Re: css* tools segfault
Stevan Bajić <[email protected]>
| Newsgroups | gmane.mail.spam.dspam.devel |
|---|---|
| Message-ID | <[email protected]> |
On Thu, 26 Aug 2010 17:18:40 +0200 Eric Bischoff <[email protected]> wrote: > > Le 26 août 2010 à 12:03, Frantisek Hanzlik a écrit : > > > Stevan Bajić wrote: > > ... > >> > >> I will try to make the other css tools to work again but as said before: this will not fix the fundamental issue with 32/64 bit. > >> > >> Looking at the current code I am not sure if the fixes will work with already broken css files? I mean: I don't know if fixing the hash driver will force users to recreate their css file and start from scratch? > >> > >> Would that be an issue or should I just go on and fix things and not care about the compatibility? > >> > >> > > > > From my opinion it isn't too important solve broken css files - dspam > > learning is quick and rebuild broken css ones with pretraining dspam > > with some ham/spam collection is solution too. > > Yes, I agree, it s*cks redoing the training all the time, but if this time files can't get broken anymore it's not a real issue. > From a mathematical viewpoint one can almost never say never. While I believe that this time a lot of things are fixed, I can not say that nothing unexpected can happen. All I can say is that the structure is now the same on 32 bit and on 64 bit. I still have not attempted to fix that endianess of the Hash driver. I have added the functions to handle the endianess but have not wired them into Hash driver. Getting that 32/64 bit thing to work is a big task. In it's core it is easy but you can not fix that problem only for the Hash driver. All DSPAM storage drivers share a certain core and when you touch/change that core then all storage drivers needs to be changed too. I am still in the middle of stabilizing the side effect of my changes. At the beginning I was thinking: Easy. Just fix that 32/64 bit issue and that's it. But doing this is everything else than a low hanging fruit. Every storage driver has his own issues. Some use signed integers while other can not, etc.... -- Kind Regards from Switzerland, Stevan Bajić ------------------------------------------------------------------------------ This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd _______________________________________________ Dspam-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspam-devel