Re: [AMBER] compile errors with AmberTools1.5 + Amber11

From: Robert Wohlhueter <bobwohlhueter.earthlink.net>
Date: Sat, 30 Apr 2011 20:04:04 -0400

Jason,

A couple of comments on your comments: By "clean slate" I mean really
clean, i.e. removing the $AMBERHOME dir entirely, re-extracting tars,
re-patching, re-configuring. And I have AMBERHOME defined in root's
.cshrc. (Nor does it make any difference if I `chown` of AMBERHOME to
non-root, and build as a non-root user. So these aren't issues.

I can't expect you to fix my Ubuntu installation of Python; it's a
standard installation via the Synaptic Package Manager, and I don't see
anything wrong with it: 2.5, 2.6, and 3.1 are all installed under
/usr/lib; all of these contain a "os.py", none contain a "sys.py",
specifying absolute path to 2.6 either via PYTHONHOME or hard-wiring in
AT15_Amber11.py doesn't free me from the "import site" failed message.

One rock left unturned is the python syntax errors which I get when
omitting the "-nopython" flag while configuring. These errors prevent me
from configuring without the "-nopython" flag, and thus force me to rely
on my (broken?) Unbuntu Python installations. These errors all have to
do with use of \n for "newline" in the python code. [SyntaxError:
("(unicode error) \\N escapes not supported (can't load > unicodedata
module)"]). My search of the internet seems to confirm this, and says
that the proper "newline" character for python using unicode is
"u2424". Is it not possible/preferable for amber to patch its python
code to reflect this?

Bob W.



On 4/29/11 3:52 PM, Jason Swails wrote:
> On Fri, Apr 29, 2011 at 4:06 PM, Robert Wohlhueter<
> bobwohlhueter.earthlink.net> wrote:
>
>> Hi Jason,
>> Well, configuring (a clean slate) with -nopython flag circumvented the
>> weird python errors, but brought me other troubles (see below):
>> apparently my local python installation is not up to snuff. On a long
>> shot, I tried the `AT15_Amber11.py` script again anyway, but got the
>> same errors as previously (can't import "site", "os").
>>
>> Not to muddy the waters, but I also tried a clean build with the latest
>> (half an hour ago) bugfix.all for AT 1.5. You might be interested that
>> it gives a "malformed patch" error (see below), but didn't affect my
>> python problem.
>>
>> Bob Wohlhueter
>>
>> *****************************************************************
>> cpp -traditional -P -DBINTRAJ lapack.f> _lapack.f
>> gfortran -c -O0 -o lapack.o _lapack.f
>> cpp -traditional -P -DBINTRAJ resp.f> _resp.f
>> gfortran -c -O0 -o resp.o _resp.f
>> gfortran \
>> lapack.o resp.o -o resp
>> gfortran -o calcpka calcpka.f90
>> ./py_setup.sh
>> Using system python
>> Error importing cpinutil.py python modules! cpinutil will not work.
>> make[1]: *** [python] Error 1
>> make[1]: Leaving directory `/usr/local/lib/amber11/AmberTools/src/etc'
>> make: *** [serial] Error 2
>>
>>
>> root.winter-linux: ...lib/amber11 (25%)# ./AT15_Amber11.py
>>
> Ooh, I would not do this as root. In several places, I think, the Amber
> install references files in $AMBERHOME/bin. If AMBERHOME is blank, this
> will just be parsed as /bin/, which is one of the last places you want to
> mess around in (especially if it's *not* your scripts, so you don't know
> what it's doing).
>
> ./AT15_Amber11.py
> 'import site' failed; use -v for traceback
>> Traceback (most recent call last):
>> File "./AT15_Amber11.py", line 8, in<module>
>> import os, sys
>> ImportError: No module named os
>>
> Not good... Either you didn't start from a clean tree and it's still trying
> to use the broken python from $AMBERHOME/bin, or your system python is
> busted (very not good). The first means just get rid of the whole
> AmberTools 1.5 directory and start over. If it's the second one, and you're
> using Ubuntu, then something is seriously wrong (I use Ubuntu all the time
> and have never seen this...).
>
>
>> ****************bugfix.all patch
>> error*************************************************
>> patching file AmberTools/src/cpptraj/src/AmberNetcdf.cpp
>> patching file AmberTools/src/mmpbsa_py/MMPBSA_mods/alamdcrd.py
>> patching file AmberTools/src/mmpbsa_py/setup.sh
>> patch: **** malformed patch at line 3176: Workarounds: none
>>
> I don't know why this is happening, but I can reproduce it. If you just
> delete the last 2 lines it goes away... but not sure what'll happen when we
> add another patch.
>
> HTH,
> Jason
>


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Apr 30 2011 - 17:30:02 PDT
Custom Search