-
Notifications
You must be signed in to change notification settings - Fork 14.3k
[analyzer] Improve some comments in ArrayBoundCheckerV2 (NFC) #83545
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
...which was introduced by one of my recent commits.
To clarify them and include some thoughts about the difficulties in resolving a certain FIXME.
@llvm/pr-subscribers-clang-static-analyzer-1 Author: None (NagyDonat) ChangesThis comment-only change fixes a typo, clarifies some comments and includes some thoughts about the difficulties in resolving a certain FIXME. Full diff: https://github.com/llvm/llvm-project/pull/83545.diff 1 Files Affected:
diff --git a/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp b/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
index fdcc46e58580b4..29eb932584027d 100644
--- a/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
+++ b/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
@@ -301,21 +301,27 @@ compareValueToThreshold(ProgramStateRef State, NonLoc Value, NonLoc Threshold,
// calling `evalBinOpNN`:
if (isNegative(SVB, State, Value) && isUnsigned(SVB, Threshold)) {
if (CheckEquality) {
- // negative_value == unsigned_value is always false
+ // negative_value == unsigned_threshold is always false
return {nullptr, State};
}
- // negative_value < unsigned_value is always false
+ // negative_value < unsigned_threshold is always true
return {State, nullptr};
}
if (isUnsigned(SVB, Value) && isNegative(SVB, State, Threshold)) {
- // unsigned_value == negative_value and unsigned_value < negative_value are
- // both always false
+ // unsigned_value == negative_threshold and
+ // unsigned_value < negative_threshold are both always false
return {nullptr, State};
}
- // FIXME: these special cases are sufficient for handling real-world
+ // FIXME: These special cases are sufficient for handling real-world
// comparisons, but in theory there could be contrived situations where
// automatic conversion of a symbolic value (which can be negative and can be
// positive) leads to incorrect results.
+ // NOTE: We NEED to use the `evalBinOpNN` call in the "common" case, because
+ // we want to ensure that assumptions coming from this precondition and
+ // assumptions coming from regular C/C++ operator calls are represented by
+ // constraints on the same symbolic expression. A solution that would
+ // evaluate these "mathematical" compariosns through a separate pathway would
+ // be a step backwards in this sense.
const BinaryOperatorKind OpKind = CheckEquality ? BO_EQ : BO_LT;
auto BelowThreshold =
|
@llvm/pr-subscribers-clang Author: None (NagyDonat) ChangesThis comment-only change fixes a typo, clarifies some comments and includes some thoughts about the difficulties in resolving a certain FIXME. Full diff: https://github.com/llvm/llvm-project/pull/83545.diff 1 Files Affected:
diff --git a/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp b/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
index fdcc46e58580b4..29eb932584027d 100644
--- a/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
+++ b/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
@@ -301,21 +301,27 @@ compareValueToThreshold(ProgramStateRef State, NonLoc Value, NonLoc Threshold,
// calling `evalBinOpNN`:
if (isNegative(SVB, State, Value) && isUnsigned(SVB, Threshold)) {
if (CheckEquality) {
- // negative_value == unsigned_value is always false
+ // negative_value == unsigned_threshold is always false
return {nullptr, State};
}
- // negative_value < unsigned_value is always false
+ // negative_value < unsigned_threshold is always true
return {State, nullptr};
}
if (isUnsigned(SVB, Value) && isNegative(SVB, State, Threshold)) {
- // unsigned_value == negative_value and unsigned_value < negative_value are
- // both always false
+ // unsigned_value == negative_threshold and
+ // unsigned_value < negative_threshold are both always false
return {nullptr, State};
}
- // FIXME: these special cases are sufficient for handling real-world
+ // FIXME: These special cases are sufficient for handling real-world
// comparisons, but in theory there could be contrived situations where
// automatic conversion of a symbolic value (which can be negative and can be
// positive) leads to incorrect results.
+ // NOTE: We NEED to use the `evalBinOpNN` call in the "common" case, because
+ // we want to ensure that assumptions coming from this precondition and
+ // assumptions coming from regular C/C++ operator calls are represented by
+ // constraints on the same symbolic expression. A solution that would
+ // evaluate these "mathematical" compariosns through a separate pathway would
+ // be a step backwards in this sense.
const BinaryOperatorKind OpKind = CheckEquality ? BO_EQ : BO_LT;
auto BelowThreshold =
|
// unsigned_value == negative_value and unsigned_value < negative_value are | ||
// both always false | ||
// unsigned_value == negative_threshold and | ||
// unsigned_value < negative_threshold are both always false |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just a minor nit: Wouldn't it be better to write <=
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm mentioning them separately because these are separate cases in this function. (It may check either "value < threshold" or "value == threshold" depending on the last argument.)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh. I understand.
This comment-only change fixes a typo, clarifies some comments and includes some thoughts about the difficulties in resolving a certain FIXME.