os: Fix File.read() in JS backends [Issue : #20501]

[GGRei]

Fixed #20501 File.read() function for the JavaScript backend.

The issue was due to incorrect array access in the JavaScript generated from V code.

To respond at ticket expectation :

• Using the js backend: expected to either read a non-zero amount of bytes, or raise an error.

The original function already includes a try-catch block to manage errors. Unless specifically requested, I don't plan to add additional error handling

- Using the js_browser: expected to either fail compilation (since File I/O is unavailable in the browser), or for os.open() to always return an error.

the existing code already handles errors. If open_file() is called in a non-NodeJS environment, an error is generated. However, it appears that the test in ticket was conducted using Node.js, which could explain the observed results.

I'm always available for any suggestions/modifications.

[spytheman]

(rebased over current master)

[spytheman]

The observed behavior for the C backend and for the JS backend is still different.
You can use this file open_and_read_from_file_test.js.v to try (and please add it to vlib/os/ for example, to prevent future regressions).

import os

fn test_read_from_file() {
	mut buf := []u8{len: 10}
	f := os.open(@FILE)!
	n := f.read(mut &buf)!
	println(buf[..n])
}	

[spytheman]

Note that the C version https://github.com/vlang/v/blob/master/vlib/os/file.c.v#L233 , reads up to buf.len bytes, and it returns how many bytes it could read.
It does not read the whole file.

[GGRei]

Done. let me know if everything is okay. Thanks.

[spytheman]

I can confirm, that the fix for File.read/1 is working.

It still prints differently, but fixing that is out of scope for this PR.

Excellent work.