doc: Fix wrong code in WITH_LOCK doxygen comment #34355

1. 
   maflcko commented at 4:52 pm on January 20, 2026: member

   The typo is harmless, but a bit confusing every time i read it
2. DrahtBot added the label Docs on Jan 20, 2026
3. 
   DrahtBot commented at 4:52 pm on January 20, 2026: contributor

   The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

   Code Coverage & Benchmarks

   For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/34355.

   Reviews

   See the guideline for information on the review process.

   Type	Reviewers
   ACK	hebasto, l0rinc

   If your review is incorrectly listed, please copy-paste <!–meta-tag:bot-skip–> into the comment that the bot should ignore.

4. in src/sync.h:284 in fa07e52616

   281+//!   WITH_LOCK(cs, int i = 1; return i;);
   282 //!
   283 //! is int, the deduced type of:
   284 //!
   285-//!   WITH_LOCK(cs, return {int j = 1; return (j);});
   286+//!   WITH_LOCK(cs, int j = 1; return (j););
   

   ---

   ---

   

   l0rinc commented at 5:10 pm on January 20, 2026:

   what’s the purpose of (j)?

   ---

   

   maflcko commented at 5:47 pm on January 20, 2026:

   It is used to explain &int, which is different than the int above.

   ---

   

   l0rinc commented at 10:20 am on January 22, 2026:

   Do you mean that

   0//!   WITH_LOCK(cs, int j = 1; return j);
   

   wouldn’t work, only

   0//!   WITH_LOCK(cs, int j = 1; return (j));
   

   ? Why do we need to group it, not sure I understand?

   ---

   

   maflcko commented at 10:53 am on January 22, 2026:

   j is different than (j). One is int, the other is &int after decltype(auto) return type deduction

    0#include <type_traits>
    1
    2#define WITH_CODE(code) [&]() -> decltype(auto) { code; }()
    3
    4int main() {
    5    // Case 1: returns int (by value)
    6    static_assert(
    7        std::is_same_v<decltype(WITH_CODE(int i = 1; return i;)), int>);
    8
    9    // Case 2: returns int& (reference to local variable)
   10    static_assert(
   11        std::is_same_v<decltype(WITH_CODE(int j = 1; return (j);)), int&>);
   12
   13    // Compile-time check for reference-to-local
   14    static_assert(std::is_lvalue_reference_v<decltype(WITH_CODE(int k = 1;
   15                                                                return (k);))>);
   16}
   

   https://godbolt.org/z/ez6d7dz3T

   ---

   

   maflcko commented at 11:41 am on January 22, 2026:

   Also, with C++23, it will be int&&: https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2266r3.html#decltype-auto
5. in src/sync.h:280 in fa07e52616

   276@@ -277,11 +277,11 @@ inline MutexType* MaybeCheckNotHeld(MutexType* m) LOCKS_EXCLUDED(m) LOCK_RETURNE
   277 //! Since the return type deduction follows that of decltype(auto), while the
   278 //! deduced type of:
   279 //!
   280-//!   WITH_LOCK(cs, return {int i = 1; return i;});
   281+//!   WITH_LOCK(cs, int i = 1; return i;);
   

   ---

   ---

   

   l0rinc commented at 5:10 pm on January 20, 2026:

   We usually don’t have a ; inside, e.g.: https://github.com/bitcoin/bitcoin/blob/7f5ebef56a0f2bad7df30252b11da3e33cb3c144/src/net.h#L429

   ---

   

   maflcko commented at 5:46 pm on January 20, 2026:

   Correct. However, the docs here exist to explain decltype(auto).

   ---

   

   maflcko commented at 1:52 pm on January 21, 2026:

   i guess i didn’t understand the comment here. the first semicolon is required, but the second can be removed, so i did that
6. l0rinc changes_requested
7. hebasto approved
8. 
   hebasto commented at 1:35 pm on January 21, 2026: member

   ACK fa07e5261666c808613d9ce9c201cc7a0ca6b80e, the current code snippets have wrong syntax.

   I agree with #34355 (review), as the semicolon is now trailing and should be dropped. Happy to re-ACK once this is addressed.

9. doc: Fix wrong code in WITH_LOCK doxygen comment fa61fadad1
10. maflcko force-pushed on Jan 21, 2026
11. hebasto approved
12. 
    hebasto commented at 2:05 pm on January 21, 2026: member

    re-ACK fa61fadad1c3df0274c3ddd351b8d68a0c4fe644.
13. 
    l0rinc commented at 10:21 am on January 22, 2026: contributor

    ACK fa61fadad1c3df0274c3ddd351b8d68a0c4fe644
14. hebasto merged this on Jan 22, 2026
15. hebasto closed this on Jan 22, 2026

    ---

16. maflcko deleted the branch on Jan 22, 2026

Contributors
maflcko DrahtBot l0rinc hebasto

Labels
Docs