| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Addresses #101
|
|
* Remove `then_some` to stay compatible with earlier versions of Rust
* Fix CI builds to build on old versions of Rust
* Trying 1.61
|
|
* Deprecate support for actix-web 2.0
* Update actix.rs
Added QsForm to support application/x-www-form-urlencoded Web forms
* Update actix.rs with working adjustments
Added Into_inner function for both QsQuery and QsForm to behave like actix_webs default extractors
Made some fixes
* Added missing code to actix.rs for QsForm
* Update actix.rs
Added new trait "IntoInner"
Fixed QsFormConfig not correctly setting new settings correctly
* Update actix.rs
Removed trait requirement for using into_inner() now functions as a function.
* Update src/actix.rs
* Remove redundant config, add tests
* Add changelog
* Bump serde qs version
* Add Debug bound to docs example
---------
Co-authored-by: nMessage <135612238+nMessage@users.noreply.github.com>
|
|
Partially address #71
|
|
|
|
* fix unit type serialization issue
`()` and `A` returned a no key error previously. This is very
unergonimic if you just have a trait bound for Serialize and want to
generate an empty querystring `?`
* add support for deserializing unit structs
* Comment updates for serializer methods
|
|
|
|
|
|
|
|
This change makes non-strict mode percent-decode utf-8 lossily.
This means that invalid UTF-8 percent encoded sequences, such as `%E9`, are
decoded as the unicode replacement character (� `U+FFFD`).
Refs: https://github.com/samscott89/serde_qs/issues/43
|
|
|
|
|
|
* Add an `actix` feature to provide a nicer error message
for people migrating from <= 0.8
* Fix tests.
* Bump version in readme.
|
|
* Remove all default actix features
* Set MSRV
* Have `actix4` feature, instead of `actix`
* enable actix tests for all features
|
|
|
|
|
|
|
|
* add tracing, add tests, default to empty string
|
|
|
|
Fixes #40.
Thanks to @faulesocke for reporting.
|
|
|
|
Fixes #38
Thanks to @pooyamb for reporting.
|
|
|
|
|
|
|
|
|
|
* update dependencies
- actix-web v2
- percent encoding v2.1
- rust 2018 edition
- remove rustfmt no longer supported rules
* ci: add feature build matrix
* fix actix unit tests
Co-authored-by: Mario Reder <mreder1289@gmail.com>
|
|
resolves #23
|
|
This allows switching to non-strict parsing mode
resolves #21
|
|
|
|
|
|
|
|
|
|
|
|
When parsing a value, always just consume all the text, and ignore
special encoded of brackets.
(Hence non-strict mode only applies to keys)
|
|
|
|
|
|
Removes the dependency on the `url` crate because we only need
the percent decoding, with our own encoding set.
All keys/values are conservatively percent-encoded, but the square
brackets are untouched. Compatible with strict decoding, and
potentially less error prone.
|
|
Permits encoded brackets, and will generally tolerate parsing
errors where possible.
Permitting, for example, `a%5B[0%5D=1` to decode as `a: vec![1]`.
Default behaviour is strict.
|
|
|
|
- Only perform the percent decoding at the final step (fixes #10).
- Reduce allocations by borrowing as much as possible.
|
|
Fixes #7.
|
|
Clean up the code in places.
|
|
This reverts commit 6429b975c6cbcf5d7e0c2b1d02026f8d46b5c50f.
|
|
When deserializing types such as TupleStructs, if the `Level`
value is a map, let's call `visit_map` instead of coercing it
to a sequence immediately.
|
|
|
|
Simplify `ser` logic, and expand `de` functionality.
|
|
|
|
|
