lib/ureq
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
682 Commits 1 Branch 0 Tags
a8ca1ec0855c4e32e34243e10d2722be6348b8e4
Commit Graph

682 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Martin Algesten
2c29cc230c Remove Sync requirement of ReadWrite trait 
3rd party TLS connections are currently required to provide a Sync
guarantee. Since ureq will only ever use the connection on a single
thread, this is not necessary.

The reason we end up with this bound is because we want Response and
Error to be Sync, in which case Rust's automatic inferral of Sync
fails.

This change "masks" the Stream in a wrapper making it Sync.

Close #474
 2022-01-30 21:50:11 +01:00
Martin Algesten
219b5edf9e Rename test function as_write_vec -> into_written_bytes 
Tests use `Response::as_write_vec` to inspect the outgoing HTTP/1.1
request line and headers. The current version has two problems:

1. Called `as_write_vec` when it actually returns a `&[u8]`.
2. Inspects/uses the `Response::stream` without consuming `Response`.

The first problem is trivial, but the second is subtle. Currently all
calls on `Response` that works with the internal `Response::stream`
consumes `self` (`into_string`, `into_reader`).

`Response` is by itself `Send + Sync`, and must be so because the
nested Stream is `Read + Write + Send + Sync`. However for
implementors of `TLSStream`, it would be nice to relax the `Sync`
requirement.

Assumption: If all fields in Response are `Sync` except
`Response::stream`, but any access to `stream` consumes `Response`, we
can consider the entire `Response` `Sync`.

This assumption can help us relax the `TlsStream` `Sync` requirement
in a later PR.
 2022-01-29 17:32:18 +01:00
Martin Algesten
6e5041044b Rename trait HttpsStream -> ReadWrite and make it public 
Also provide an example of how to use it.
 2022-01-22 10:41:22 +01:00
Daniel
140aa5901f Add tcp no_delay option (#465) 
* Add tcp no_delay option
 2022-01-06 23:22:33 +01:00
Martin Algesten
007ce44ea7 Bump version to 2.4.0 2021-12-23 08:38:00 +01:00
Martin Algesten
adbd9144af Update changelog for 2.4.0 2021-12-23 08:38:00 +01:00
Martin Algesten
3b68dce23a Make Middleware always Send + Sync + 'static 2021-12-22 07:58:45 +01:00
Martin Algesten
81c0e66ec7 MiddlewareNext iterator without Box 
Thanks @jsha for showing me the way.
 2021-12-22 07:58:45 +01:00
Martin Algesten
43680335dc Middleware chain based on iterator 2021-12-22 07:58:45 +01:00
Martin Algesten
a2d62368f3 Remove Option wrapper around middleware 2021-12-22 07:58:45 +01:00
Martin Algesten
09ecb6ffd6 Implement middleware function 2021-12-22 07:58:45 +01:00
Martin Algesten
7b2f28bbc2 Tidy up Response::url initialization 2021-12-22 07:58:45 +01:00
Martin Algesten
87cf59d399 Doc fix 2021-12-22 07:58:45 +01:00
Martin Algesten
98ac6313e5 Agent/AgentBuilder impl Debug manually 2021-12-22 07:58:45 +01:00
Jacob Hoffman-Andrews
c55ac31087 Fix quoting in test runner 2021-12-20 09:20:52 +01:00
Jacob Hoffman-Andrews
352bf19f40 Fix doctests 2021-12-20 09:19:25 +01:00
Jacob Hoffman-Andrews
2df70168c4 Use ? instead of unwrap in examples (#458) 
This is recommended by the Rust API Guidelines:

https://rust-lang.github.io/api-guidelines/documentation.html#examples-use--not-try-not-unwrap-c-question-mark

One exception: When we need to unwrap an Option, the examples still use
.unwrap(). The alternative would be something like
`.ok_or(SomeErrorHere)?`, which feels like an awkward way to deal with
an Option. This might get better with NoneError:

