💬 Sjors commented on pull request "ci: build multiprocess on most jobs":
(https://github.com/bitcoin/bitcoin/pull/30975#issuecomment-2650047665)
Same error, but seems to happen at a later step:

```
[15:53:41.937] [ 67%] Linking CXX executable bitcoin-node
[15:53:41.939] cd /ci_container_base/ci/scratch/build-x86_64-pc-linux-gnu/src && /usr/bin/cmake -E cmake_link_script CMakeFiles/bitcoin-node.dir/link.txt --verbose=1
[15:53:41.955] /usr/bin/clang++-20 -ftrivial-auto-var-init=pattern -O2 -g -fsanitize=fuzzer,address,undefined,float-divide-by-zero,integer -fstack-protector-all -fcf-protection=full -fstack-clash-protection -Wl,-z,rel
...

💬 laanwj commented on pull request "Limit retries in GetRNDRRS to avoid infinite loop":
(https://github.com/bitcoin/bitcoin/pull/31826#discussion_r1950400146)
yielding a 1000 timeshares to check if the RNG is working is probably overkill, i'd say if it fails like, 10 times in a row that's already enough indication we shouldn't use it?

💬 laanwj commented on pull request "Limit retries in GetRNDRRS to avoid infinite loop":
(https://github.com/bitcoin/bitcoin/pull/31826#discussion_r1950408783)
As we already use `GetRNDR` in `SeedHardwareFast` which is called way more often than `SeedHardwareSlow` (including *in* `SeedSlow`) i'm not sure if calling it here has any additional value. Let's just skip this loop if rndrrs isn't supported.

💬 laanwj commented on pull request "Limit retries in GetRNDRRS to avoid infinite loop":
(https://github.com/bitcoin/bitcoin/pull/31826#issuecomment-2650101231)
Please update the PR title to reflect the new set of changes.

💬 maflcko commented on issue "Networking tests fail on emulated big-endian systems":
(https://github.com/bitcoin/bitcoin/issues/31812#issuecomment-2650105950)
Ah, I see you are manually running the tests. I thought you were running the s390x ci task.

In the task, the rpc_bind test is excluded:

https://github.com/bitcoin/bitcoin/blob/79f02d56ef78c22caff39713d660cabf54d452ac/ci/test/00_setup_env_s390x.sh#L14

The comment links to https://github.com/bitcoin/bitcoin/issues/17765#issuecomment-602068547

So the exclusion back then may have been due to a permission error on aarch64 or s390x, looking at the discussion. Also, the qemu wrapping back then was
...

💬 Sjors commented on pull request "guix: Notarize MacOS app bundle and codesign all MacOS and Windows binaries":
(https://github.com/bitcoin/bitcoin/pull/31407#issuecomment-2650126441)
Guix hashes for macOS and Windows, before code-sign:

```
bd11279800610b70a1feed4451d9fa3a137ed75f26c794363c5ef82d820c592e guix-build-46e44a35b858/output/arm64-apple-darwin/SHA256SUMS.part
3a911d1a009222c042a84864dccd58462ce3167ea5f407830bf98d370fa356e4 guix-build-46e44a35b858/output/arm64-apple-darwin/bitcoin-46e44a35b858-arm64-apple-darwin-codesigning.tar.gz
2b1c4d096f27ae2a650ce3fe954304c53bb51b7ddac5dd533cabcc862a75f684 guix-build-46e44a35b858/output/arm64-apple-darwin/bitcoin-46e44a
...

💬 hebasto commented on pull request "guix: Notarize MacOS app bundle and codesign all MacOS and Windows binaries":
(https://github.com/bitcoin/bitcoin/pull/31407#discussion_r1950441666)
```suggestion
tar xf bitcoin-${VERSION}-win64-codesigning.tar.gz
```

💬 Sjors commented on pull request "ci: build multiprocess on most jobs":
(https://github.com/bitcoin/bitcoin/pull/30975#issuecomment-2650167452)
The combination did the trick! It's not easy being green.

✅ fanquake closed an issue: "ci: intermittent p2p_tx_download.py timeout (test_preferred_tiebreaker_inv)"
(https://github.com/bitcoin/bitcoin/issues/31833)

🚀 fanquake merged a pull request: "test: add missing sync to p2p_tx_download.py"
(https://github.com/bitcoin/bitcoin/pull/31837)

💬 hebasto commented on pull request "guix: Notarize MacOS app bundle and codesign all MacOS and Windows binaries":
(https://github.com/bitcoin/bitcoin/pull/31407#discussion_r1950471063)
```suggestion
tar xf bitcoin-${VERSION}-${ARCH}-apple-darwin-codesigning.tar.gz
```

💬 Sjors commented on pull request "multiprocess: add bitcoin-mine test program":
(https://github.com/bitcoin/bitcoin/pull/30437#issuecomment-2650217902)
It would be nice to "Build bitcoin-mine executable." to the summary and add a similar guard as in #31834.

