This makes sure we can conclude from `(int)myShort == 0` that `myShort`
is 0 even though we can no longer conclude from `(short)myInt == 0` that
`myInt` is 0. Without this, we lost a good result in the test for
`InfiniteLoopWithUnsatisfiableExitCondition.ql`.
Reported by an LGTM customer here: https://discuss.lgtm.com/t/2-false-positives-in-c-for-comparison-is-always-same/1943.
Even though the comparison is pointless in the preprocessor configuration in effect during extraction, it is not pointless in other preprocessor configurations. Similar to ExprHasNoEffect, we'll now exclude results in functions that contain preprocessor-excluded code. I factored the similar code already used in ExprHasNoEffect in a non-recursive version into Preprocessor.qll, leaving the recursive version in ExprHasNoEffect.ql. I believe the recursive version is too aggressive for PointerlessComparison, which does no interprocedural analysis.
C promotion rules. The following issues are now flagged:
(1) passing a larger type than the receiver can accept
(e.g., long long -> int)
(2) passing a type of different signedness than the
parameter specified.
The IR construction code wasn't handling lambda expressions, so I added `TranslatedLambdaExpression`. It's pretty straightforward: it creates a temporary variable, initializes it with an `Uninitialized` instruction, then initializes the individual captured fields with the initializer list supplied in the AST.
When testing the case of a lambda with no captures, I noticed that we weren't handling initialization of empty structs with an initializer list correctly, so I fixed that along the way.
I was getting confused by the bad indentation for wrapped lines in
TranslatedInitialization.qll, so I fixed that up in a separate commit.
The `Expr.getType` predicate returns a pointer type since that's the
type of the `new`-expression as a whole. To find the class type, we use
`NewExpr.getAllocatedType`.
This commit reduces the number of alerts in a Qt snapshot from 229 to
51, and it removes the two false positives in
https://github.com/Subsurface-divelog/subsurface.
My recent changes to suppress FPs in `ReturnStackAllocatedMemory.ql`
caused us to lose all results where there was a `Conversion` at the
initial address escape. We cannot handle conversions in general, but
this commit restores the good results for the trivial types of
conversion that we can handle.
The main source of slowness in `BrokenCryptoAlgorithm.ql` was that the
regexp on function (macro) names was evaluated once per call
(invocation) instead of once per name. Factoring out separate predicates
for the problematic functions (macros) fixes this.
On https://github.com/ericniebler/range-v3, this change reduces the run
time of the two slowest predicates from
BrokenCryptoAlgorithm::InsecureMacroSpec#class#f .... 35.1s
BrokenCryptoAlgorithm::InsecureFunctionCall#class#f . 12.8s
to
BrokenCryptoAlgorithm::getAnInsecureFunction#f . 1.2s
BrokenCryptoAlgorithm::getAnInsecureMacro#f .... 12ms
Instead of `algorithmBlacklistRegex` having 2 * 5 results, it now has
only one result, which is a single regex that represents the union of
the previous 2 * 5 regexes. This means that `BrokenCryptoAlgorithm.ql`
has much less regex matching to do.
On https://github.com/ericniebler/range-v3, this change reduces the run
time of the two slowest predicates from
BrokenCryptoAlgorithm::InsecureMacroSpec#class#f .... 2m21s
BrokenCryptoAlgorithm::InsecureFunctionCall#class#f . 54.5s
to
BrokenCryptoAlgorithm::InsecureMacroSpec#class#f .... 35.1s
BrokenCryptoAlgorithm::InsecureFunctionCall#class#f . 12.8s
Since we cannot track data flow from a fully-converted expression but
only the unconverted expression, we should check whether the address
initially escapes into the unconverted expression, not the
fully-converted one.
This fixes most of the false positives observed on lgtm.com.
Also added `PhiInstruction::getAnInputOperand()`, and renamed `PhiInstruction::getAnOperandDefinitionInstruction()` to `getAnInput()` for consistency with other `Instruction` classes.