https://docs.rs/rustc-std-workspace-std/1.0.1/std/option/struct.NoneError.html

I also rearranged some examples that used turbofish to use type
annotations. I think type annotations are more familiar to early Rust
users (and they use fewer characters and less punctuation, which is
always nice).
 2021-12-19 20:04:30 -08:00
Jacob Hoffman-Andrews
2563df4f62 Bump rustls to >=0.20.1 
Without this we get a build error because we depend on an impl of Error
on InvalidDnsNameError.
 2021-12-20 02:39:38 +01:00
Martin Algesten
738cb84aec Enable 'gzip' by default 2021-12-20 02:38:18 +01:00
Martin Algesten
4aa5480543 Improve doc after review 2021-12-19 21:17:26 +01:00
Martin Algesten
8f476e9d29 cargo fmt 2021-12-19 21:17:26 +01:00
Martin Algesten
75f6be8ff8 Format error message like suggested in review 2021-12-19 21:17:26 +01:00
Martin Algesten
d1a2c630f5 Update src/rtls.rs 
Co-authored-by: Jacob Hoffman-Andrews <github@hoffman-andrews.com>
 2021-12-19 21:17:26 +01:00
Martin Algesten
324c5c074f Update src/error.rs 
Co-authored-by: Jacob Hoffman-Andrews <github@hoffman-andrews.com>
 2021-12-19 21:17:26 +01:00
Martin Algesten
f3857eed00 Ensure we provide a Transport::message() when we can 2021-12-19 21:17:26 +01:00
Martin Algesten
6a094db668 Provide access methods for error::Transport 
.message() and .kind() so that the user can get the details of a transport
error. This also documents the motivation for the different fields
 2021-12-19 21:17:26 +01:00
Martin Algesten
158a6c6f03 Clarify serde doc on Response::into_json (#451) 
Close #428
 2021-12-19 11:41:44 -08:00
Malloc Voidstar
598ebf4393 Remove Content-Encoding and length when decompressing 
This should lower the chance of breakage. Probably also more proper for a client library.
 2021-12-19 14:01:56 +01:00
Malloc Voidstar
6281a0bea6 Add tests for automatic decompression 2021-12-19 14:01:56 +01:00
Malloc Voidstar
9a070fb1ad And to lib.rs, woops 2021-12-19 14:01:56 +01:00
Malloc Voidstar
024bcaac99 Add gzip/brotli to README 2021-12-19 14:01:56 +01:00
Malloc Voidstar
23e993689a Rename Decompressor to BrotliDecoder to align with gzip 2021-12-19 14:01:56 +01:00
Malloc Voidstar
873e6066f3 Add support for gzip and brotli 
Automatically sends the Accept-Encoding header on requests.

Not runtime-configurable, only with Cargo features.
 2021-12-19 14:01:56 +01:00
Martin Algesten
0f0dec5f32 Fixes after feedback 2021-12-19 11:00:39 +01:00
Martin Algesten
2b0eca9827 Move auth header on redirect to unit construction 
The auth header stripping was in the wrong place (when serializing the request),
rather than in the construction of the Unit, where it ought to be.

This also makes redirect header retention testable.
 2021-12-19 11:00:39 +01:00
Martin Algesten
f941717a5d Mark RedirectAuthHeaders as non_exhaustive 2021-12-19 11:00:39 +01:00
Martin Algesten
27533cf31b Use match expression against RedirectAuthHeaders to ensure exhaustive 2021-12-19 11:00:39 +01:00
Martin Algesten
c59632bd97 Use Url (instead of String) in internal history var 2021-12-19 11:00:39 +01:00
Martin Algesten
2a5ab43c8c Expose RedirectAuthHeaders enum at crate root 2021-12-19 11:00:39 +01:00
Martin Algesten
8d33318a23 Rename setter to be consistent with other builder settter 2021-12-19 11:00:39 +01:00
Martin Algesten
b033e0f614 Doc fixes to RedirectAuthHeaders 2021-12-19 11:00:39 +01:00
llde
26a3715f62 Revert move of variable 2021-12-19 11:00:39 +01:00
llde
653f791638 Create new configuration option for redirect preserving authorization header in Agent. Handle new option in Unit 2021-12-19 11:00:39 +01:00
llde
38ad90307d Preserve Authorization in same host redirects, when scheme and port are equals 2021-12-19 11:00:39 +01:00
soruh
61e2402045 deprecate using SerdeValue, SerdeMap and serde_to_value 
as they are just wrapper around serde_json::{Value, Map, to_value}.
Instead expose the serde_json and serde crates used.
 2021-12-17 20:12:07 +01:00
soruh
59f1fab4d3 allow send_json to send any serde::Serialize value 2021-12-17 20:12:07 +01:00
Jacob Hoffman-Andrews
56276c3742 Add support for alternate TLs implementations. 2021-12-17 17:47:30 +01:00
Malloc Voidstar
1c1dfaa691 Bump rustls to 0.20.1; add src to rustls error (#438) 2021-12-17 00:32:00 -08:00
Martin Algesten
876d5757a7 Update changelog 2021-11-28 22:45:16 +01:00
Martin Algesten
3709221a28 Handle 400-error from client.badssl.com 2021-11-28 22:45:16 +01:00