05:34:41 <gingeropolous> i remember I failed to update xmrchain testnet on time, and it found the right chain
05:35:11 <gingeropolous> and i think it "just worked"
21:44:54 <dartian[m]> I have a monero idea, but I wonder if it would sacrifice privacy. Certainly it would only best be used with "trusted" remote nodes.
21:44:54 <dartian[m]> Basically, the idea is for remote nodes to *only* send the view keys for a block instead of the whole transaction (obviously only for blocks with view keys).
21:44:54 <dartian[m]> A wallet then creates a list of blocks with matching view keys, and requests *only* those blocks from the remote node.
21:44:54 <dartian[m]> The privacy risk is because then the node can tell which blocks you have transactions in. Random blocks could be added to obfuscate, but privacy is still reduced. 
21:46:16 <dartian[m]> The desired benefit would be less data to download from remote nodes
21:46:40 <dartian[m]> * remote nodes to sync a wallet 
21:47:46 <UkoeHB> where is the remote node getting these view keys?
21:48:59 <dartian[m]> Sorry... View tags. 
21:49:19 <dartian[m]> s/keys/tags/, s/keys/tags/
21:51:25 <UkoeHB> I see, so 1) view tag scanning with {tx pubkeys, view tags} sent by remote node, 2) request all output info for blocks that contain at least one view tag match. It would be less data on average...
21:52:38 <UkoeHB> Not sure how much you'd gain from that in practice
22:01:08 <dartian[m]> <UkoeHB> "I see, so 1) view tag scanning..." <- That's exactly right. I'm not sure if sync time is mostly most commonly limited by computational processing of blocks or not, but it would probably help in cases with slower connections at least 
22:09:09 <UkoeHB> Hmm it might be a worthwhile optimization for seraphis remote-scanning services, to speed up legacy scanning for their users. Those services will already be seeing all view tag matches for seraphis enotes. Will take a note of this
23:13:26 <kayabanerve[m]> <jeffro256[m]> "It would be good news if I can..." <- I will be mad if your do, not that my opinion has too much weight here. We've generally macro'd anything which has to be endian-dependent, and those cases come up infrequently. It's generally whenever we're raw accessing memory to void the need to copy.
23:14:04 <kayabanerve[m]> PPC, CELL, and SPARC are BE. There's also BE ARM 
23:14:40 <kayabanerve[m]> Wow that's an incredibly old post I'm sorry for necrobumping it. My client dropped me up three weeks ago not three hours ago :(
23:14:55 <kayabanerve[m]> Please ignore me there