Skip to content

Python: Modernize 4 queries for missing/multiple calls to init/del methods #19932

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

Open
wants to merge 22 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
22 commits
Select commit Hold shift + click to select a range
271f32e
Move missing/multiple calls to init/del queries to folder
joefarebrother Jun 30, 2025
a2fc14a
Update missing call to init
joefarebrother Jun 30, 2025
6f9983a
Add missing call to del
joefarebrother Jun 30, 2025
adcfdf1
Modernize multple calls to init/del
joefarebrother Jul 1, 2025
caddec4
Update alert messages
joefarebrother Jul 1, 2025
71d1179
Fix FPs and typo
joefarebrother Jul 1, 2025
085df26
Move tests and add inline expectation postprocessing
joefarebrother Jul 2, 2025
2faf67d
Update test outputs + fix semantics
joefarebrother Jul 2, 2025
1b4e2fe
Change implenetation of missing calls to use getASuperCallTarget, and…
joefarebrother Jul 3, 2025
16b90a1
Fixes
joefarebrother Jul 3, 2025
b3056fc
Update tests for calls to init + fixes
joefarebrother Jul 3, 2025
73057d3
Add additional test case + update missing del tests
joefarebrother Jul 3, 2025
2e6f35b
Remove case excluding classes with a __new__ method; as it doesn't ma…
joefarebrother Jul 3, 2025
c5b79fa
Update multiple calls queries to include call targets in alert message
joefarebrother Jul 4, 2025
804b9ef
Update tests and add an additional test
joefarebrother Jul 4, 2025
6ca4f32
qhelp: move examples to subfolder
joefarebrother Jul 4, 2025
2e5f470
Update qhelp + alert messages
joefarebrother Jul 4, 2025
d2c68de
Update integration test output
joefarebrother Jul 4, 2025
f1026e4
Add change note
joefarebrother Jul 4, 2025
c47e6e3
Add qldoc
joefarebrother Jul 4, 2025
4b49ac3
Fix changenote formatting
joefarebrother Jul 4, 2025
d163bdf
Fix typos
joefarebrother Jul 4, 2025
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
ql/python/ql/src/Classes/CallsToInitDel/MissingCallToDel.ql
ql/python/ql/src/Classes/CallsToInitDel/MissingCallToInit.ql
ql/python/ql/src/Classes/CallsToInitDel/SuperclassDelCalledMultipleTimes.ql
ql/python/ql/src/Classes/CallsToInitDel/SuperclassInitCalledMultipleTimes.ql
ql/python/ql/src/Classes/ConflictingAttributesInBaseClasses.ql
ql/python/ql/src/Classes/DefineEqualsWhenAddingAttributes.ql
ql/python/ql/src/Classes/EqualsOrHash.ql
ql/python/ql/src/Classes/InconsistentMRO.ql
ql/python/ql/src/Classes/InitCallsSubclass/InitCallsSubclassMethod.ql
ql/python/ql/src/Classes/MissingCallToDel.ql
ql/python/ql/src/Classes/MissingCallToInit.ql
ql/python/ql/src/Classes/MutatingDescriptor.ql
ql/python/ql/src/Classes/SubclassShadowing.ql
ql/python/ql/src/Classes/SuperclassDelCalledMultipleTimes.ql
ql/python/ql/src/Classes/SuperclassInitCalledMultipleTimes.ql
ql/python/ql/src/Classes/WrongNameForArgumentInClassInstantiation.ql
ql/python/ql/src/Classes/WrongNumberArgumentsInClassInstantiation.ql
ql/python/ql/src/Exceptions/CatchingBaseException.ql
Expand Down
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
ql/python/ql/src/Classes/CallsToInitDel/MissingCallToDel.ql
ql/python/ql/src/Classes/CallsToInitDel/MissingCallToInit.ql
ql/python/ql/src/Classes/CallsToInitDel/SuperclassDelCalledMultipleTimes.ql
ql/python/ql/src/Classes/CallsToInitDel/SuperclassInitCalledMultipleTimes.ql
ql/python/ql/src/Classes/ConflictingAttributesInBaseClasses.ql
ql/python/ql/src/Classes/DefineEqualsWhenAddingAttributes.ql
ql/python/ql/src/Classes/EqualsOrHash.ql
ql/python/ql/src/Classes/InconsistentMRO.ql
ql/python/ql/src/Classes/InitCallsSubclass/InitCallsSubclassMethod.ql
ql/python/ql/src/Classes/MissingCallToDel.ql
ql/python/ql/src/Classes/MissingCallToInit.ql
ql/python/ql/src/Classes/MutatingDescriptor.ql
ql/python/ql/src/Classes/SubclassShadowing.ql
ql/python/ql/src/Classes/SuperclassDelCalledMultipleTimes.ql
ql/python/ql/src/Classes/SuperclassInitCalledMultipleTimes.ql
ql/python/ql/src/Classes/WrongNameForArgumentInClassInstantiation.ql
ql/python/ql/src/Classes/WrongNumberArgumentsInClassInstantiation.ql
ql/python/ql/src/Exceptions/CatchingBaseException.ql
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -851,9 +851,14 @@ Class getNextClassInMroKnownStartingClass(Class cls, Class startingClass) {
)
}

