Wayne Boucher wrote:
>Hello,
>
>CCPNMR_TOP_DIR, well more particularly the PYTHONPATH, should point to
>wherever the code is that you want to use. The standard analysis script
>created in the ccpnmr/bin directory says something like:
>
>export CCPNMR_TOP_DIR=/Users/wb104/ccpnmr
>export PYTHONPATH=.:$CCPNMR_TOP_DIR/ccpnmr1.0/python
>export LD_LIBRARY_PATH=/sw/lib:/sw/lib:/usr/X11R6/lib
>export TCL_LIBRARY=/sw/lib/tcl8.4
>export TK_LIBRARY=/sw/lib/tk8.4
>/sw/bin/python -i -O
>$CCPNMR_TOP_DIR/ccpnmr1.0/python/ccpnmr/analysis/AnalysisGui.py $1 $2 $3 $4 $5
>
>and the PYTHONPATH line and last lines both point to the ccpnmr1.0
>directory, and normally that is (directly) inside ccpnmr. What the update
>script is supposed to do is change the ccpnmr1.0 directory to
>ccpnmr1.0_1.0.x (for whatever x the old version was) and then place the
>new code in ccpnmr1.0. That way the above analysis script works
>automatically on the new version.
>
>
Hi Wayne
what I have is
$CCPNMR_TOP_DIR/bin/analysis:
#!/bin/tcsh
source $CCPNMR_TOP_DIR/bin/setup
$CCPNMR_TOP_DIR/bin/python -i -O
$CCPNMR_TOP_DIR/ccpnmr1.0/python/ccpnmr/analysis/AnalysisGui.py $1 $2 $3
$4 $5
and then in $CCPNMR_TOP_DIR/bin/setup
setenv ANALYSIS_STARTUP $CCPNMR_TOP_DIR/bin/startup.py
setenv PYTHONPATH $CCPNMR_TOP_DIR/ccpnmr1.0/python:$CCPNMR_TOP_DIR/macros
setenv http_proxy 'http://www-cache.leeds.ac.uk:3128'
so I all I have done is replaced the name ccpnmr by 1.0.12 the fact
that it is an old version number is a red herring all the changes with
the ccpnmr1.0 directory changed to ccpnmr1.0_1.0.12 has worked ....
thus in ~/programs/ccpn/1.0.12 I have
analysis.prof
bin
ccpnmr1.0
ccpnmr1.0_1.0.11
ccpnmr1.0_1.0.12
doc
INSTALL
....
etc
>The update script does not touch the bin directory (at least not
>currently, when we up the version to 1.1.x we will have to do something
>about that). So if the bin/analysis script mentions 1.0.12 (or whatever)
>explicitly then analysis will continue to run that rather than the new
>code. So you will need to edit that script by hand (for now, until we
>figure out if we want to try to do that as well in future, although it's a
>bit of a difficult exercise since we don't know how people have
>hand-edited it).
>
>
no thats not a problem as 1.0.12 is now $CCPNMR_TOP_DIR
>The symbolic linking failing is (probably) a separate issue, I'm not quite
>sure why that is happening. (It might be to do with the non-standard
>directory structure or it might be a problem with the "source" command in
>your Linux. On the latter point, in the next release the "make links"
>command will use "sh" instead of "source", since the former seems to
>be more standard.)
>
>
>
now this is the real problem... I don't think that the path should be
the problem as I have the same slightly deviant directory structure at
work in leeds (the ubuntu box is at home) and
it seems to work well
when you say you are changing source to sh is that wiithon a python
script somewhere?
eitherway when you have made your changes I can check if they work with
my debina/ubuntu box...
regards
gary
n.b. shouldn we move this off lists it maybe a bit boring tfoe the rest
of the world ;-)
>Wayne
>
>On Thu, 15 Feb 2007, Gary S. Thompson wrote:
>
>
>
>>Wayne Boucher wrote:
>>
>>
>>
>>>Hmmm, that looks like it was trying to start up the old (1.0.12) code.
>>>Which is not good! But it looks like that directory structure is
>>>non-standard. So the code should be in ccpnmr/ccpnmr1.0 and when the
>>>upgrade happens that directory should be renamed as
>>>ccpnmr/ccpnmr1.0_1.0.12, and the 1.0.13 version then gets to become the
>>>new ccpnmr/ccpnmr1.0. My guess is that the upgrade script got confused by
>>>the different directory structure and failed in parts, including that
>>>symbolic linking.
