💬 andrewtoth commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399074127)
This is to not enter if there is only a coinbase tx, since it has no inputs to fetch. If there were 2 txs, and the second has 1000 inputs, we would still want to enter here.

💬 l0rinc commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399077894)
My benchmarks so far indicate the opposite: after 3-4 threads there is no benefit to the parallelization (either on SSD or HDD). I will remeasure your new changes after you give me the 👍

💬 andrewtoth commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399079463)
I don't think we can do that. We need to set these here for other threads to read. These are only read from other threads, never written to. We also only read from other threads after the main thread has released the counting_semaphore, so we know the pointers are synced across the threads.

💬 andrewtoth commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399082258)
Err, we want to throw a runtime error here to test the try/catch in the inputfetcher.

💬 alexanderwiederin commented on pull request "kernel: Introduce initial C header API":
(https://github.com/bitcoin/bitcoin/pull/30595#discussion_r2399082925)
remove `kernel` to align with https://github.com/bitcoin/bitcoin/pull/30595/commits/43f850c1ae9a66d8de599b81d3d78e7178c12251

💬 andrewtoth commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399083511)
I rewrote this part, these are gone now.

💬 l0rinc commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399084818)
But prefiltering would allow sorting, which should untangle the threads.
The threads will access the same files (which are more likely to be different from the files the other threads are requesting), so they may profit from cache locality if the OS supports it - that's why I suggested giving it a try.

💬 vasild commented on pull request "net: replace manual reference counting of CNode with shared_ptr":
(https://github.com/bitcoin/bitcoin/pull/32015#issuecomment-3361550525)
`8218daac0f...d026d5475e`: rebase due to conflicts

This is now a little bit simplified after https://github.com/bitcoin/bitcoin/pull/32326 was merged - no need to touch `FindNode()` anymore.

💬 maflcko commented on pull request "build: Drop support for EOL macOS 13":
(https://github.com/bitcoin/bitcoin/pull/33489#issuecomment-3361550259)
> Looking around for more versions and checking [446e73c](https://github.com/bitcoin/bitcoin/commit/446e73cc0bb7cd628039eaf9b1bcc93db23b924f) it seem to me we could update a few more places:

This is about something else: The sdk and xcode version used for the depends/guix build. So it should be a separate pull request. Generally, bumping the sdk/xcode in depends is a lengthy/tedious process, so it should be done only when needed. My recommendation would be to bump it to 26.x within the next y
...

💬 andrewtoth commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399092048)
> We may have fewer file system locks if the threads are accessing different regions

Ok, but that is not the same as this InputFetcher construction being lock free.

💬 l0rinc commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399094785)
I really dislike that, will try to come up with a lock-free version later (maybe next week)

💬 l0rinc commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399100147)
I think I would prefer a less opinionated version, as long as it's still correct. No need to optimize for the consensus failure speed in my opinion, I would prefer simpler code for a change as risky as this one.

💬 l0rinc commented on pull request "build: Drop support for EOL macOS 13":
(https://github.com/bitcoin/bitcoin/pull/33489#issuecomment-3361587199)
> This is about something else:

What's the reason for keeping a version in the bug issue template that we don't support?

> I don't see how this is related to this pull?

If we recommend a newer `AppleClang` version, it may not work with fuzzing, that's all.

💬 l0rinc commented on pull request "validation: fetch block inputs on parallel threads >10% faster IBD":
(https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399116295)
No, that's orthogonal, it's another area where we could possibly reduce contention.

💬 maflcko commented on pull request "build: Drop support for EOL macOS 13":
(https://github.com/bitcoin/bitcoin/pull/33489#issuecomment-3361601497)
> > This is about something else:
>
> What's the reason for keeping a version in the bug issue template that we don't support?

It is just a placeholder, so shouldn't matter. I can bump it on the next push, if there is one, or it can be done as a follow-up.

💬 fanquake commented on issue "Release Schedule for 30.0":
(https://github.com/bitcoin/bitcoin/issues/32275#issuecomment-3361619599)
Moved the release date forward a week, as we still need to do at least an rc3.

💬 maflcko commented on pull request "ci: Turn CentOS config into Alpine musl config":
(https://github.com/bitcoin/bitcoin/pull/33480#issuecomment-3361650414)
Another (tiny) benefit of this pull request is that the task will now finally be able to run on my riscv64 box (modulo the two python wheels)

💬 l0rinc commented on pull request "build: Drop support for EOL macOS 13":
(https://github.com/bitcoin/bitcoin/pull/33489#issuecomment-3361731762)
code review ACK 1aaaaa078bb2efed126e3f41ecf7c81ccf005818

💬 l0rinc commented on pull request "coins: use number of dirty cache entries in flush warnings/logs":
(https://github.com/bitcoin/bitcoin/pull/33512#discussion_r2399206305)
You also answered something similar in https://github.com/bitcoin/bitcoin/pull/31132#discussion_r2399049434

💬 fanquake commented on pull request "init: Fix Ctrl-C shutdown hangs during wait calls":
(https://github.com/bitcoin/bitcoin/pull/33511#issuecomment-3361852609)
> This should be backported in my opinion.

Not going to block `30.0` if it's not ready, but marked for backport.