Closed GoogleCodeExporter closed 9 years ago
Tim, this should be an easy one to knock off.
What's the command for clearing everything in Memops so this bug doesn't occur?
Should I ask Rasmus?
Original comment by jurge...@gmail.com
on 4 Jan 2010 at 12:06
From Rasmus:
There might be something, but it is hard to check when you do not have the test
system to hand.
It is unlikely to be CCPN holding on to old data. You are making a new project,
right? That should effectively reset everything.
It looks to me like you have two MolSystem files in the load directory(s) with
the same MolSystem.code ('1brv'). Either that or you are somehow creating a new
one with a code that is already in use in a file. The
solution would be to get rid of one of the files or move it away. So: look in
the project (or memops/Implementation/projname.xml) to make sure where the
project is picking up files; then check there which files
have MolSystem.code '1brv' and figure out how to get rid of them.
In more general terms, any file in the relevant repository will be read when
apropriate and is conceptually part of the project.
If all else fails you could try to send me the project and code, but it would
be easier for you to do it yourself.
Original comment by jurge...@gmail.com
on 21 Apr 2010 at 2:33
Rasmus is right. The way I was untarring the CCPN .tgz it would insert new
MolSystem files from e.g. 1brv_blabla
into 1brv directory. Fixed in r766 by explicitly removing the CCPN untarred
directory after using it. In general
that is not needed but for testing I like it. It's because the rather unusual
feature of initCcpn to untar whatever is
in the tar and then copy to the right directory.
Thanks Rasmus!
Original comment by jurge...@gmail.com
on 21 Apr 2010 at 3:03
Original issue reported on code.google.com by
jurge...@gmail.com
on 30 Oct 2009 at 2:04