Simon Marlow wrote: > On 02/02/2010 13:50, Sittampalam, Ganesh wrote: > >> Another problem with conflict marking is that sometimes a simple >> 2-way conflict can leave you with 3 or more conflicting options. By >> "2-way" >> I mean merging two different linear branches without conflicts. It's >> just silly that this happens. > > That does sound strange, any idea why it happens?
I haven't managed to understand the existing conflict marking code, but essentially the problem is that the conflict representation just lists conflicting patches without giving information about where they came from, and repartitioning the conflicts into two independent chains is tricky, especially if you don't even know it's actually possible. Even when I rewrote it to try quite hard to find good solutions, I ran into cases (found with QuickCheck) where you have to spend a lot of CPU time searching, or accept something suboptimal. Ganesh =============================================================================== Please access the attached hyperlink for an important electronic communications disclaimer: http://www.credit-suisse.com/legal/en/disclaimer_email_ib.html =============================================================================== _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
