The new-dataflow ImportResolution module only used
semmle.python.essa.SsaDefinitions for the 5-line helper predicate
SsaSource::init_module_submodule_defn. Inline it locally and drop the
dependency on legacy SsaDefinitions. This is the only remaining direct
import of semmle.python.essa.* in the new dataflow stack, so dropping
it makes the layering cleaner.
Semantic noop on the current SSA: SsaSourceVariable.getName() and
GlobalVariable.getId() both project the same DB column
(variable(_,_,result)), and the old call's 'init.getEntryNode() = f'
join was just constraining init = package via Scope.getEntryNode()'s
functional uniqueness. RA dump of accesses.ql confirms only the
expected predicate-rename shuffle; all 70 dataflow + ApiGraphs library
tests pass.
This factors out commit 8cab5a20f2 from the larger shared-CFG
migration #21925.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The internal predicates that identify `@staticmethod`, `@classmethod` and
`@property` decorators previously required the decorator's `NameNode` to
satisfy `isGlobal()` (i.e. no SSA def reaches the decorator's name use).
That filter was correct but unnecessarily indirect: these three names
are builtins, and even when a class body redefines one, the class body
has not started executing at the decorator position, so Python uses the
builtin.
Match the decorator's AST `Name` directly instead, dropping the CFG/SSA
detour. The slight semantic change — `isGlobal()` would have rejected
module-level shadowing of these builtins — is negligible in practice
and explicitly documented in the change note.
`hasContextmanagerDecorator` and `hasOverloadDecorator` keep the
`NameNode.isGlobal()` check because their target names (`contextmanager`,
`overload`) are imported, not builtin, and local shadowing is a real
concern.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Adds a new `isLazy` predicate to the relevant classes, and adds the
relevant dbscheme (and up/downgrade) changes. On upgrades we do nothing,
and on downgrades we remove the `is_lazy` bits.
This takes care of most of the false negatives from the preceding
commit.
Additionally, we add models for some known wrappers of `socket.socket`
from the `gevent` and `eventlet` packages.
For module-level metaclass declarations, we now also check that the
right hand side in a `__metaclass__ = type` assignment is in fact the
built-in `type`.
These could arguably be moved to `Class` itself, but for now I'm
choosing to limit the changes to the `DuckTyping` module (until we
decide on a proper API).
This module (which for convenience currently resides inside
`DataFlowDispatch`, but this may change later) contains convenience
predicates for bridging the gap between the data-flow layer and the old
points-to analysis.
The fix may look a bit obscure, so here's what's going on.
When we see `from . import helper`, we create an `ImportExpr` with level
equal to 1 (corresponding to the number of dots). To resolve such
imports, we compute the name of the enclosing package, as part of
`ImportExpr.qualifiedTopName()`. For this form of import expression, it
is equivalent to `this.getEnclosingModule().getPackageName()`. But
`qualifiedTopName` requires that `valid_module_name` holds for its
result, and this was _not_ the case for namespace packages.
To fix this, we extend `valid_module_name` to include the module names
of _any_ folder, not just regular package (which are the ones where
there's a `__init__.py` in the folder). Note that this doesn't simply
include all folders -- only the ones that result in valid module names
in Python.
Adds `hasOverloadDecorator` as a predicate on functions. It looks for
decorators called `overload` or `something.overload` (usually
`typing.overload` or `t.overload`). These are then filtered out in the
predicates that (approximate) resolving methods according to the MRO.
As the test introduced in the previous commit shows, this removes the
spurious resolutions we had before.
Moves the classes/predicates that _actually_ depend on points-to to the
`LegacyPointsTo` module, leaving behind a module that contains all of
the metrics-related stuff (line counts, nesting depth, etc.) that don't
need points-to to be evaluated.
Consequently, `Metrics` is now no longer a private import in
`python.qll`.
On `keras-team/keras`, this was producing ~200 million intermediate
tuples in order to produce a total of ... 2 tuples.
After the refactor, max intermediate tuple count is ~80k for the
charpred (and 4 for the new helper predicate).
These were causing the repo `gufolabs/noc` to spend ~30 seconds
evaluating `ControlFlowNode.strictlyDominates`. Just in case, I added
`overlay[caller] to the other instances of `pragma[inline]` as well.
