vrypan |--o--|
@vrypan.eth
https://frame.weponder.io/api/polls/30021
2 replies
0 recast
4 reactions
jtgi
@jtgi
vrypan you can’t name it a cup size. other names: - farchive - forever - durable - glacier - bedrock
3 replies
0 recast
6 reactions
jtgi
@jtgi
some unix inspired names - fsync - far - fcp maybe fcp my new favorite
1 reply
0 recast
0 reaction
vrypan |--o--|
@vrypan.eth
fcp is a good one!
1 reply
0 recast
1 reaction
vrypan |--o--|
@vrypan.eth
fcp it is. @jtgi you are its godfather. cc @stephancill `fcp fc://hubble.local:2283/fname localfile.data` I'll tidy it up a bit, but it feels stable. https://github.com/vrypan/fcp
2 replies
1 recast
4 reactions
Kasra Rahjerdi
@jc4p
yo this is sick, is there any reading docs on the grpc connecting and throughput?
1 reply
0 recast
0 reaction
vrypan |--o--|
@vrypan.eth
Probably someone from the MM team can provide more info. However, if you set --page-size=1000, downloading a full, normal profile will only take a few seconds. Uploading requires a hub with write access, and as expected, it's slower. page-size is a tradeoff between local memory usage and number of requests. I.e. if you request 1000 records from the hub, your machine will hold 1000 records in memory per request, marshal them and write them to disk. Given that my whole profile export is 3.4MB, well, this should not be an issue for any desktop computer :-)
1 reply
0 recast
0 reaction
Kasra Rahjerdi
@jc4p
are you communicating using a custom grpc client or the shuttle one in the monorepo?
1 reply
0 recast
0 reaction
vrypan |--o--|
@vrypan.eth
fcp *is* a gRPC client.
1 reply
0 recast
0 reaction