>>>
>>>
>>>
>>>
>>>
>>Hi Wayne
>> what I have is indeed named a little different. For historical and
>>organisational reasons I have a main ccpnmr directory and then one
>>directory per version below it (e.g. 1.0.12 1.0.13 etc) However
>>CCPNMR_TOP_DIR does correctly point to 1.0.12... is this something that
>>needs to be fixed? or is it me just being silly ;-)
>>
>>
>>regards
>>gary
>>
>>
>>
>>>More about the symbolic linking. The C code is compiled as shared
>>>libraries, to give *.so files. This happens in the sub-directories of
>>>ccpnmr1.0/c. Because of the way Python works with imports (it's a long
>>>story), in order to get these to be importable into Analyhsis we create
>>>symbolic links to these shared libraries in ccpnmr1.0/python. There are
>>>three sub-directories to worry about, as you discovered, i.e. memops, ccp
>>>and ccpnmr. They each have a directory called c inside which the symbolic
>>>links live. There is a file called linkSharedObjs in each of those
>>>directories which can be run (executed) to create the symbolic links,
>>>should the installation script fail, which it seems to have done for you.
>>>
>>>Hope this makes sense.
>>>
>>>Wayne
>>>
>>>On Wed, 14 Feb 2007, gary thompson wrote:
>>>
>>>
>>>
>>>
>>>
>>>>---------- Forwarded message ----------
>>>>From: gary thompson <[log in to unmask]>
>>>>Date: 14-Feb-2007 11:56
>>>>Subject: memcache crash after upgrade a story
>>>>To: ccpnmr <[log in to unmask]>
>>>>
>>>>
>>>>Dear All
>>>>I have just done an automatic upgrade from 1.0.12 to 1.0.13 and get
>>>>the following on startup:
>>>>
>>>>Error, the Analysis module will not work, something is wrong with the C code.
>>>>Traceback (most recent call last):
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/AnalysisGui.py",
>>>>line 71, in ?
>>>> from ccpnmr.analysis.AnalysisPopup import AnalysisPopup
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/AnalysisPopup.py",
>>>>line 91, in ?
>>>> from ccpnmr.analysis.Analysis import Analysis
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/Analysis.py",
>>>>line 83, in ?
>>>> from memops.c.MemCache import MemCache
>>>>ImportError: No module named MemCache
>>>>
>>>>this is under ubuntu now I have had a look in
>>>>ccpnmr1.0/python/memops/c and their is a file linkSharedObjs and
>>>>running this seems to sort this problem but then I got
>>>>
>>>>Error, the Analysis module will not work, something is wrong with the C code.
>>>>Traceback (most recent call last):
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/AnalysisGui.py",
>>>>line 71, in ?
>>>> from ccpnmr.analysis.AnalysisPopup import AnalysisPopup
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/AnalysisPopup.py",
>>>>line 91, in ?
>>>> from ccpnmr.analysis.Analysis import Analysis
>>>> File "/home/garyt/programs/ccpnmr/1.0.12/ccpnmr1.0/python/ccpnmr/analysis/Analysis.py",
>>>>line 85, in ?
>>>> from ccpnmr.c.ContourFile import ContourFile, StoredContourFile
>>>>ImportError: No module named ContourFile
>>>>
>>>>so i went round and found all the linkSharedObjs scripts and ran them
>>>>and now everything seems to work ;-)
>>>>
>>>>so in ccpnmr1.0/python I ran 'find . -name '*linkS*' -print':
>>>>
>>>>./ccp/c/linkSharedObjs
>>>>./ccpnmr/c/linkSharedObjs
>>>>./memops/c/linkSharedObjs
>>>>
>>>>It seems there is something broken? ;-)
>>>>
>>>>regards
>>>>gary
>>>>
>>>>
>>>>
>>>>
>>>>
>>>.
>>>
>>>
>>>
>>>
>>>
>>--
>>-------------------------------------------------------------------
>>Dr Gary Thompson
>>Astbury Centre for Structural Molecular Biology,
>>University of Leeds, Astbury Building,
>>Leeds, LS2 9JT, West-Yorkshire, UK Tel. +44-113-3433024
>>email: [log in to unmask] Fax +44-113-2331407
>>-------------------------------------------------------------------
>>
>>
>>
>
>.
>
>
>
--
-------------------------------------------------------------------
Dr Gary Thompson
Astbury Centre for Structural Molecular Biology,
University of Leeds, Astbury Building,
Leeds, LS2 9JT, West-Yorkshire, UK Tel. +44-113-3433024
email: [log in to unmask] Fax +44-113-2331407
-------------------------------------------------------------------
|