private Function findFunctionAccordingToMroKnownStartingClass(
Class cls, Class startingClass, string name
) {
/**
* Gets a potential definition of the function `name` of the class `cls` according to our approximation of
* MRO for the class `startingCls` (see `getNextClassInMroKnownStartingClass` for more information).
*
* Note: this is almost the same as `findFunctionAccordingToMro`, except we know the
* `startingClass`, which can give slightly more precise results.
*/
Function findFunctionAccordingToMroKnownStartingClass(Class cls, Class startingClass, string name) {
result = cls.getAMethod() and
result.getName() = name and
cls = getADirectSuperclass*(startingClass)
Expand All @@ -866,7 +871,7 @@ private Function findFunctionAccordingToMroKnownStartingClass(

/**
* Gets a potential definition of the function `name` according to our approximation of
* MRO for the class `cls` (see `getNextClassInMroKnownStartingClass` for more information).
* MRO for the class `startingCls` (see `getNextClassInMroKnownStartingClass` for more information).
*
* Note: this is almost the same as `findFunctionAccordingToMro`, except we know the
* `startingClass`, which can give slightly more precise results.
Expand Down
213 changes: 213 additions & 0 deletions python/ql/src/Classes/CallsToInitDel/MethodCallOrder.qll
Original file line number Diff line number Diff line change
@@ -0,0 +1,213 @@
/** Definitions for reasoning about multiple or missing calls to superclass methods. */

import python
import semmle.python.ApiGraphs
import semmle.python.dataflow.new.internal.DataFlowDispatch
import codeql.util.Option

/** Holds if `meth` is a method named `name` that transitively calls `calledMulti` of the same name via the calls `call1` and `call2`. */
predicate multipleCallsToSuperclassMethod(
Function meth, Function calledMulti, DataFlow::MethodCallNode call1,
DataFlow::MethodCallNode call2, string name
) {
exists(Class cls |
meth.getName() = name and
meth.getScope() = cls and
call1.getLocation().toString() < call2.getLocation().toString() and
calledMulti = getASuperCallTargetFromCall(cls, meth, call1, name) and
calledMulti = getASuperCallTargetFromCall(cls, meth, call2, name) and
nonTrivial(calledMulti)
)
}

/** Gets a method transitively called by `meth` named `name` with `call` that it overrides, with `mroBase` as the type determining the MRO to search. */
Function getASuperCallTargetFromCall(
Class mroBase, Function meth, DataFlow::MethodCallNode call, string name
) {
exists(Function target | target = getDirectSuperCallTargetFromCall(mroBase, meth, call, name) |
result = target
or
result = getASuperCallTargetFromCall(mroBase, target, _, name)
)
}

/** Gets the method called by `meth` named `name` with `call`, with `mroBase` as the type determining the MRO to search. */
Function getDirectSuperCallTargetFromCall(
Class mroBase, Function meth, DataFlow::MethodCallNode call, string name
) {
meth = call.getScope() and
getADirectSuperclass*(mroBase) = meth.getScope() and
meth.getName() = name and
call.calls(_, name) and
mroBase = getADirectSubclass*(meth.getScope()) and
exists(Class targetCls |
// the differences between 0-arg and 2-arg super is not considered; we assume each super uses the mro of the instance `self`
superCall(call, _) and
targetCls = getNextClassInMroKnownStartingClass(meth.getScope(), mroBase) and
result = findFunctionAccordingToMroKnownStartingClass(targetCls, mroBase, name)
or
// targetCls is the mro base for this lookup.
// note however that if the call we find uses super(), that still uses the mro of the instance `self` will sill be used
// assuming it's 0-arg or is 2-arg with `self` as second arg.
callsMethodOnClassWithSelf(meth, call, targetCls, _) and
result = findFunctionAccordingToMroKnownStartingClass(targetCls, targetCls, name)
)
}

/** Gets a method that is transitively called by a call to `cls.<name>`, with `mroBase` as the type determining the MRO to search. */
Function getASuperCallTargetFromClass(Class mroBase, Class cls, string name) {
exists(Function target |
target = findFunctionAccordingToMroKnownStartingClass(cls, mroBase, name) and
(
result = target
or
result = getASuperCallTargetFromCall(mroBase, target, _, name)
)
)
}

