Create a tool for stubs in cstruct v4 #72
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #72 +/- ##
==========================================
- Coverage 92.33% 91.50% -0.84%
==========================================
Files 20 21 +1
Lines 2180 2508 +328
==========================================
+ Hits 2013 2295 +282
- Misses 167 213 +46
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
dissect/cstruct/types/base.py
Outdated
| def _type_stub(cls, name: str = "") -> str: | ||
| return f"{name}: {cls.__name__}" | ||
|
|
||
| def to_stub(cls, name: str) -> str: |
There was a problem hiding this comment.
Maybe name this differently? Also I think this might as well be a private method. Maybe _to_type_stub? to_stub might be confusing because it also makes me think of the compiler.
There was a problem hiding this comment.
i renamed it to to_type_stub as it is a function that gets called externally.
There was a problem hiding this comment.
I don't agree on the "externally" part, I'd classify this usage as internal and would never expect anyone to ever call this externally. This is not part of any expected usage API.
|
I added some more tests, and noticed there was still some issues with import stuff from a file. As all signatures/names dissapear once the stub file is there. I attempted to fix this by adding more checks in |
|
I have made a lot of suggested changes with 7c93f22 |
There was a problem hiding this comment.
PR Overview
This PR introduces a new stub generation tool for cstruct v4 and updates many core type modules. Key changes include a transition from using public ‘name’ properties to the internal ‘_name’ attribute for fields, significant type‐annotation improvements (including generics and the addition of BaseArray), and refactoring in parsing and compiler functionalities for better consistency and type safety.
Reviewed Changes
| File | Description |
|---|---|
| dissect/cstruct/types/structure.py | Updated field name references to use internal “_name” attribute. |
| dissect/cstruct/types/base.py | Enhanced type annotations and introduced BaseArray. |
| dissect/cstruct/types/pointer.py | Added generics and refined pointer arithmetic operations. |
| dissect/cstruct/parser.py | Simplified struct registration logic by adding a “register_if_name” flag. |
| pyproject.toml | Added a new command line script “cstruct-stubgen” and updated ruff settings. |
Copilot reviewed 31 out of 31 changed files in this pull request and generated no comments.
Comments suppressed due to low confidence (2)
dissect/cstruct/types/structure.py:97
- Ensure that all references to a field's name consistently use the internal '_name' attribute. Verify that no residual use of 'field.name' remains, as mismatches could lead to runtime errors.
if field._name in lookup and field._name != "_":
dissect/cstruct/parser.py:225
- [nitpick] The combined use of 'register' and 'register_if_name' in the _struct method makes the registration logic complex and potentially error-prone. Consider refactoring this logic to separate struct creation from type registration more clearly for improved maintainability.
if register or register_if_name:
A small fix was to just return 0 if the cls is a BaseType inside `MetaType`
In the case multiple cstruct definitions are in one file, it will not overwrite one another.
Otherwise the generated class stub wouldn't be a valid class definition
This is for those structures that define a type and a variable with a name.
E.G.
```c
typedef struct _NAME {
...
} NAME;
```
now generates:
```python
class _NAME(...): ...
NAME: _NAME
```
Co-authored-by: Erik Schamper <1254028+Schamper@users.noreply.github.com>
| elif isinstance(typedef, str): | ||
| stub = f"{name}: TypeAlias = {typedef}" |
There was a problem hiding this comment.
I don't think this code will ever get reached. As the TypeAlias where there for the uint8 and those kinda types. These are now defined in cstruct proper no?
|
|
||
| if field.name: | ||
| result[field.name] = value | ||
| if field._name: |
There was a problem hiding this comment.
isn't field._name always defined as it either uses name or type.__name__?
| if field._name: | ||
| sizes[field._name] = stream.tell() - offset | ||
| result[field._name] = value |
The only thing that does not work as intended yet is the
dereferenceof pointer. As in, it doesn't return the desired type information but justT