💬 TheCharlatan commented on pull request "doc: add LLVM instruction for macOS < 13":
(https://github.com/bitcoin/bitcoin/pull/29934#discussion_r1577616561)
I tested this without specifying a llvm version on macos 12.7.4 monterey. It did resolve the compilation errors. I would prefer this without a version, since it is yet another thing to keep track of in our versioning zoo, and I don't think it is preferable for developers to be instructed to use an older compiler. Also not sure if the guide should cover an OS version that is no longer supported by its package manager. If something eventually breaks for a developer, this could be revisited.

💬 fanquake commented on pull request "ci: Drop no longer needed `-I` flag in "tidy" task":
(https://github.com/bitcoin/bitcoin/pull/29929#issuecomment-2074566263)
Did this change with LLVM version, or IWYU version or changing the base image?

💬 maflcko commented on pull request "refactor: Use our own implementation of urlDecode":
(https://github.com/bitcoin/bitcoin/pull/29904#discussion_r1577622427)
How important is it to preserve invalid encodings? For example the decode of `%1 ` (percent, 1, space) differs between this implementation and the one of libevent.

💬 hebasto commented on pull request "ci: Drop no longer needed `-I` flag in "tidy" task":
(https://github.com/bitcoin/bitcoin/pull/29929#issuecomment-2074570556)
> Did this change with LLVM version, or IWYU version or changing the base image?

My guess is LLVM version, but I didn't check.

🚀 fanquake merged a pull request: "ci: Drop no longer needed `-I` flag in "tidy" task"
(https://github.com/bitcoin/bitcoin/pull/29929)

💬 fanquake commented on pull request "build: add `-Wundef`":
(https://github.com/bitcoin/bitcoin/pull/29876#discussion_r1577642260)
> but probably not worth doing this last-minute change for autotools.

Not sure. Everything not done for Autotools, but done in CMake is just more bebaviour/divergence to maintain, so I'd prefer to do as much as possible prior.

💬 fanquake commented on pull request "[WIP] libevent @ master + use CMake":
(https://github.com/bitcoin/bitcoin/pull/29835#issuecomment-2074617393)
> and if we liked the CMake build parts we'd patch them in and drop the rest.

Yep. From the OP:
> Depending on the outcome, we could possibly pull a (minified) patch set into depends.

💬 hebasto commented on pull request "build: Consider `SOURCE_DATE_EPOCH` in Guix environment only":
(https://github.com/bitcoin/bitcoin/pull/29761#issuecomment-2074627637)
> > However, using it outside Guix scripts for local builds on macOS is not meaningful.
>
> Why? Using Guix should not be a requirement for SOURCE_DATE_EPOCH to be useful.

Not a requirement. Both Guix and SOURCE_DATE_EPOCH are tools of the same domain, which is build reproducibility.

> I would think that given the pinned compilers, and provisioned std lib, the macOS cross build would actually pretty be close to deterministic outside of Guix.

Is anyone interesting / working on that?

...

💬 fanquake commented on pull request "build: Consider `SOURCE_DATE_EPOCH` in Guix environment only":
(https://github.com/bitcoin/bitcoin/pull/29761#issuecomment-2074632025)
> Is anyone interesting / working on that?

I'm saying it's likely already the case, without further changes.

> Under peer reviewing, I guess :)

Ok. Concept NACK I guess. I don't see the motivation.

✅ hebasto closed a pull request: "build: Consider `SOURCE_DATE_EPOCH` in Guix environment only"
(https://github.com/bitcoin/bitcoin/pull/29761)

💬 vasild commented on pull request "doc: suggest only necessary Qt packages for installation on FreeBSD":
(https://github.com/bitcoin/bitcoin/pull/29932#discussion_r1577686920)
> 1. Why `qt5-qmake`?

This PR originated from https://github.com/hebasto/bitcoin/pull/133#discussion_r1551480291 where I was testing bitcoin+cmake on freebsd. `qt5-qmake` is needed because of this error:

```
CMake Error at /usr/local/lib/cmake/Qt5Core/Qt5CoreConfig.cmake:14 (message):
The imported target "Qt5::Core" references the file

"/usr/local/lib/qt5/bin/qmake"

but this file does not exist. Possible reasons include:

* The file was deleted, renamed, or moved to a
...

💬 vasild commented on pull request "doc: suggest only necessary Qt packages for installation on FreeBSD":
(https://github.com/bitcoin/bitcoin/pull/29932#issuecomment-2074666837)
`b1e98c1b36...f4362367d7`: remove `qt5-qmake` and `qt5-network`, see https://github.com/bitcoin/bitcoin/pull/29932#discussion_r1575115643

💬 fanquake commented on issue "guix: SOURCE_DATE_EPOCH is already set in some environments":
(https://github.com/bitcoin/bitcoin/issues/29935#issuecomment-2074668778)
> As we run Guix shell in a container, it seems reasonable to rename SOURCE_DATE_EPOCH in the guix-build and guix-codesign scripts, and pass it to the container using its original name:

Not sure. We already have FORCE_DIRTY_WORKTREE. Seems fine to just make SOURCE_DATE_EPOCH function in the same way, rather than new vairables / more build options / things being less-standard.

💬 kosuodhmwa commented on issue "~/.bitcoin (which is a softlink to a separate vmware virtual drive) dir is now almost 1tb":
(https://github.com/bitcoin/bitcoin/issues/29909#issuecomment-2074669556)
Thank you very much, will look on it ASAP! :-)

💬 maflcko commented on pull request "doc: suggest only necessary Qt packages for installation on FreeBSD":
(https://github.com/bitcoin/bitcoin/pull/29932#issuecomment-2074673320)
Please update to the corresponding version of the OS in line 3 of the file

💬 fanquake commented on issue "RFC: In guix compile the GUI sequentially from everything else?":
(https://github.com/bitcoin/bitcoin/issues/29914#issuecomment-2074679694)
> fully static binaries

Yea. I think I'm just going to start PR'ing related changes, so we can move forward with static builds separately from the GUI, and keep all it's deps / build tools out of the env.

💬 laanwj commented on issue "guix: SOURCE_DATE_EPOCH is already set in some environments":
(https://github.com/bitcoin/bitcoin/issues/29935#issuecomment-2074684256)
That sounds like a good approach to me too.

🚀 fanquake merged a pull request: "netbase: clean up Proxy logging"
(https://github.com/bitcoin/bitcoin/pull/29882)

💬 fanquake commented on pull request "script: throw disabled err for op_ver and its variants":
(https://github.com/bitcoin/bitcoin/pull/28169#issuecomment-2074717810)
I agree. At a minimum, given the PR description is near impossible to follow, it should be re-written. Especailly given the commit only contains `feat: OP_VERIF and OP_VERNOTIF throw disabled error`.

📝 hebasto opened a pull request: "ci: Workaround Microsoft mirror issue for GitHub Action"
(https://github.com/bitcoin/bitcoin/pull/29951)
Upstream:
- issue -- https://github.com/actions/runner-images/issues/9733
- code snippet -- https://github.com/actions/runner-images/issues/9733#issuecomment-2074590278