after a bunch of fiddling with
>>>> the Makefile) I was able to get FLASH to build and then run successfully.
>>>>  But when I tried the flmake approach, the build failed with an odd error.
>>>>  The error was in the linking.
>>>>
>>>> Comparing the successful Makefile and unsuccessful Makefile I got this
>>>> from diff:
>>>> 51,52c51,52
>>>> < LIB := $(LIB_HDF5) $(LIB_MPI) $(LIB_OPT)
>>>> < CFLAGS := $(CFLAGS_HDF5) $(CFLAGS_OPT)
>>>> ---
>>>> > LIB := $(LIB_HDF5) $(LIB_MPI) $(LIB)
>>>> > CFLAGS := $(CFLAGS_HDF5) $(CFLAGS_MPI) $(CFLAGS_OPT)
>>>> I edited the unsuccesful one and replaced the $(LIB) with $(LIB_OPT)
>>>> and removed $(CFLAGS_MPI) (though I think the latter has no effect) and
>>>> voila, success again.
>>>>
>>>> So my question is, is that a bug in flmake that causes the generated
>>>> Makefile to use $(LIB) in that place instead of $(LIB_OPT)?  Or do I have
>>>> some odd configuration that causes that.  I'm not so clear on how the
>>>> Makefile is generated.
>>>>
>>>> Jon
>>>> ________________________________________________________
>>>> Jonathan D. Slavin                 Harvard-Smithsonian CfA
>>>> jslavin at cfa.harvard.edu       60 Garden Street, MS 83
>>>> phone: (617) 496-7981       Cambridge, MA 02138-1516
>>>> fax: (617) 496-7577            USA
>>>> ________________________________________________________
>>>>
>>>>
>>>
>>>
>>> --
>>> Anthony Scopatz
>>> Post-doctoral Scholar
>>> The FLASH Center for Computational Science
>>> Dept. of Astronomy and Astrophysics
>>> The University of Chicago
>>> cell: (512) U2-PU239
>>> office: (773) 834-2705
>>> fax:(773) 834-3230
>>> http://flash.uchicago.edu
>>>
>>
>>
>
>
> --
> Anthony Scopatz
> Post-doctoral Scholar
> The FLASH Center for Computational Science
> Dept. of Astronomy and Astrophysics
> The University of Chicago
> cell: (512) U2-PU239
> office: (773) 834-2705
> fax:(773) 834-3230
> http://flash.uchicago.edu
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20130614/a4ca05c7/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: setup.log
Type: application/octet-stream
Size: 2492 bytes
Desc: not available
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20130614/a4ca05c7/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: flmake.log
Type: application/octet-stream
Size: 2741 bytes
Desc: not available
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20130614/a4ca05c7/attachment-0001.obj>