fix(process): improve stdout encoding handling in read_line functions #1471

[Tunglies]

I developed on macOS, need to make sure encoding behavior same as Windows.
Sidecar:

use encoding_rs;
use std::io::{self, Write};
use std::thread;
use std::time::Duration;

fn main() {
    let stdout = io::stdout();
    let mut handle = stdout.lock();
    let utf8_output = "[download] Destination: PEDRO - Jaxomy, Agatino Romero, Raffaella Carrà (TikTok Song) [XzilCu9PcZk].f616.mp4\n";
    const TARGET_ENCODING: &'static encoding_rs::Encoding = encoding_rs::WINDOWS_1252;
    let (encoded_bytes, _, malformed) = TARGET_ENCODING.encode(utf8_output);
    if malformed {
        eprintln!("[SIM] Warning: some characters cannot be encoded in Windows-1252!");
    }
    loop {
        match handle.write_all(encoded_bytes.as_ref()) {
            Ok(_) => {
                eprintln!(
                    "[SIM] Simulated output: successfully wrote {} bytes.",
                    encoded_bytes.len()
                );
            }
            Err(e) => {
                eprintln!("[SIM] Error writing to stdout: {}", e);
                break;
            }
        }
        thread::sleep(Duration::from_millis(1000));
    }
    let check_a_grave = TARGET_ENCODING.encode("à").0;
    eprintln!(
        "[SIM] Check: 'à' is encoded in Windows-1252 as bytes {:?}",
        check_a_grave
    );
}

User side, code snippet

let command = app_handle
            .clone()
            .shell()
            .sidecar(EXTERNAL_BIN)
            .expect("bin");
        let (mut rx, _) = command.spawn().expect("spawn");

        tauri::async_runtime::spawn(async move {
            while let Some(event) = rx.recv().await {
                if let CommandEvent::Stdout(line) = event {
                    let output = match String::from_utf8(line) {
                        Ok(s) => s,
                        Err(e) => {
                            println!("Invalid UTF-8 sequence: {}", e);
                            String::from_utf8_lossy(&e.into_bytes()).into_owned()
                        },
                    };
                    println!("{}", output);
                }
            }
        });

Output with mainline

[download] Destination: PEDRO - Jaxomy, Agatino Romero, Raffaella Carr� (TikTok Song) [XzilCu9PcZk].f616.mp4

Output with PR

[download] Destination: PEDRO - Jaxomy, Agatino Romero, Raffaella Carrà (TikTok Song) [XzilCu9PcZk].f616.mp4

[Tunglies]

CC @FabianLars, this is initial start. Potential to continue improve with your advice.

there's something for us to fix #1471 (comment)

If auto encoding is required, current hardcoded for encoding need change to system encoding page detect, and brings more cross-platform support.

[github-actions]

Package Changes Through f5858bf

There are 21 changes which include dialog-js with minor, dialog with minor, log with minor, log-js with minor, localhost with patch, barcode-scanner with patch, barcode-scanner-js with patch, deep-link with patch, deep-link-js with patch, shell with patch, shell-js with patch, http with patch, http-js with patch, nfc with patch, nfc-js with patch, updater with minor, updater-js with minor, upload with minor, upload-js with minor, websocket with patch, websocket-js with patch

Planned Package Versions

The following package releases are the planned based on the context of changes in this pull request.

package	current	next
api-example	2.0.38	2.0.39
api-example-js	2.0.34	2.0.35
deep-link-example-js	2.2.8	2.2.9
barcode-scanner	2.4.2	2.4.3
barcode-scanner-js	2.4.2	2.4.3
deep-link	2.4.5	2.4.6
deep-link-js	2.4.5	2.4.6
dialog	2.4.2	2.5.0
dialog-js	2.4.2	2.5.0
http	2.5.4	2.5.5
http-js	2.5.4	2.5.5
localhost	2.3.1	2.3.2
log	2.7.1	2.8.0
log-js	2.7.1	2.8.0
nfc	2.3.3	2.3.4
nfc-js	2.3.3	2.3.4
shell	2.3.3	2.3.4
shell-js	2.3.3	2.3.4
single-instance	2.3.6	2.3.7
updater	2.9.0	2.10.0
updater-js	2.9.0	2.10.0
upload	2.3.2	2.4.0
upload-js	2.3.2	2.4.0
websocket	2.4.1	2.4.2
websocket-js	2.4.1	2.4.2

Add another change file through the GitHub UI by following this link.

---

Read about change files or the docs at github.com/jbolda/covector

[FabianLars]

@Legend-Master Do you know if this could also cause issues? This feels like it depends on the spawned process eg iirc node also uses utf8 on all platforms by default 🤔

[Legend-Master]

Since this is the raw output of the command in Vec<u8>, I don't think we should alter/preprocess it here to be honest

And for the encoding, I think it really depends on the command/program you spawned, it's not something we could assume (by the way, windows-1252 code page / encoding isn't the one used in all applications, it's only for apps using ANSI with ANSI Latin 1; Western European (Windows))

https://learn.microsoft.com/en-us/windows/win32/intl/code-page-identifiers

[Tunglies]

I found re-export usage use tauri_plugin_shell::process::Encoding;, use that we can do

let decoder = encoding_rs::WINDOWS_1252;
let (cow, _, had_errors) = recoder.decode(&line);
let output: String = cow.into_owned();

it's not something we could assume

Shall we leave this and documents on the usage of re-export Encoding? #1471 isn't a bug, user should consider encoding when use.

[Legend-Master]

Our JavaScript API defaulting to assume UTF-8 which I think is more or less in line with the Rust behavior (since JS one defaults to string while the Rust one is Vec<u8>), I think maybe we mirror the SpawnOptions.encoding option from JS? (and of course documentations)

#1471 isn't a bug, user should consider encoding when use.

Yeah, the program can output whatever they want technically, we can only provide sensible defaults that works for most

[Tunglies]

sturcture Command {
 encoding: Encoding
 ...
}
app_handle.sidecar(BIN).output_line_encoding(Encoding::Type)

Some thing like this? The .spawn only returns rx and child. Additional field of Command structure needed. And this should only takes effect on not-raw output.

[Legend-Master]

Ah, I see, let's just do the documentation work for now then