jtriley.eth

@jtriley

28 Following

411 Followers

jtriley.eth pfp

finally getting some good flight time, gonna start uploading footage soon :3

1 reply

0 recast

4 reactions

jtriley.eth pfp

rwandan espresso & müesli gm :3

1 reply

0 recast

1 reaction

jtriley.eth pfp

cooked my esc on my last flight, time for a swap ._.

1 reply

0 recast

5 reactions

jtriley.eth pfp

this is gorgeous 🔥🔥🔥

0 reply

0 recast

1 reaction

jtriley.eth pfp

defs go with a ready-to-fly or bind-and-fly kit for your first. on battery life, it's best to just buy a lot of batteries lol (though make sure it matches the drone's spec!) but before you fly anything you defs want to train first. i like drl simulator. just make sure you use "raw input" once you're familiar with it

1 reply

0 recast

1 reaction

jtriley.eth pfp

all good except 1, receive always runs on zero calldata, even with zero wei ._.

1 reply

0 recast

2 reactions

jtriley.eth pfp

what gets logged if you send: 1. no calldata, no wei? 2. no calldata, one wei? 3. calldata, no wei? 4. calldata, wei?

1 reply

0 recast

7 reactions

jtriley.eth pfp

bc you have to know when to terminate without a jumpi, so you jump to wherever you start and decrement until the last call frame there are ways to get around this without reverse encoding but it incurs cost at runtime and in calldatasize

1 reply

0 recast

1 reaction

jtriley.eth pfp

removal of the loop cleans it up a lot, just 16 calls delimited by jump labels :)

1 reply

0 recast

1 reaction

jtriley.eth pfp

to compensate for dynamically sized payloads, the encoding consists of a header and body where the header contains the call target, value, and payload calldata pointer and size. the body is each payload tightly packed. until next time 🫡 3/3

0 reply

0 recast

5 reactions

jtriley.eth pfp

encoding reversal is conceptually related to turing incompleteness, instead of iterating left to right, calldata is read right to left where the jump target specifies the rightmost element and reads left until there are no more calls to make. 2/3

1 reply

0 recast

3 reactions

jtriley.eth pfp

the rippler branches only in the initial jump, there are no loops, no calldata structure checks, no external call status checks. turing completeness is gas inefficiency. 1/3

1 reply

0 recast

5 reactions

jtriley.eth pfp

announcing the rippler, a fast as fuck multicaller. the rippler uses a custom encoding scheme, packing calls in reverse order and prefixing the calldata with a 2 byte jump target indicating the number of calls to make. 0/3 https://github.com/jtriley-eth/the-rippler

2 replies

1 recast

18 reactions

jtriley.eth pfp

0 reply

0 recast

1 reaction

jtriley.eth pfp

a list built on top of w h a t n o w

1 reply

0 recast

2 reactions

jtriley.eth pfp

it also carries fragmentation issues. a lot of evm's have different precompiles, if l1 eth adds an opcode or precompile, do the others add it? most evms don't even have push0 yet evm needs to ossify quite soon else we need a way for external evms to reach consensus on changes 1/1

0 reply

0 recast

2 reactions

jtriley.eth pfp

precompiles carry the same implications as opcodes, but their cost is massive comparatively bc you have to make an external call to precompiles. it would be better to have an opcode instead, technically, but opcode additions are a massive undertaking. 0/1

1 reply

0 recast

1 reaction

jtriley.eth pfp

b4ufly is a great app for planning an fpv session you need to be in unrestricted airspace with no NOTAMs to fly. special events like airshows may affect where you can fly dji by default will not let you fly in restricted airspace but this is not always the case, better to be safu

0 reply

0 recast

0 reaction

jtriley.eth pfp

don't forget to submit your tcr

0 reply

0 recast

1 reaction

jtriley.eth pfp

1 reply

0 recast

2 reactions