a few days ago there was a question on if lcr module is able to serialize parallel forking of branches with the same q value. i replied that there is no such capability in lcr module.
now that i have thought about the matter a bit, in my opinion the best solution is to write a separate serialize function that rewrites the q values rather than trying adding yet another feature to already too complex lcr module.
-- juha
Hi Juha,
Well,.... this links to a discussion from more than an year ago - moving the parallel2serial converter and serial forking support from lcr to core. Already there are 2 core functions that are doing the same (more or less) and I will really like to see this step completed and remove the serialization support from lcr module - as you said, the module grows, and it should keep focused on the LCR functionality.
Regarding the q issue - I agree with your approach - instead of hacking the serialization engine, better play with q values before serialization.
Is any common will on the first item (serialization to be moved out from LCR)? Because, if so, we should work out a plan of action.
Regards, Bogdan
Juha Heinanen wrote:
a few days ago there was a question on if lcr module is able to serialize parallel forking of branches with the same q value. i replied that there is no such capability in lcr module.
now that i have thought about the matter a bit, in my opinion the best solution is to write a separate serialize function that rewrites the q values rather than trying adding yet another feature to already too complex lcr module.
-- juha
Users mailing list Users@lists.openser.org http://lists.openser.org/cgi-bin/mailman/listinfo/users
Bogdan-Andrei Iancu writes:
Is any common will on the first item (serialization to be moved out from LCR)? Because, if so, we should work out a plan of action.
i don't anything against it as long as core version would implement the same serialization capability that lcr.
i made a huge hack in lcr module in order to get per branch stuff working, which does not exist in core version. i don't myself remember the details, but i think i proposed adding some more stuff to transaction state in order to remove need for the hacks.
in my opinion, it does not make sense to migrate anything to core, before the branch stuff is fixed.
-- juha
-
Bogdan-Andrei Iancu -
jh＠tutpro.com