💬 MarcoFalke commented on pull request "ci: Use qemu-user through container engine":
(https://github.com/bitcoin/bitcoin/pull/28087#issuecomment-1637960373)
If it doesn't work locally, you can check the enabled architectures via `ls /proc/sys/fs/binfmt_misc/`

💬 MarcoFalke commented on issue "Intermittent failures in interface_usdt_mempool.py":
(https://github.com/bitcoin/bitcoin/issues/27380#issuecomment-1637971779)
Ok, I'll disable the broken test due to lack of progress here.

📝 MarcoFalke opened a pull request: "test: Disable known broken USDT test"
(https://github.com/bitcoin/bitcoin/pull/28088)
The failure is known and running into more failures doesn't help anyone. Not disabling the test would be a waste of CPU and developer time.

https://github.com/bitcoin/bitcoin/issues/27380

💬 MarcoFalke commented on issue "Spurious (?) valgrind failure for p2p_compactblocks.py":
(https://github.com/bitcoin/bitcoin/issues/27741#issuecomment-1637988681)
Tracked it down to `CXXFLAGS="-O1 -g -fipa-sra "` so far.

👍 dergoegge approved a pull request: "fuzz: Bump FuzzedDataProvider.h"
(https://github.com/bitcoin/bitcoin/pull/28086#pullrequestreview-1532614758)
utACK fa367422efa3c00f27dab2b58f2080303ed18b91

💬 Crypt-iQ commented on pull request "http: add evhttp_connection_set_closecb to avoid g_requests hang":
(https://github.com/bitcoin/bitcoin/pull/27909#issuecomment-1637992727)
I modified the diff above and printed out the `evhttp_request` memory address as well as the `output_buffer` memory address. I noticed that the`evhttp_request` address is re-used when there's a crash and the `output_buffer` address is not consistent between callbacks.

<details>
<summary>diff</summary>
<br>

```
diff --git a/src/httpserver.cpp b/src/httpserver.cpp
index 849e9b482..9fefad6bd 100644
--- a/src/httpserver.cpp
+++ b/src/httpserver.cpp
@@ -211,6 +211,7 @@ std::string Reques
...

💬 MarcoFalke commented on issue "Spurious (?) valgrind failure for p2p_compactblocks.py":
(https://github.com/bitcoin/bitcoin/issues/27741#issuecomment-1637998300)
This can be fixed, like most other C++ bugs can be fixed: "Add or remove `*` or `&` in random places, until it works."


Diff:

```diff
while (range.first != range.second) {
- auto [node_id, list_it] = range.first->second;
+ auto& [node_id, list_it] = range.first->second;

```

Or, with added `const`:

```diff
diff --git a/src/net_processing.cpp b/src/net_processing.cpp
index 8da2c70..f26522b 100644
--- a/src/net_processing.cpp
+++ b/src/net_processing.cpp

...

💬 vasild commented on pull request "net: support unix domain sockets for -proxy and -onion":
(https://github.com/bitcoin/bitcoin/pull/27375#discussion_r1265260663)
> Transport endpoint is not connected

Is this still a problem or it got resolved?

💬 fanquake commented on pull request "fuzz: Bump FuzzedDataProvider.h":
(https://github.com/bitcoin/bitcoin/pull/28086#issuecomment-1638004857)
Pulls in the latest changes from upstream: https://github.com/llvm/llvm-project/blob/main/compiler-rt/include/fuzzer/FuzzedDataProvider.h.

🚀 fanquake merged a pull request: "fuzz: Bump FuzzedDataProvider.h"
(https://github.com/bitcoin/bitcoin/pull/28086)

🚀 fanquake merged a pull request: "doc: update windows `-fstack-clash-protection` doc"
(https://github.com/bitcoin/bitcoin/pull/28084)

💬 pinheadmz commented on pull request "net: support unix domain sockets for -proxy and -onion":
(https://github.com/bitcoin/bitcoin/pull/27375#discussion_r1265268347)
Resolved!

🚀 fanquake merged a pull request: "fuzz: Flatten all FUZZ_TARGET macros into one"
(https://github.com/bitcoin/bitcoin/pull/28065)

🚀 fanquake merged a pull request: "guix: Remove librt usage from release binaries"
(https://github.com/bitcoin/bitcoin/pull/28069)

💬 MarcoFalke commented on pull request "fuzz: wallet, add target for `Crypter`":
(https://github.com/bitcoin/bitcoin/pull/28074#discussion_r1265311577)
Needs rebase

👍 0xB10C approved a pull request: "test: Disable known broken USDT test"
(https://github.com/bitcoin/bitcoin/pull/28088#pullrequestreview-1532737766)
ACK on disabling the test for now - I'll revisit #27380 when I have more bandwith it.

🤔 MarcoFalke reviewed a pull request: "Make poly1305 support incremental computation + modernize"
(https://github.com/bitcoin/bitcoin/pull/27993#pullrequestreview-1527628367)
Sure, happy to leave comments, but I think it may be better to fix them in a follow-up, unless you really want to here.

💬 MarcoFalke commented on pull request "Make poly1305 support incremental computation + modernize":
(https://github.com/bitcoin/bitcoin/pull/27993#discussion_r1262023316)
```suggestion
BOOST_CHECK_EQUAL(HexStr(total_tag), "64afe2e8d6ad7bbdd287f97c44623d39");
```

style nit, if you retouch.

💬 MarcoFalke commented on pull request "Make poly1305 support incremental computation + modernize":
(https://github.com/bitcoin/bitcoin/pull/27993#discussion_r1265337111)
nit: If you want to avoid the cast here, you could make `ConsumeRandomLengthByteVector` a template on `B` and then directly consume into a `std::vector<std::byte>` and use that.

💬 MarcoFalke commented on pull request "Make poly1305 support incremental computation + modernize":
(https://github.com/bitcoin/bitcoin/pull/27993#discussion_r1265338188)
```suggestion
std::vector<std::byte> msg(i, std::byte{uint8_t(i)});
```

(nit, just personal preference)