TLS connection times out

[pl4nf0ur]

This very simple example fails due to a TLS timeout with both the rustls backend as well as the native-tls:

let agent: ureq::Agent = ureq::Agent::config_builder()
    .timeout_connect(Some(Duration::from_secs(5)))
    .build()
    .into();

agent.get("https://google.com").call().unwrap();

->

called `Result::unwrap()` on an `Err` value: Timeout(Connect)

I tried doing the same w/ reqwest via:

reqwest::blocking::Client::new()
    .get("https://google.com")
    .send()
    .unwrap();

as well as w/ curl, and they both work.

Interestingly, if I send the request to https://httpbin.org it works, but most websites cause the request to time out.

[algesten]

Hi there! I don't think we get enough information to do anything here.

Could you run ureq with RUST_LOG=trace and post the results?

[mrsarm]

I got the same @algesten , this was blocking me to migrate a ureq v2 codebase that was running fine in the same machine. So I tested the same example here and I got the same Timeout error indeed, that it's the same I was getting in my project so I created a Github project https://github.com/mrsarm/ureqtest with the same minimal code posted here, and indeed I get the same error again, here is the stacktrace:

$ RUST_BACKTRACE=1 cargo run
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.02s
     Running `target/debug/ureqtest`
GET https://google.com ...

thread 'main' panicked at src/main.rs:13:27:
called `Result::unwrap()` on an `Err` value: Timeout(Connect)
stack backtrace:
   0: __rustc::rust_begin_unwind
             at /rustc/1159e78c4747b02ef996e55082b704c09b970588/library/std/src/panicking.rs:697:5
   1: core::panicking::panic_fmt
             at /rustc/1159e78c4747b02ef996e55082b704c09b970588/library/core/src/panicking.rs:75:14
   2: core::result::unwrap_failed
             at /rustc/1159e78c4747b02ef996e55082b704c09b970588/library/core/src/result.rs:1765:5
   3: core::result::Result<T,E>::unwrap
             at /home/user/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/result.rs:1167:23
   4: ureqtest::main
             at ./src/main.rs:13:27
   5: core::ops::function::FnOnce::call_once
             at /home/user/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/ops/function.rs:253:5
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

But even more surprising was that when I ran the same test on Github Actions so you could see there the error and from a different environment, it did work ! so I ran it again from my machine... same error again and again, until I did change my Wi-Fi and it worked 🤷

Then I switched back to the original network and did work as well, but there were no network issues, as I mentioned, in another project where I was migrating from ureq v2 to v3 the v2 version worked, and as soon I switched to the v3 branch with the code migrated, it failed over and over, so my assumption is that there must be something in ureq, or more probably in its dependencies that has some sensitivity with bad connections that normally other peaces of software (like ureq v2) don't, but what is the problem exactly and how to reproduce it exactly may be almost impossible.