gistfile1.txt

12:11 <technoweenie> hey, how can i simulate a riak version conflict to test conflict resolution? 

12:11 <benblack> turn on allow_multi 

12:11 <benblack> and do multiple puts to the same key without passing a clock 

12:11 <technoweenie> oh ok, that simple then 

12:12 <technoweenie> i'd have to fetch it twice so they have the same vclock right? 

12:12 <benblack> fetch it twice? 

12:12 <technoweenie> if i'm going to do 2 puts 

12:13 <technoweenie> nvm i'm looking at this from a higher level. if we're talking just http 
then my question is moot 

12:13 <ericflo> technoweenie: No you explicitly don't want it to have the same vclock 

12:13 <technoweenie> hm ok 

12:14 <ericflo> technoweenie: both PUTs should have no X-Riak-Vclock header 

12:14 <ericflo> technoweenie: which means it won't be able to determine causality 

12:15 <ericflo> technoweenie: which means it'll necessarily create siblings (conflict) 

12:15 <ericflo> correct me if I'm wrong benblack 

12:16 <technoweenie> i guess i figured if you give the same vclock, it'd do the same thing. 

12:16 <ericflo> technoweenie: Oh I see what you're saying, yeah, that should work too 

12:16 <technoweenie> i'm imagining 2 clients getting an object w/ a vclock, and then putting new 
values by passing the same vclock 

12:17 <benblack> that is equivalent 

12:17 <technoweenie> ok cool 

12:17 <technoweenie> thanks guys