/** Holds if `meth` does something besides calling a superclass method. */
predicate nonTrivial(Function meth) {
exists(Stmt s | s = meth.getAStmt() |
not s instanceof Pass and
not exists(DataFlow::Node call | call.asExpr() = s.(ExprStmt).getValue() |
superCall(call, meth.getName())
or
callsMethodOnClassWithSelf(meth, call, _, meth.getName())
)
) and
exists(meth.getANormalExit()) // doesn't always raise an exception
}

/** Holds if `call` is a call to `super().<name>`. No distinction is made between 0- and 2- arg super calls. */
predicate superCall(DataFlow::MethodCallNode call, string name) {
exists(DataFlow::Node sup |
call.calls(sup, name) and
sup = API::builtin("super").getACall()
)
}

/** Holds if `meth` calls `super().<name>` where `name` is the name of the method. */
predicate callsSuper(Function meth) {
exists(DataFlow::MethodCallNode call |
call.getScope() = meth and
superCall(call, meth.getName())
)
}

/** Holds if `meth` calls `target.<name>(self, ...)` with the call `call`. */
predicate callsMethodOnClassWithSelf(
Function meth, DataFlow::MethodCallNode call, Class target, string name
) {
exists(DataFlow::Node callTarget, DataFlow::ParameterNode self |
call.calls(callTarget, name) and
self.getParameter() = meth.getArg(0) and
self.(DataFlow::LocalSourceNode).flowsTo(call.getArg(0)) and
callTarget = classTracker(target)
)
}

/** Holds if `meth` calls a method named `name` passing its `self` argument as its first parameter, but the class it refers to is unknown. */
predicate callsMethodOnUnknownClassWithSelf(Function meth, string name) {
exists(DataFlow::MethodCallNode call, DataFlow::Node callTarget, DataFlow::ParameterNode self |
call.calls(callTarget, name) and
self.getParameter() = meth.getArg(0) and
self.(DataFlow::LocalSourceNode).flowsTo(call.getArg(0)) and
not exists(Class target | callTarget = classTracker(target))

Check warning

Code scanning / CodeQL

Omittable 'exists' variable Warning

This exists variable can be omitted by using a don't-care expression
in this argument
.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe this is a little more readable, conveying the intent that we're looking for cases where no class can be tracked here.

)
}

/** Holds if `base` does not call a superclass method `shouldCall` named `name` when it appears it should. */
predicate missingCallToSuperclassMethod(Class base, Function shouldCall, string name) {
shouldCall.getName() = name and
shouldCall.getScope() = getADirectSuperclass+(base) and
not shouldCall = getASuperCallTargetFromClass(base, base, name) and
nonTrivial(shouldCall) and
// "Benefit of the doubt" - if somewhere in the chain we call an unknown superclass, assume all the necessary parent methods are called from it
not callsMethodOnUnknownClassWithSelf(getASuperCallTargetFromClass(base, base, name), name)
}

/**
* Holds if `base` does not call a superclass method `shouldCall` named `name` when it appears it should.
* Results are restricted to hold only for the highest `base` class and the lowest `shouldCall` method in the hierarchy for which this applies.
*/
predicate missingCallToSuperclassMethodRestricted(Class base, Function shouldCall, string name) {
missingCallToSuperclassMethod(base, shouldCall, name) and
not exists(Class superBase |
// Alert only on the highest base class that has the issue
superBase = getADirectSuperclass+(base) and
missingCallToSuperclassMethod(superBase, shouldCall, name)
) and
not exists(Function subShouldCall |
// Mention in the alert only the lowest method we're missing the call to
subShouldCall.getScope() = getADirectSubclass+(shouldCall.getScope()) and
missingCallToSuperclassMethod(base, subShouldCall, name)
)
}

/**
* If `base` contains a `super()` call, gets a method in the inheritance hierarchy of `name` in the MRO of `base`
* that does not contain a `super()` call, but would call `shouldCall` if it did, which does not otherwise get called
* during a call to `base.<name>`.
*/
Function getPossibleMissingSuper(Class base, Function shouldCall, string name) {
missingCallToSuperclassMethod(base, shouldCall, name) and
exists(Function baseMethod |
baseMethod.getScope() = base and
baseMethod.getName() = name and
// the base method calls super, so is presumably expecting every method called in the MRO chain to do so
callsSuper(baseMethod) and
// result is something that does get called in the chain
result = getASuperCallTargetFromClass(base, base, name) and
// it doesn't call super
not callsSuper(result) and
// if it did call super, it would resolve to the missing method
shouldCall =
findFunctionAccordingToMroKnownStartingClass(getNextClassInMroKnownStartingClass(result
.getScope(), base), base, name)
)
}

private module FunctionOption = Option<Function>;

