Closed hexonaut closed 3 years ago
Hmm if you are using dapptools
for checking this gas usage you need to be careful, as you won't catch the set storage to 0 refund. I think the final gas cost should be the same for both type operations.
Best way to test this specific case is deploying both contracts in kovan and actually checking a real tx.
I'd advise doing it before merging as I think it looks better the 0/1 switch and it also avoids the set in the constructor.
Just ran a test: https://kovan.etherscan.io/address/0xb89a0B376c12f9A27f481614cB01Acd6dBc79a2d#code
Existing lock1
: https://kovan.etherscan.io/tx/0xff41332b0600d4ec0cbfe8928d1e77fbd9bcc707494e72cdcbd8860bd5851c86
Proposed new lock2
: https://kovan.etherscan.io/tx/0xf0b40fcf158daa25f234b2a7a3aafa3039a09ea6f23557727f059a841d2d4765
Difference in gas is 22
which is probably just due to function call ordering or something. Looks like you are right @gbalabasquer. I'll close this PR.
PR for https://github.com/hexonaut/dss-flash/issues/9
Making this switch saves 19,200 gas in my personal test.