I've been reading some (old) discussions about the potential attack vector of using ERC20 approve (for example: https://github.com/ethereum/EIPs/issues/20#issuecomment-263524729) but I've been noticing that in practice, most DeFi apps (like the 1inch router or the pancake swap router for example) simply have the user do something like this
instead of following the flow that seems to be recommended, e.g.
I'm wondering why this practice is seen as ok.
I personally think this is the better way, and since the approval is for a routing contract, (in my mind) there is essentially zero risk of this attack vector. Not to mention it's way less cumbersome for actual users. But these are old discussions, and I'm wondering what the current thoughts about this issue are from solidity devs.
My main question is this: theoretically, if I had a my own contract holding $100 million WETH (or some other ERC20 token), is it really okay to say something like:
I understand for the average user, approving infinite amounts is probably not a huge deal, but what about a contract that was holding a sizable amount of money. Is it still safe? Is there any threat of this attack vector if the approval is for another contract itself? Any thoughts are greatly appreciated :)