[Thread Prev][Thread Next][Index]

Re: [ferret_users] build pyferret_v710



Hi Martin,

Since you are getting to the 'yes? ' prompt, it appears the main pyferret module itself is not the source of the problem.  Since it appears it is when you try to open a displayed window, it probably is the pipedviewer python code (which is part of PyFerret) causing the error.

In this release, all the python code was updated to work under either Python2.x or Python3.x, and use PyQt5, if available, otherwise PyQt4.  So one thing to see is if PyQt5 happens to be installed in your python2.7, something like:

$ python2.7
>>> import PyQt5
>>> ^D

If the import was successful (no ImportError raised) then that would be one difference from earlier code.  My understanding is that PyQt5 really needs Python3.x, so this combination could be the problem.  There were also some changes to the python code for the 2.x and 3.x compatibility, but I am not sure why those changes would give the error message you show. 

The viewers have little tests that are easily run:

$ cd $FER_DIR/lib/python2.7/site-packages/pipedviewer/
$ python2.7 pipedviewerpq.py
(This will display a little dialog box with a button.  Every time you press the button it will print a drawing command to the terminal as well as issue that command to the viewer.  It should display a window, draw various things to that window, save the final plot as a PDF and PNG, and exit.)

I suspect that when you run this test you will immediately get the error message.  If you do have both PyQt5 and PyQt4 in Python2.7, you might try commenting out the attempt to import PyQt5, and just import PyQt4 instead (in all the .py files in the directory).

Karl


On Wed, Mar 1, 2017 at 9:18 AM, Martin Schmidt <martin.schmidt@xxxxxxxxxxxxxxxxx> wrote:
Hi,
I am trying to build the new pyferret. The machine is the data processing server of a compute center. Hence, the linux release is not the most recent one and many libraries and python are not in the system path but are available via a "modules" environment.

Anyway, with gcc-49 and python-2.7.6 I could build pyferret_v700, which works very well. But trying to repeat the compilation with the new ferret release I end up with
        NOAA/PMEL TMAP
        PyFerret v7.1 (optimized)
        Linux 3.0.101-0.47.86.1.11753.0.PTF-defa - 03/01/17
         1-Mar-17 18:07

yes? set win 1
/sw/tools/python/2.7.6/generic/bin/python: symbol lookup error: /sw/tools/python/2.7.6/generic/lib/libxcb-xkb.so.1: undefined symbol: xcb_sumof

Before I pack and send the site_specific.mk, platform_specific.mk.x86_64-linux and the Makefile, which are mostly unchanged, I would like to ask, if there are changes in pyferret which may explain the failure. Is there eventually a workaround?

Best,
Martin Schmidt



--
Karl M. Smith, Ph.D.
JISAO Univ. Wash. and PMEL NOAA
"The contents of this message are mine personally and do
not necessarily reflect any position of the Government
or the National Oceanic and Atmospheric Administration."

[Thread Prev][Thread Next][Index]
Contact Us
Dept of Commerce / NOAA / OAR / PMEL / Ferret

Privacy Policy | Disclaimer | Accessibility Statement