💬 maflcko commented on pull request "ci: Skip read-write of default env vars":
(https://github.com/bitcoin/bitcoin/pull/31678#issuecomment-2650253088)
> I want to check if this helps [#30852 (comment)](https://github.com/bitcoin/bitcoin/issues/30852#issuecomment-2558198430) when used with #31545 by not having to keep the base images around anymore.

@0xB10C Are you waiting for me to rebase this once more? I've held back, because there is already one review, but I am happy to rebase.

💬 maflcko commented on issue "Increasing self-hosted runner raw performance":
(https://github.com/bitcoin/bitcoin/issues/30852#issuecomment-2650258443)
As for the raw performance, the runners were upgraded, and the massive scheduling delays of last year (up to ~5hours (?) in peak times, when 10 pull requests or commits were pushed at the same time, resulting in about 100 CI tasks at the same time), it should be a bit smoother this year. I don't have full metrics, but https://cirrus-ci.com/github/bitcoin/bitcoin/master shows "Duration Chart" as a proxy, with the end-to-end maximum duration. Currently, it shows a range of 44 minutes up to 2h:44mi
...

💬 fanquake commented on pull request "guix: remove test-security/symbol-check scripts":
(https://github.com/bitcoin/bitcoin/pull/31818#issuecomment-2650289845)
Guix Build:
```bash
4756d1c55efefe2dd3429d93060555617bb00709772a3c13638aab8c8dd9af2c guix-build-76c090145e9b/output/aarch64-linux-gnu/SHA256SUMS.part
f9fe3c2cb751060366eae0070a1c91897ac49c74bfc53baafca65cb17413ce7e guix-build-76c090145e9b/output/aarch64-linux-gnu/bitcoin-76c090145e9b-aarch64-linux-gnu-debug.tar.gz
6532576a94838e11f8aa98b037f116af32b71eb6f446f67ffc451207b442fd36 guix-build-76c090145e9b/output/aarch64-linux-gnu/bitcoin-76c090145e9b-aarch64-linux-gnu.tar.gz
7ed8370d29955a04
...

💬 laanwj commented on pull request "depends: Make default `host` and `build` comparable":
(https://github.com/bitcoin/bitcoin/pull/30584#issuecomment-2650297413)
Tested ACK b28917be363fb5a82effffeadbe4ba27bb1c70ce
- Successfully tested depends build with `gmake NO_QT=1` on OpenBSD 7.6
- Differences in cmakelists.txt show it's no longer doing a cross-build by default (eg it does `try_run()`):

<details>
<summary>libevent</summary>

```diff
--- cmake-libevent-base.txt Tue Feb 11 10:23:06 2025
+++ cmake-libevent-patched.txt Tue Feb 11 10:33:37 2025
@@ -1,5 +1,5 @@
# This is the CMakeCache file.
-# For build in directory: /home/user/bitcoin/depe
...

💬 vasild commented on pull request "test: add end-to-end tests for CConnman and PeerManager":
(https://github.com/bitcoin/bitcoin/pull/26812#issuecomment-2650310815)
`752fbab26a...53d2c8e0e2`: rebase due to conflicts

This PR is now smaller because part of it was merged via https://github.com/bitcoin/bitcoin/pull/30205, thanks!

💬 laanwj commented on pull request "guix: Notarize MacOS app bundle and codesign all MacOS and Windows binaries":
(https://github.com/bitcoin/bitcoin/pull/31407#discussion_r1950557551)
Yes that seems something to consider when and if we cross that bridge. Currently an `.ots` is already generated for the final SHA256SUMS at release time, not sure to move that to another phase in the process.

🤔 hebasto reviewed a pull request: "guix: Notarize MacOS app bundle and codesign all MacOS and Windows binaries"
(https://github.com/bitcoin/bitcoin/pull/31407#pullrequestreview-2608297032)
Concept ACK 46e44a35b85830a60cf622e039db19ccf1989008.

I have reviewed the code and it looks OK. However, I am not entirely confident in 8400ada306063f1412ef3ace57e255783db879ef due to my lack of familiarity with the signapple tool.

Additionally, I did not review the changes to the signapple tool itself.

🤔 Prabhat1308 reviewed a pull request: "fuzz: coinselection: cover `SetBumpFeeDiscount`"
(https://github.com/bitcoin/bitcoin/pull/31806#pullrequestreview-2608300048)
ACK [0289f03](https://github.com/bitcoin/bitcoin/pull/31806/commits/0289f03790151135afbd5281a45a9f6256f0a235)

`result_{selection-algo-type}->SetBumpFeeDiscount()` sets discount amount before Recalculating the Waste . This will catch any `underflow` scenarios if occuring while subtracting the discount from waste.
Also adding it increased fuzz coverage for earlier untouched part.