Paths

Table of Contentst

Merge #14885: rpc: Assert named arguments are unique in RPCHelpMan
ClosedPublic
Actions

Authored by nakihito on Mar 27 2020, 00:22.

Details

Reviewers

Fabien
deadalnix

Group Reviewers

Restricted Owners Package	(Owns No Changed Paths)
Restricted Project

Commits

rSTAGING43d44b2e899b: Merge #14885: rpc: Assert named arguments are unique in RPCHelpMan
rABC43d44b2e899b: Merge #14885: rpc: Assert named arguments are unique in RPCHelpMan

Summary

e09a5875ca rpc: Assert named arguments are unique in RPCHelpMan (João Barbosa)

Pull request description:

Prevents an obvious mistake.

Tree-SHA512: 32c24a1934b17ab6f0d5cd31bdf0388e93ee5156ccc1b4f78eb9fd7f1d4b27a4b978b594ff11812bc9f20987c9fc36bf4497ddaedf18cf6bcbea19c050571334

Backport of Core PR14885

Depends on D5746

Test Plan

ninja
./bitcoind
mkdir internalerror
./bitcoin-cli help > RPCs
./getrpchelps.sh RPCs internalerror
grep -rnI Internal\ error internalerror

The grep command's output should be empty.

Script:

getrpchelps.sh231 BDownload

Change the name of an argument for an RPC to be the same as another
i.e. waitforblock in rpc/blockchain.cpp from

{"blockhash", RPCArg::Type::STR_HEX, /* opt */ false,

 {"timeout", RPCArg::Type::STR_HEX, /* opt */ false,
ninja
./bitcoind
./bitcoin-cli help waitforblock

Confirm bitcoind does not crash and that bitcoin-cli outputs the following error:

Internal bug detected: 'named_args.insert(arg.m_name).second'
You may report this issue here: https://github.com/Bitcoin-ABC/bitcoin-abc/issues

Diff Detail

Repository

rABC Bitcoin ABC

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

nakihito created this revision.Mar 27 2020, 00:22

Owners added a reviewer: Restricted Owners Package.Mar 27 2020, 00:22

Herald added a reviewer: Restricted Project. · View Herald TranscriptMar 27 2020, 00:22

[Bot Message]
One or more PR numbers were detected in the summary.
Links to those PRs have been inserted into the summary for reference.

Build Bitcoin-ABC / Diffs / Diff Testing (build-diff) started.

Build Bitcoin-ABC / Diffs / Diff Testing (build-diff) passed.

Harbormaster completed remote builds in B10005: Diff 17236.Mar 27 2020, 00:32

Fabien requested changes to this revision.Mar 27 2020, 13:25

Fabien added a subscriber: Fabien.

Fabien added inline comments.

src/rpc/util.cpp

124 ↗

(On Diff #17236)

From the developer notes:

- Assertions should not have side-effects

  - *Rationale*: Even though the source code is set to refuse to compile
    with assertions disabled, having side-effects in assertions is unexpected and
    makes the code harder to understand

If assert is compiled out the insertion is lost.

This revision now requires changes to proceed.Mar 27 2020, 13:25

Going to have to go through some dependencies to remove the assert.

Rebased of D5746 to remove assert.

Build Bitcoin-ABC / Diffs / Diff Testing (build-diff) started.

nakihito planned changes to this revision.Apr 16 2020, 19:46

nakihito edited the test plan for this revision. (Show Details)Apr 16 2020, 19:55

nakihito edited the test plan for this revision. (Show Details)

nakihito edited the summary of this revision. (Show Details)

nakihito added a parent revision: D5746: Merge #17192: util: Add CHECK_NONFATAL and use it in src/rpc.

nakihito requested review of this revision.Apr 16 2020, 19:57

Build Bitcoin-ABC / Diffs / Diff Testing (build-diff) passed.

Harbormaster completed remote builds in B10254: Diff 18873.Apr 16 2020, 19:58

deadalnix requested changes to this revision.Apr 17 2020, 00:33

deadalnix added a subscriber: deadalnix.

deadalnix added inline comments.

src/rpc/util.cpp
292 ↗	(On Diff #18873)	Why not an assert as in the original PR?

This revision now requires changes to proceed.Apr 17 2020, 00:33

nakihito requested review of this revision.Apr 17 2020, 18:09

nakihito added inline comments.

src/rpc/util.cpp
292 ↗	(On Diff #18873)	https://reviews.bitcoinabc.org/D5586?id=17236#inline-34624 From the developer notes: Assertions should not have side-effects Rationale: Even though the source code is set to refuse to compile with assertions disabled, having side-effects in assertions is unexpected and makes the code harder to understand If assert is compiled out the insertion is lost.