Repository navigation
Fix #15158 mismatchingContainers message refers to data pointers as iterators - #8980
chrchr-github wants to merge 10 commits into
Conversation
|
Test results for commit cd98726 (tools/test-my-pr.py, main compared to this PR): Test: http://ec2-16-170-140-253.eu-north-1.compute.amazonaws.com/pr-8980/ Posted automatically by the cppcheck PR test runner. +N: warnings only with this PR, -N: warnings only with main. The AI review is written by Claude and can be wrong. |
|
See https://trac.cppcheck.net/ticket/15166 for the preexisting FP found by AI. |
dmcppcheck
left a comment
dmcppcheck
left a comment
There was a problem hiding this comment.
AI review
The new wording is still wrong in one case. LifetimeKind::Object is assumed to always mean a container, so taking the addresses of members of two plain structs gives "Pointers of different containers 's1' and 's2'". The objects are not containers, and fixing that kind of misleading message is the point of #15158. Apart from this I found no problems. TestStl passes, and the other variations I tried (vector/string data(), &a[0], unique_ptr<int[]>::get(), arrays, real iterators) get sensible wording.
Testing
I built cppcheck and testrunner in pr/. ./testrunner TestStl passed: 123 tests, 0 failures.
I checked the changed message with /work/pr/cppcheck -q --enable=all --library=std t1.cpp:
#include <algorithm>
#include <vector>
#include <string>
#include <memory>
struct S { int x; int y; };
void f2(std::vector<int>& a, std::vector<int>& b) { std::sort(a.data(), b.data() + 1); }
void f3(S s1, S s2) { std::sort(&s1.x, &s2.y); }
void f5(std::unique_ptr<int[]>& a, std::unique_ptr<int[]>& b) { std::sort(a.get(), b.get()); }
void f6(std::string& a, std::string& b) { std::sort(a.begin(), b.end()); }
void f9() { int a1[10], b1[10]; std::sort(&a1[0], &b1[10]); }Output (abridged):
t1.cpp:6: error: Pointers of different containers 'a' and 'b' are used together. [mismatchingContainers]
t1.cpp:7: error: Pointers of different containers 's1' and 's2' are used together. [mismatchingContainers] <-- S is not a container
t1.cpp:8: error: Pointers of different objects 'a' and 'b' are used together. [mismatchingContainers]
t1.cpp:9: error: Iterators of different containers 'a' and 'b' are used together. [mismatchingContainers]
t1.cpp:10: error: Addresses of different objects 'a1' and 'b1' are used together. [mismatchingContainers]
Only the f3 message is wrong. The rest read correctly.
Automated review of cd98726 by Claude. The pull request was built and tested locally.
| const std::string expr1(tok1 ? tok1->expressionString() : std::string("v1")); | ||
| const std::string expr2(tok2 ? tok2->expressionString() : std::string("v2")); | ||
| const std::string dataStr = (ltk == ValueFlow::Value::LifetimeKind::Iterator) ? "Iterators" : (ltk == ValueFlow::Value::LifetimeKind::Address ? "Addresses" : "Pointers"); | ||
| const std::string objectStr = (ltk == ValueFlow::Value::LifetimeKind::Iterator || ltk == ValueFlow::Value::LifetimeKind::Object) ? "containers '" : "objects '"; |
There was a problem hiding this comment.
AI review
This treats LifetimeKind::Object as if it always meant a container, but Object is also the kind for a pointer to a member of any object. For example:
struct S { int x; int y; };
void f3(S s1, S s2) { std::sort(&s1.x, &s2.y); }This now reports Pointers of different containers 's1' and 's2' are used together., and S is not a container. Choosing between "containers" and "objects" based on whether the token is actually a container (e.g. astIsContainer(tok1)) would fix this.
No description provided.