Improve IBD check about bwt HOT 5 CLOSED

electrs is also using initialblockdownload (and is behaving similarly annoying with stale regtest chains), perhaps @romanz has some insight on how that compares to using headers == blocks?

Yes, handling regtest "staleness" indeed an issue...
I think that the solution in 49dc380 should work well.

from bwt.

This is happening because bitcoind considers itself in IBD when no new blocks were received for more than nMaxTipAge. Generating a new block should get it out of IBD mode. (previously also brought up in #35 (comment))

I'm not sure what's the right way to handle this. We could check for headers == blocks instead (I actually have this commented out in the code), but this might have unintended consequences, as you mentioned. IsInitialBlockDownload does a bunch more checks that may be necessary.

electrs is also using initialblockdownload (and is behaving similarly annoying with stale regtest chains), perhaps @romanz has some insight on how that compares to using headers == blocks?

One option that seems safe is keeping the current check for non-regtest chains, and using headers == blocks for regtest only.

from bwt.

One option that seems safe is keeping the current check for non-regtest chains, and using headers == blocks for regtest only.

I went ahead and implemented that in 49dc380. @justinmoon Can you confirm this works for you as intended?

from bwt.

Thanks @romanz!

from bwt.

I haven't tested this, but I wrote the exact same code locally a few days ago and it fixed it. This looks good.

from bwt.