pull down to refresh

I'm guilty of not being as much of a voyeur on core as I'd like, but saw this today.
This merged the final step of issue#20160:
Here is a proposed timeline for the move from legacy wallets to descriptor wallets, and at the same time, bdb to sqlite, culminating with the removal of both legacy wallets and bdb. Even though the wallet type and the wallet storage format are orthogonal to each other, as noted in #19077, I would like to tie them together for the sake of simplicity when it comes to how we make new wallets and the migration of one type to another.
The timeline is crazy. Super in awe of the discipline that required. ~30 TODOs gradually done over the course of 5 years!
100 sats \ 2 replies \ @optimism 19h
2 weeks ago a dude came to me at a rum shop asking me if I could help him with his 2014 android wallet backup (to which he also forgot his password.) The same will happen for bdb wallets, so we better make sure we have tools to recover.
reply
100 sats \ 1 reply \ @k00b OP 18h
Sounds like we do!
A bare minimum of legacy wallet code is kept in order to perform wallet migration.
reply
100 sats \ 0 replies \ @optimism 18h
As long as people didn't configure --with-incompatible-bdb on <= 28.0 (back when there was autotools), yes. There will always be people that fall through the gap but luckily most bitcoiners will help you f2f.
reply
0 sats \ 1 reply \ @Jerrian 18h
This is some next-level engineering discipline quietly knocking out 30+ intricate TODOs over five years, culminating in a clean transition from legacy to descriptor wallets and from bdb to sqlite. The patience, foresight, and long-term commitment here is honestly awe-inspiring. It’s a masterclass in how to evolve critical infrastructure without breaking users or trust. Core devs rarely get the spotlight, but this is one of those moments that deserves serious respect.
reply
No one wants to read AI comments. Please stop.
reply