[CIG-SHORT] PyLith tutorial at CIG16: things you need to do **before** the tutorial

[Brad Aagaard]

Ravi,

I am looking for the **PyLith** config.log, not the installer 
config.log. It will be in the pylith-build directory.

Thanks,
Brad


On 06/13/2016 08:44 AM, Ravi Kanda wrote:
> Brad,
>
> The "=yes" in the infinite loop is not printed within the config.log,
> only in the terminal (screen).  I left config running for 15 min, and
> was still printing this string to the screen.  I had attached both the
> config.log and a stdout capture (with millions of "=yes" lines stripped
> out) to my original message - perhaps they got stripped off?  Attaching
> them again ...
>
> Thanks,
> Ravi.
> ====================================
>
> On 06/13/2016 09:29 AM, Brad Aagaard wrote:
>> Ravi,
>>
>> Please send the config.log file for the PyLith configure that is in an
>> infinite loop, so we can diagnose this further. It it is large, gzip
>> the log file.
>>
>> Thanks,
>> Brad
>>
>>
>> On 06/13/2016 08:13 AM, Ravi Kanda wrote:
>>> Hi Brad,
>>>
>>> I am trying to install Pylith on Ubuntu 14 with the latest source
>>> installer tarball (2.1.0-3).  I already have previous (apt-get)
>>> installations of mpi2, python/numpy, netcdf4, proj4, and cppunit. So, I
>>> am using the "DESKTOP-LINUX-MINIMAL" options as described in the
>>> "INSTALL" file.  Running config from directory $PYLITHROOT/src/pylith,
>>> which also contains the pylith installer directory,
>>> 'pylith-installer-2.1.0-3'.  I am running into config problems (attached
>>> CONFIG.LOG), which I hope is due to a stupid mistake on my part:
>>>
>>> (1) I get two "command not found" or "No such file or directory" errors
>>> when I unset or set, respectively, the CPPFLAGS env variable related to
>>> the following:
>>>       - configure:22278: checking cppunit/TestRunner.h presence
>>>       - configure:23244: checking netcdfcpp.h presence
>>> The appropriate include-dir for the above on my system is /usr/include,
>>> which I thought was 'standard'.  Also, in both cases, I see something
>>> like the following below each of these items (line numbers purged):
>>>      "... accepted by the compiler, rejected by the preprocessor!
>>>       ... WARNING: ... : proceeding with the compiler's result"
>>>
>>> (2)  Then configure seems to go into an "infinite loop" after looking
>>> for 'netcdfcpp.h', printing out lines containing only the string "=yes"
>>> (see attached STDOUT capture).  When I look at system processes in
>>> 'top', I see an executable with the name 'yes' running away.
>>>
>>> Any tips will be greatly appreciated!
>>>
>>> Thanks,
>>> Ravi.
>>> ===============================================
>>>
>>> On 06/06/2016 10:48 AM, Brad Aagaard wrote:
>>>> PyLith Tutorial: Things you need to do **before** the tutorial
>>>>
>>>>   IMPORTANT: We expect you to make use of the extensive documentation
>>>>   and online instructions **BEFORE** the tutorial. This will allow us
>>>>   to make better use of the limited time for tutorials at the
>>>>   CIG16 meeting.
>>>>
>>>>   1. Install PyLith, ParaView, and CUBIT/Trelis. See the instructions
>>>>   at the bottom of the tutorial webpage. PyLith and Paraview are
>>>>   open-source with binary packages available; Trelis has a 30-day
>>>>   free trial. Some institutions may already have Trelis licenses, so
>>>>   ask around your department to see if you can use an existing
>>>>   license.
>>>>
>>>>   For the tutorial, we will be using PyLith v2.2.0, ParaView 5.0.1,
>>>>   CUBIT 15.1 (US gov't agencies), and Trelis 16.0.2 (everyone
>>>>   else). PyLith v2.2.0 will be released shortly before the
>>>>   tutorial. It will include some bug fixes and improvements to the
>>>>   manual. We recommend getting familiar with v2.1.0 as it is very
>>>>   similar.
>>>>
>>>>   2. Read the software manuals!
>>>>
>>>>   3. View the videos from last year's tutorials that are available
>>>>   online at https://wiki.geodynamics.org/software:pylith:start (2015
>>>>   Crustal Deformation Modeling Tutorial). Work through the examples
>>>>   that are discused.
>>>>
>>>>   4. If possible, come to the workshop with an idea of what you want
>>>>   to model. In general, we recommend starting with a small, 2-D
>>>>   simplified version to play with during the workshop and then
>>>>   gradually adding complexity to achieve the desired final
>>>>   model. Participants who follow this advice get the most out of the
>>>>   tutorials!
>>>>
>>>> Regards,
>>>> Brad Aagaard
>>>> Charles Williams
>>>> Matt Knepley
>>>
>>
>