Hi Giacomo,
I'm sure that you know this, but once you have changed the environment.txt file, you MUST run (in the directory with environment.txt)
$ make clean
$ make
$ make links
otherwise your changes won't have any effect.
So are you now getting the same errors as before, or new (different) ones?
Now that we have the Tcl/Tk parts fixed we need to make sure that the python is also ok.
How are you launching Analysis? If from the bin/analysis file, could you have a look at the bin/analysis file and double check which Python executable is being used to run Analysis ?
It should be the one that you compiled the code against, ie NOT the system python.
One final thing to check is the C shared objects themselves. First change to the directory where they are:
$ cd <ccpn path>/c/memops/global
and then try
ldd TkHandler.so
This will display something like
djo35@mammoth:~/Code/ccpnmr/ccpn/c/memops/global$ ldd TkHandler.so
To make sure 100% that you are using the libraries that you expect to be using.
For something completely different;
Using the pre-compiled versions will save you this headache. If compiling from source you really need to keep an eye on your libraries. i must also re-iterate that v1 CCPN Analysis will not be supported indefinitely, and switching to v2 will allow you to use all the new functionality, use a more stable analysis and still get continued CCPN support.
Hope that helps,
Dan
On 30 Nov 2009, at 09:36, Giacomo Quilici wrote:
> Hi,
> The suggested method has been used for checking versions of the software installed.
> The "good" news are in the fact that the bug is, indeed, in the usage of version 8.5 ot Tk vs the 2.4
> version of Python. Hence I recompiled analysis after a check on the environment.txt file I edited: here
> the bad news started, because the version of Tk loaded in Python 2.4 still resulted to be the 8.5 (that,
> BTW, is the one installed in the system, that is Slackware Linux) instead of the required 8.4 by means of
> the following lines :
>
> # Tcl
> TCL_DIR = /usr/local/common/src/ccpnmr_old_GL/tcl8.4
> TCL_LIB = -ltcl8.4
> TCL_INCLUDE_FLAGS = -I$(TCL_DIR)/include
> TCL_LIB_FLAGS = -L$(TCL_DIR)/lib
>
> # Tk
> TK_DIR = /usr/local/common/src/ccpnmr_old_GL/tk8.4
> TK_LIB = -ltk8.4
> TK_INCLUDE_FLAGS = -I$(TK_DIR)/include
> TK_LIB_FLAGS = -L$(TK_DIR)/lib
>
> I tried also to use the easy installation procedure (python2.4/bin/python installCode.py) selecting where
> to get libs etc, but still no way: the program seemed to work properly, but no peaks appeared in the
> windows or could be displayed.
> I have checked into Python which environments are imported this time and TclTk installed locally results
> to be loaded , that is 8.4 (not the one in standard path for the machine).
> Thanks for your attention
>
> Giacomo
>
>
>
>
>
> From: [log in to unmask]
> To: [log in to unmask]
> Subject: FW: errors in opening ccpn version 1
> Date: Thu, 26 Nov 2009 15:15:16 +0100
>
>
>
>
>
>
>
>
> > Date: Thu, 26 Nov 2009 11:44:47 +0000
> > From: [log in to unmask]
> > Subject: Re: errors in opening ccpn version 1
> > To: [log in to unmask]
> >
> > Hello Giacomo,
> >
> > This is the root of your error:
> >
> > > ( 3 times ) TkHandler.error: could not get main window, possibly different Python Tcl/Tk and Analysis
> > > Tcl/Tk Exception in Tkinter callback
> >
> > And occurs because the version of Tcl/Tk used when compiling python is different from the one used when you compiled analysis.
> >
> > First, make sure that the python you are using to run analysis is definitely the one that you compiled against (sometimes system python can sneak in here).
> >
> > So: try
> >
> > $ /<path to python>/bin/python -i /<ccpnmr root>/AnalysisGui.py
> >
> > and see if that produces the same error.
> >
> > Next, try running that python 2.4 by itself:
> >
> > $ /<path to python>/bin/python
> >
> > and importing Tkinter - the Tcl/Tk interface
> >
> > >>> import Tkinter
> > >>> Tkinter._test()
> >
> > This should tell you which Tcl/Tk your python 2.4 was compiled against. Hopefully it matches Tk 8.4
> >
> > Once you've tried this, let us know what happened, although I must add that we can't continue to support Analysis v1 forever and further support will be limited to v2 at some point. The project upgrade mechanisms should work, and if they don't we would be most interested to know why (and then fix). If you could send a project (off list) that fails for you we would be glad to fix it.
> >
> > Hope that helps,
> >
> > Dan
> >
> >
> > On 26 Nov 2009, at 11:27, Giacomo Quilici wrote:
> >
> > > Good morning, we have some problem in opening old projects of ccpnmr version1. We know that it's possible to convert them into version 2 and in some cases we did it succesfully, but they are a lot and some of them are huge, so it could be better use them with the old version 1.
> > > We had to remove the software and now we find some trouble in the installation:
> > >
> > >
> > >
> > > The CCPNMR v.1 has been installed from scratch (many prjects are still in the old format) compiling
> > >
> > > Python 2.4
> > > Tcl 8.4
> > > Tk 8.4
> > > Mesa 6.0
> > >
> > > and updated.
> > > the program starts correctly, converts paths (or minor modification are required) and the project is
> > > opened. the spectra, peaks etc are still there (e.g., peaks can be browsed) but there is no way to see
> > > anything on the screen.
> > > the errors that are (many times) prompted in the shell are the following:
> > >
> > > ( 3 times ) TkHandler.error: could not get main window, possibly different Python Tcl/Tk and Analysis
> > > Tcl/Tk Exception in Tkinter callback
> > >
> > >
> > > ( 22 times ) AttributeError: WindowCanvas instance has no attribute 'canvas_width'
> > >
> > > Exception in Tkinter callback
> > >
> > > Is the thing correctable by means of manual editing of files ? Or what do we require for fixing the
> > > problem ?
> > > anyway we will try to convert the project from version 1 to version 2 via the web interface.
> > >
> > > Thanks for your attention
> > >
> > > Giacomo Quilici
> > >
> > > Dulbecco Telethon Institute c/o S. Raffaele Scientific Institute
> > > Center of Genomics, BioInformatics and BioStatistics
> > > Biomolecular NMR Laboratory 1B4
> > > Via Olgettina 58
> > > 20132 Milano
> > > Italy
> > >
> > >
> > >
> > >
> > >
> > > Crea e condividi i tuoi filmati con Movie Maker
> >
> > Daniel O'Donovan
> > [log in to unmask]
> >
> >
> >
>
> Crea e condividi i tuoi filmati con Movie Maker
> Crea e condividi i tuoi filmati con Movie Maker
Daniel O'Donovan
[log in to unmask]
|