Ryo Onodera

I hope this should be fixed by somehow, or maybe I'll create a small pull request...

bump! I'll handle this this week hopefully. I investigated that calling `validate` inside this strategy is the culprit of preventing fallback. Are you still around? :) @cairo140

@cschiewek ping! If I create a pull request for this, could you review and merge? (also, I really hope other already-existing pull requests are handled too...)

@NicolasJacob well, random seed is actually possible to some extent, btw. :) ``` Console $ seq 10 | shuf --random-source=

@taiki-e Hi, I'm glad to see steady progress of this. We're potentially looking forward to using this for Solana's critically-important Accounts DB indexes if possible, which is 10 million entries...

oh, also, i have a small reproducible program and monkey-patched `crossbeam-epoch` and `rayon`. please let me know if it's needed to drive this issue forward.

> Possible solutions: > ... Yet another solution would probably be starting to accept a new type parameter (with default type to be compatible) over `Global`/`Local`. And that type parameter...

@taiki-e hey, I know your status show `busy`. But I waited for a week. Could you take a look at this issue, if possible? Or, could you refer this to...

really appreciate for spending on this issue. :) > If we handle this on our end, it seems increasing PINNINGS_BETWEEN_COLLECT or to make it configurable is okay. thx for confirmation....

> > Yet another solution would probably be starting to accept a new type parameter (with default type to be compatible) over `Global`/`Local`. And that type parameter will be responsible...