Fix false positive in a py/unreachable-statement #1779
Conversation
|
|
xcorail
force-pushed
the
py-unreachable-stmt
branch
from
715c244 to
c1381f3
Compare
Added basic support for the using stmt, checked stmt, unchecked stmt Note that the translations do not use the compiler generated element framework and hence they are just rough approximations. For accuracy, in the future their translation should use it.
The most common reason for the C# autobuilder to fail is because it cannot determine a single unique .sln or .proj file to build, instead reporting multiple sln or proj files at the same shortest depth. This commit changes this to build all such files, rather than reporting an error.
CPP: Add a test of ConditionalDeclExpr.
These files have come out of autoformat since the big commit that autoformatted everything.
In theory this bug could associated CaptureOutNodes with the wrong transitively called callable. However, in practice I could not create a test case that revealed incorrect behaviour. I've included one such test case in the commit. I believe that the cause of this is that OutNode::getACall() is not actually used in the data flow libraries. Instead, DataFlowDispatch::Cached::getAnOutNode is the predicate which is used to associated OutNode's with DataFlowCall's in practice, and that is always used in a context that correctly binds the runtime target of the call.
C++: Autoformat five files
…ent-tuple-size Python: Modernise the `py/mixed-tuple-returns` query.
…check-performance Approved by asger-semmle
Approved by asger-semmle
As the linearity of the disjuncts is different, this enables us to pick better join orders for each disjunct separately.
The user knows that an expression functionally determines its hashCons value, and that an expression functionally determines its number of children, but this is not provable from the definitions, and so not usable by the optimiser. By storing the result of those known-functional calls in a variable, rather than repeating the call, we enable better join orders.
This is helpful for turning real-world cases into test cases.
|
@taus-semmle I applied your suggestions, thanks. |
|
You may want to rebase onto |
A bit of manually applied "magic" ensures that we don't end up considering _all_ variables that are shared by two comparisons, but only those where the first comparison is complex. This is in general a much smaller number.
|
I... seem to have pushed directly to your branch. I apologise for the transgression -- I really did try creating a PR against this, but it seems something went wrong in the process. |
| @@ -22,6 +35,14 @@ predicate typing_import(ImportingStmt is) { | |||
| ) | |||
| } | |||
|
|
|||
| /* This predicate holds when a previous complex comparison occured before | |||
There was a problem hiding this comment.
Just noticed this typo: occurred
C#: Handle unbound types in conversion library
Java/C++/C#: Misc. dataflow fixes.
Java: Improve guards for equal ssa variables.
No worries. Do we still need the |
A bit of manually applied "magic" ensures that we don't end up considering _all_ variables that are shared by two comparisons, but only those where the first comparison is complex. This is in general a much smaller number.
|
Nope. We do not need |
|
Hmmmm ... @taus-semmle the rebase messed up this PR :-( |
This is supposed to fix the FP reported in the forum: https://discuss.lgtm.com/t/python-false-positive-unreachable-statement/2240
The comparison
start_bracket_index >= 0 > end_bracket_indexis not well interpreted, and the subsequent comparisonstart_bracket_index > end_bracket_indexis considered as unreachable. It appears that there is not even aFlowNodecorresponding to this line, as if the CFG had been optimized and pruned of unreachable nodes.So my change is to detect if in the same module, there is a previous chained comparison involving the same variables as the unreachable one. If there is one, then I don't report the statement as unreachable, because the CFG could have been erroneously pruned.