txgraph: drop move assignment operator #33862

1. 
   sipa commented at 4:25 pm on November 12, 2025: member

   This removes the only place where move-assignment of TxGraph::Ref is used (in tests), and drops supports for it.

   Suggested in #33629 (review)

2. txgraph: drop move assignment operator aef40b93cf
3. 
   DrahtBot commented at 4:25 pm on November 12, 2025: 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/33862.

   Reviews

   See the guideline for information on the review process.

   Type	Reviewers
   ACK	l0rinc, ajtowns

   If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update.

4. 
   sipa commented at 4:25 pm on November 12, 2025: member

   cc @ajtowns
5. glozow added the label Refactoring on Nov 12, 2025
6. glozow added the label Mempool on Nov 12, 2025
7. in src/txgraph.cpp:3454 in aef40b93cf

   3463-    other.m_graph = nullptr;
   3464-    other.m_index = GraphIndex(-1);
   3465-    return *this;
   3466-}
   3467-
   3468 TxGraph::Ref::Ref(Ref&& other) noexcept
   

   ---

   ---

   

   l0rinc commented at 5:33 pm on November 12, 2025:

   This seems used to me, but https://storage.googleapis.com/oss-fuzz-coverage/bitcoin-core/reports/20251112/linux/src/bitcoin-core/src/txgraph.cpp.html#L3469 seems confused by it.
8. l0rinc approved
9. 
   l0rinc commented at 5:35 pm on November 12, 2025: contributor

   code review ACK aef40b93cf057d2a27d61881b0858d491206bcd3

   Simplifies the public interface of TxGraph by deleting the move assignment operator only used in tests. The usage seems correct as it is currently, but left a question about the move-constructor which the fuzzers seem to be confused about.

10. 
    instagibbs commented at 5:40 pm on November 12, 2025: member

    link to discussion, I think #33629 (review)
11. 
    ajtowns commented at 11:32 am on November 13, 2025: contributor

    ACK aef40b93cf057d2a27d61881b0858d491206bcd3 – matches what I was thinking

    Adding some assertions in Ref::Ref(Ref&&) triggers when I run the fuzz binary over some txgraph corpus data I generated previously, so #33862 (review) doesn’t seem like a real problem.

Contributors
sipa DrahtBot l0rinc instagibbs ajtowns

Labels
Refactoring Mempool