💬 sipa commented on pull request "net: don't lock cs_main while reading blocks in net processing":
(https://github.com/bitcoin/bitcoin/pull/26326#issuecomment-2089075429)
> we should followup this with sending NotFounds for all the blocks that we are unable (or unwilling) to provide

We can't just change the protocol like that; other software than Bitcoin Core might not expect `notfound` for block messages. But I do believe it would be useful to have *some* way of signalling, and using the information, that a block is unavailable. It would need a short BIP, and a way to negotiate support, but it would mean that over time we may also be able to prefer downloadin
...

💬 IAmAdamRest commented on issue "Manually Banning Peers Results in Crash":
(https://github.com/bitcoin/bitcoin/issues/29916#issuecomment-2089079287)
I decided to try on my ubuntu machine as well and had this same issue but I am not a power user by any means. I will look up this gdb thing to see if I can get a log.
happened right after i banned a peer that kept connecting and disconnecting and connecting and disconnecting like hundreds of times over the course of a few minutes. each time the cpu would get high and i had enough it was an
...

💬 sr-gi commented on pull request "fuzz: txorphan tests fixups":
(https://github.com/bitcoin/bitcoin/pull/29974#discussion_r1586788475)
The reason why I choose to do it this way is because I don't think we generally want to be generating the same transaction often, but it won't be an issue if the case is hit once in a while (really infrequently really).

I don't have strong opinions on what the odds for that should be, so I'm happy to change it if someone does, otherwise I'll leave it as is

⚠️ laanwj opened an issue: "Linux debug information for release is broken"
(https://github.com/bitcoin/bitcoin/issues/30016)
The debug information in the shipped `.dbg` artifacts seems to be unusable. The `.debug_info` section is corrupt:
```sh
$ tar -zxvf bitcoin-27.0-x86_64-linux-gnu-debug.tar.gz
$ objdump -Wi bitcoin-27.0/bin/bitcoin-cli.dbg

Section '.debug_info' has an invalid size: 0.

bitcoin-27.0/bin/bitcoin-cli.dbg: file format elf64-x86-64

Section '.debug_info' has an invalid size: 0.
```
i've checked various Linux architectures (at least x86_64, ARM, RISC-V) and the same issue exists there.
...

💬 laanwj commented on issue "Linux debug information for release is broken":
(https://github.com/bitcoin/bitcoin/issues/30016#issuecomment-2089161664)
Okay, may be a `objdump` bug actually?

`readelf -wi` (which should be equivalent to `objdump -Wi`) parses the same section fine. So does `gdb` it seems?

(FWIW i did try various versions of `objdump` before reporting this)

✅ laanwj closed an issue: "Linux debug information for release is broken"
(https://github.com/bitcoin/bitcoin/issues/30016)

💬 laanwj commented on issue "Linux debug information for release is broken":
(https://github.com/bitcoin/bitcoin/issues/30016#issuecomment-2089162681)
Closing for now.

💬 cbergqvist commented on pull request "net: Favor peers from addrman over fetching seednodes":
(https://github.com/bitcoin/bitcoin/pull/29605#discussion_r1586817396)
Accidental indentation change?

🤔 cbergqvist requested changes to a pull request: "net: Favor peers from addrman over fetching seednodes"
(https://github.com/bitcoin/bitcoin/pull/29605#pullrequestreview-2034448189)
Re-tested 33ddd1b4c1cb1165b5068fbf7a9461e295f6cef1. Bumping of timeout seems blocking.

💬 cbergqvist commented on pull request "net: Favor peers from addrman over fetching seednodes":
(https://github.com/bitcoin/bitcoin/pull/29605#discussion_r1586847044)
Switched to using `Span` for a slightly smaller change suggestion in f9bfc588f2e81a7febe233f591b75e41a52db8b4.

💬 cbergqvist commented on pull request "net: Favor peers from addrman over fetching seednodes":
(https://github.com/bitcoin/bitcoin/pull/29605#discussion_r1586850195)
Had to bump the timeout 20 -> 30 seconds for the test to pass. Example commit with extra logging: 46bac29032c0148ce5288faadbcdc3961d522060

💬 Symphonic3 commented on issue "Change estimate_mode default to "ECONOMICAL" in these RPC calls":
(https://github.com/bitcoin/bitcoin/issues/30009#issuecomment-2089182279)
Willing to create a PR for this if desired!

👍 cbergqvist approved a pull request: "test: Handle functional test disk-full error"
(https://github.com/bitcoin/bitcoin/pull/29335#pullrequestreview-2034535106)
reACK 357ad110548d726021547d85b5b2bfcf3191d7e3. Looks good!

Tested on RAM disk of 1GB, 1430MB (`doas mount -t tmpfs -o size=1430m tmpfs /mnt/tmp/`) and varied the `--jobs` and `--nocleanup`, observing expected warnings + error on disk full.

💬 AngusP commented on pull request "p2p: index TxOrphanage by wtxid, allow entries with same txid":
(https://github.com/bitcoin/bitcoin/pull/30000#discussion_r1586884665)
(Ignore me if I'm understanding, still picking things up 😅) -- coinbase transactions are [required to have all `0` Wtxids](https://github.com/bitcoin/bips/blob/master/bip-0141.mediawiki#commitment-structure) (`0x0000....0000`) which means for an `INV` message for a witness coinbase you have to use the non-witness txid to reference the transaction? - so `AlreadyHaveTx` in that case is explicitly *not* using a Wtxid? (Or are coinbases never `INV`'d?)

Also perhaps not great/confusing to treat W
...

💬 andrewtoth commented on pull request "net: don't lock cs_main while reading blocks in net processing":
(https://github.com/bitcoin/bitcoin/pull/26326#issuecomment-2089285750)
@sr-gi thank you for your review!

> that we should followup this with sending NotFounds for all the blocks that we are unable (or unwilling) to provide

After https://github.com/bitcoin/bitcoin/pull/28120 I don't think it is very likely for this to occur by updated nodes. For older nodes and other software that requests blocks past the prune threshold it would require a protocol change as @sipa mentioned.

📝 hebasto opened a pull request: "refactor, fuzz: Make 64-bit shift explicit"
(https://github.com/bitcoin/bitcoin/pull/30017)
This PR fixes MSVC warning [C4334](https://docs.microsoft.com/en-us/cpp/error-messages/compiler-warnings/compiler-warning-level-3-c4334) in the fuzzing code. Similar to https://github.com/bitcoin/bitcoin/pull/26252.

All `DisableSpecificWarnings` dropped from `fuzz.vcxproj` as all remained are inherited from `common.init.vcxproj`.

Required to simplify warning suppression porting to the CMake-based build system.

💬 furszy commented on pull request "net: don't lock cs_main while reading blocks in net processing":
(https://github.com/bitcoin/bitcoin/pull/26326#issuecomment-2089357427)
Little add about the notfound response; I already have a draft for it. I'm busy with other stuff this week but will open the PR next week :).

📝 ajtowns opened a pull request: "Implement BIP 118 validation (SIGHASH_ANYPREVOUT)"
(https://github.com/bitcoin/bitcoin/pull/30018)
Forward port of #34

✅ ajtowns closed a pull request: "Implement BIP 118 validation (SIGHASH_ANYPREVOUT)"
(https://github.com/bitcoin/bitcoin/pull/30018)

💬 ajtowns commented on pull request "Implement BIP 118 validation (SIGHASH_ANYPREVOUT)":
(https://github.com/bitcoin/bitcoin/pull/30018#issuecomment-2089358941)
Grr :(