Add config setting to avoid adding trailing slash to URLs #1719
Conversation
CodSpeed Performance ReportMerging #1719 will not alter performanceComparing Summary
|
| Bound(Bound<'py, PyUrl>), | ||
| Owned(PyUrl), |
There was a problem hiding this comment.
@davidhewitt is there a better way of doing this, or a better name?
There was a problem hiding this comment.
The point of this is to avoid allocating the Python object until it's absolutely needed. Maybe call it LazyUrlObject and have the variants be
| Bound(Bound<'py, PyUrl>), | |
| Owned(PyUrl), | |
| Object(Bound<'py, PyUrl>), | |
| Struct(PyUrl), |
| impl From<PyUrl> for Url { | ||
| fn from(value: PyUrl) -> Url { | ||
| value.lib_url | ||
| } | ||
| } | ||
|
|
||
| #[pymethods] | ||
| impl PyUrl { |
There was a problem hiding this comment.
Possibly query #[getter] also needs this treatment?
| Bound(Bound<'py, PyUrl>), | ||
| Owned(PyUrl), |
There was a problem hiding this comment.
The point of this is to avoid allocating the Python object until it's absolutely needed. Maybe call it LazyUrlObject and have the variants be
| Bound(Bound<'py, PyUrl>), | |
| Owned(PyUrl), | |
| Object(Bound<'py, PyUrl>), | |
| Struct(PyUrl), |
| EitherUrl::Py(py_url) => Ok(py_url), | ||
| EitherUrl::Rust(rust_url) => Bound::new(py, PyUrl::new(rust_url)), | ||
| EitherUrl::Bound(py_url) => Ok(py_url), | ||
| EitherUrl::Owned(py_url) => Bound::new(py, py_url), | ||
| } | ||
| } | ||
| } | ||
|
|
||
| impl CopyFromPyUrl for EitherUrl<'_> { |
There was a problem hiding this comment.
Seeing this again, this looks like it probably should have been AsRef<Url> and AsMut<Url> implementations.
| if self.remove_trailing_slash && s.ends_with('/') { | ||
| s = &s[..s.len() - 1]; | ||
| } | ||
| s | ||
| } | ||
|
|
||
| pub fn __repr__(&self) -> String { |
There was a problem hiding this comment.
Does __repr__() also need to be updated?
| @@ -490,7 +514,11 @@ fn unicode_url(lib_url: &Url) -> String { | |||
| s | |||
| } | |||
| _ => s, | |||
| }; | |||
| if remove_trailing_slash && s.ends_with('/') { | |||
There was a problem hiding this comment.
I think the point is to preserve the input, not always add or always strip. Can we keep track of the input ending with a trailing slash?
If not the config option is confusing: it should be remove_trailing_slash: bool = False to match the implementation.
There was a problem hiding this comment.
It seems that it is working as expected but I'm not entirely sure why:
u = Url('https://example.com/', add_trailing_slash=False)
str(u)
#> 'https://example.com/'
Preparation for fixing pydantic/pydantic#7186