/** An optional `Function`. */
class FunctionOption extends FunctionOption::Option {
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
this.asSome()
.getLocation()
.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
or
this.isNone() and
filepath = "" and
startline = 0 and
startcolumn = 0 and
endline = 0 and
endcolumn = 0
}

/** Gets the qualified name of this function. */
string getQualifiedName() {
result = this.asSome().getQualifiedName()
or
this.isNone() and
result = ""
}
}

/** Gets the result of `getPossibleMissingSuper`, or None if none exists. */
bindingset[name]
FunctionOption getPossibleMissingSuperOption(Class base, Function shouldCall, string name) {
result.asSome() = getPossibleMissingSuper(base, shouldCall, name)
or
not exists(getPossibleMissingSuper(base, shouldCall, name)) and
result.isNone()
}
52 changes: 52 additions & 0 deletions python/ql/src/Classes/CallsToInitDel/MissingCallToDel.qhelp
Original file line number Diff line number Diff line change
@@ -0,0 +1,52 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<qhelp>

<overview>
<p>
Python, unlike some other object-oriented languages such as Java, allows the developer complete freedom in
when and how superclass finalizers are called during object finalization.
However, the developer has responsibility for ensuring that objects are properly cleaned up, and that all superclass <code>__del__</code>
methods are called.
</p>
<p>
Classes with a <code>__del__</code> method (a finalizer) typically hold some resource such as a file handle that needs to be cleaned up.
If the <code>__del__</code> method of a superclass is not called during object finalization, it is likely that
that resources may be leaked.
</p>

<p>A call to the <code>__init__</code> method of a superclass during object initialization may be unintentionally skipped:
</p>
<ul>
<li>If a subclass calls the <code>__del__</code> method of the wrong class.</li>
<li>If a call to the <code>__del__</code> method of one its base classes is omitted.</li>
<li>If a call to <code>super().__del__</code> is used, but not all <code>__del__</code> methods in the Method Resolution Order (MRO)
chain themselves call <code>super()</code>. This in particular arises more often in cases of multiple inheritance. </li>
</ul>


</overview>
<recommendation>
<p>Ensure that all superclass <code>__del__</code> methods are properly called.
Either each base class's finalize method should be explicitly called, or <code>super()</code> calls
should be consistently used throughout the inheritance hierarchy.</p>


</recommendation>
<example>
<p>In the following example, explicit calls to <code>__del__</code> are used, but <code>SportsCar</code> erroneously calls
<code>Vehicle.__del__</code>. This is fixed in <code>FixedSportsCar</code> by calling <code>Car.__del__</code>.
</p>

<sample src="examples/MissingCallToDel.py" />

</example>
<references>

<li>Python Reference: <a href="https://docs.python.org/3/reference/datamodel.html#object.__del__">__del__</a>.</li>
<li>Python Standard Library: <a href="https://docs.python.org/3/library/functions.html#super">super</a>.</li>
<li>Python Glossary: <a href="https://docs.python.org/3/glossary.html#term-method-resolution-order">Method resolution order</a>.</li>

</references>
</qhelp>
49 changes: 49 additions & 0 deletions python/ql/src/Classes/CallsToInitDel/MissingCallToDel.ql
Original file line number Diff line number Diff line change
@@ -0,0 +1,49 @@
/**
* @name Missing call to superclass `__del__` during object destruction
* @description An omitted call to a superclass `__del__` method may lead to class instances not being cleaned up properly.
* @kind problem
* @tags quality
* reliability
* correctness
* performance
* @problem.severity error
* @sub-severity low
* @precision high
* @id py/missing-call-to-delete
*/

import python
import MethodCallOrder

Function getDelMethod(Class c) {
result = c.getAMethod() and
result.getName() = "__del__"
}

from Class base, Function shouldCall, FunctionOption possibleIssue, string msg
where
not exists(Function newMethod | newMethod = base.getAMethod() and newMethod.getName() = "__new__") and
exists(FunctionOption possiblyMissingSuper |
missingCallToSuperclassMethodRestricted(base, shouldCall, "__del__") and
possiblyMissingSuper = getPossibleMissingSuperOption(base, shouldCall, "__del__") and
(
not possiblyMissingSuper.isNone() and
possibleIssue = possiblyMissingSuper and
msg =
"This class does not call $@ during finalization. ($@ may be missing a call to super().__del__)"
or
possiblyMissingSuper.isNone() and
(
possibleIssue.asSome() = getDelMethod(base) and
msg =
"This class does not call $@ during finalization. ($@ may be missing a call to a base class __del__)"
or
not exists(getDelMethod(base)) and
possibleIssue.isNone() and
msg =
"This class does not call $@ during finalization. (The class lacks an __del__ method to ensure every base class __del__ is called.)"
)
)
)
select base, msg, shouldCall, shouldCall.getQualifiedName(), possibleIssue,
possibleIssue.getQualifiedName()
Loading
Loading