💬 fjahr commented on pull request "checkqueue: make the queue non-optional for CCheckQueueControl and drop legacy locking macro usage":
(https://github.com/bitcoin/bitcoin/pull/32467#discussion_r2085377969)
double-nit: maybe `m_queue`

💬 fjahr commented on pull request "checkqueue: make the queue non-optional for CCheckQueueControl and drop legacy locking macro usage":
(https://github.com/bitcoin/bitcoin/pull/32467#issuecomment-2873868970)
Code review ACK 5aca850c205a20a0f198827f9797fb8053f2b3dd

👍 pinheadmz approved a pull request: "config: allow setting -proxy per network"
(https://github.com/bitcoin/bitcoin/pull/32425#pullrequestreview-2834257645)
ACK 2059eaa692ff614c41dee50f4bc2d2946e5d42af

Built and ran all tests on macos/arm64. Reviewed both commits and left a few nits.
Played with the feature on mainnet, and connected to all networks:

```
ipv4 ipv6 onion cjdns total block manual
in 0 0 0 0 0
out 1 1 2 7 11 0 7
total 1 1 2 7 11
```

Even used tor's ipv4 and unix sockets for separate proxies ;-)


...

💬 pinheadmz commented on pull request "config: allow setting -proxy per network":
(https://github.com/bitcoin/bitcoin/pull/32425#discussion_r2085337724)
4af5b4d4298b4c69cd9131e87fe50e9386def13a

The default port here is a little funny since a user could set `-proxy=127.0.0.1=cjdns` right?

💬 pinheadmz commented on pull request "config: allow setting -proxy per network":
(https://github.com/bitcoin/bitcoin/pull/32425#discussion_r2085310297)
https://github.com/bitcoin/bitcoin/pull/32425/commits/4af5b4d4298b4c69cd9131e87fe50e9386def13a

nit: could move `Proxy onion_proxy;` from L1590 into this group

🤔 ryanofsky reviewed a pull request: "multiprocess: Add bitcoin wrapper executable"
(https://github.com/bitcoin/bitcoin/pull/31375#pullrequestreview-2834221930)
Rebased 81c0b9edfe533afbb2f4dda56142afdedffdb347 -> 27874b9c6e10ac5b6e71bb5e17f44c58691684fe ([`pr/wrap.29`](https://github.com/ryanofsky/bitcoin/commits/pr/wrap.29) -> [`pr/wrap.30`](https://github.com/ryanofsky/bitcoin/commits/pr/wrap.30), [compare](https://github.com/ryanofsky/bitcoin/compare/pr/wrap.29-rebase..pr/wrap.30)) due to conflict with #28710, also made many suggested improvements, particularly dropping windows escaping code and replacing with call to subprocess library

---

re:
...

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085359836)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077503435

> nit: Could use util/string.h:

That's slightly simpler, but doesn't seem enough to want to construct and unneeded vector and pull in an extra dependency here. Worth keeping in mind if this code needs to be extended though.

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085292210)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2055703942

Thanks, fixed both

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085306496)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2073194474

Agree it would be nice to add a friendlier error message that would suggest how to fix the problem. Added this to the list of followups in the PR description that could make the command nicer & easier to use. My goal is to start off with the simplest working implementation and more niceties from there.

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085373623)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077548018

> nit: Could verify we succeeded:

Good point. An assert seems extreme here since there could be some OS variation, but this case wasn't being handled well in the calling function and could do things like try to execute the same path twice. Add better handling now in caller.

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085361665)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077515690

> nit: Could keep style consistent with function body rather than wrapped function.

Thanks, I can never remember where the spaces go.

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085351896)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077299356

Good catch. If the vector was reallocated and string objects were using a small string optimization, the previous code could easily result in passing pointers to already freed memory. Fixed by adding a separate loop to construct new_argv. Also removed the c cast as that was unnecessary.

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085362931)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077517153

Good catch, removed

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085288505)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2048440135

> if you retouch

Thanks, updated

💬 ryanofsky commented on pull request "multiprocess: Add bitcoin wrapper executable":
(https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2085384043)
re: https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077800405

> nit: Child executable names leak through

Good suggestion, added it to the list in the PR description.

🤔 pablomartin4btc reviewed a pull request: "test: Remove legacy wallet RPC overloads"
(https://github.com/bitcoin/bitcoin/pull/32452#pullrequestreview-2834574797)
cr ACK b104d442277090337ce405d92f1398b7cc9bcdb7

Commit reorg proposed by @Sjors made sense.

On first commit, [test: Remove unnecessary importprivkey from wallet_createwallet](https://github.com/bitcoin/bitcoin/pull/32452/commits/94c87bbbd06eb9a57930b9f59315533cfbe8f460), the reasoning is more that `importprivkey` was only compatible with legacy wallets.

On the second commit, [test: Replace importprivkey with wallet_importprivkey](https://github.com/bitcoin/bitcoin/pull/32452/commits/fcc
...

🤔 sipa reviewed a pull request: "cluster mempool: add txgraph diagrams/mining/eviction"
(https://github.com/bitcoin/bitcoin/pull/31444#pullrequestreview-2834591758)
> Unrelated to this PR and more of an FYI, in src/test/fuzz/txgraph.cpp pick_fn I think that:

Sure, that would be correct, but very confusing, I think!

The number of possible choices is `tx_count[0] + sims[0].removed.size() + 1`, and this value is stored in the variable `choices`, representing the number of choices.

A number is then generated in the range between `0` and `choices - 1`, inclusive.

Of course, operationally, the `+ 1` and subsequence `- 1` are superfluous, but then the
...

💬 sipa commented on pull request "cluster mempool: add txgraph diagrams/mining/eviction":
(https://github.com/bitcoin/bitcoin/pull/31444#discussion_r2085515956)
> Perhaps mention that GetCurrentChunk should only be called once (which is unusual for a getter), even though I can't think of an scenario where you would actually want to call this more than once.

Why? It's perfectly allowed to call it more than once (though, kind of pointless).

> However I think the best thing to do is actually not use m_cur_cluster for dual purposes and instead have a dedicated variable that tracks when singletons and last chunks don't need to be added to m_excluded_cl
...

💬 sipa commented on pull request "cluster mempool: add txgraph diagrams/mining/eviction":
(https://github.com/bitcoin/bitcoin/pull/31444#discussion_r2085526718)
Remember that an `Assume()` is **never** necessary. Its only purpose is helping you, the reviewer, gain confidence that a property is likely true.

As for what it isn't here, I didn't think it'd be useful, as `ClearChunkData` is a very low-level internal function. I put the `Assume()` calls around maintenance largely in higher-level functions, but if you feel it would help you to have more, feel free to suggest them.

💬 sipa commented on pull request "cluster mempool: add txgraph diagrams/mining/eviction":
(https://github.com/bitcoin/bitcoin/pull/31444#discussion_r2085528194)
I prefer avoiding early returns if they don't really simplify the code.