Egon Willighagen написа:
On Nov 26, 2007 5:31 PM, Nina Jeliazkova <> wrote:
 At the beginning I would like to say that having SMARTS implementation in
CDK is really a great step forward, congratulations to all contributors.

 I've started trying CDK smarts,  with the intention to replace Joelib
Smarts I am currently using in Toxtree.

 Unfortunately, I've experienced StackOverflow error, and after some
debugging it turns out that if IAtomContainer clone method is executed after
matching a smarts it goes into an infinite cycle.

:) Great...

OK, noted. Please file it as bug report. This needs to be looked at ASAP.

This has probably nothing to do with the SMARTS code itself,
though...but the code merely unlucky to do something no one else has
done before...

Right, it seems it is due to the Ringset, stored as an atom property , which apparently Smartsparser was first to do ... and I was unlucky to need cloning the molecule afterwards.



Dr. Nina Nikolova-Jeliazkova
Institute for Parallel Processing
Bulgarian Academy of Sciences
Acad. G. Bonchev St 25-A
1113 Sofia, Bulgaria
Tel: +359 886 802011
ICQ: 10705013
PGP Public Key
	8E99 8BAD D804 1A43 27B7  7F87 CF04 C7D1 EEAB A669