fix(status): Enable to convert from i64 to hex_status by casting instead of parsing status.

[moko256]

This PR will enable to convert from i64 to hex_status by casting instead of parsing status.

Description

In elvish shell on Windows, the status code is passed as u32, and it can't parse as i32, so hex_status is same as status.
This PR will replace parsing status to i32 by casting from i64 to i32 and accept this.
The status code out of i32 will be rounded by underflow/overflow.

Motivation and Context

Closes #
Related #3459

Screenshots (if appropriate):

In powershell	In elvish before fix	In elvish after fix

[status]
disabled = false
format = "[$symbol$status $hex_status]($style) "

How Has This Been Tested?

• I have tested using MacOS
• I have tested using Linux
• I have tested using Windows

Checklist:

• I have updated the documentation accordingly.
• I have updated the tests accordingly.

[davidkna]

I think this will just work and avoid truncation:

Suggested change

	let hex_status = format!("0x{:X}", exit_code_int as u32);
	let hex_status = format!("0x{:X}", exit_code_int);

[moko256]

Avoiding truncation makes hex_status longer when status code is negative number like this:

✖-1 0xFFFFFFFFFFFFFFFF ❯

✖-1073741510 0xFFFFFFFFC000013A ❯

On Windows, status code is in range of u32 or i32 (e.g, MSVC exit() receive int), so I think this cast is necessary to make easy to see.

[davidkna]

In that case, should this also apply to int too? Furthermore, maybe this should be restricted to Windows (with another type ExitCodeTarget?) and I feel like this should be documented in the status docs.

[moko256]

Whether i32 or u32 the status fits in is consistent in one shell, and in supported platform the status code fits in 32bit.
Until the platform that have larger status code than 32bit will be supported, I think that assuming that the status code fits in 32bit is no problem and hex_status doesn't have to be restricted only for Windows. Similally, I think it doesn't have to apply to int too.
But I think that the size of status code that the module receives should be documented.

[moko256]

Instead of casting here, it is also possible to use i32 as ExitCode like this:

diff --git a/src/modules/status.rs b/src/modules/status.rs
index 5f643a95..31d26c2c 100644
--- a/src/modules/status.rs
+++ b/src/modules/status.rs
@@ -6,7 +6,7 @@ use crate::configs::status::StatusConfig;
 use crate::formatter::{string_formatter::StringFormatterError, StringFormatter};
 use crate::segment::Segment;
 
-type ExitCode = i64;
+type ExitCode = i32;
 type SignalNumber = u32;
 #[derive(PartialEq)]
 enum PipeStatusStatus<'a> {
@@ -101,15 +101,16 @@ fn format_exit_code<'a>(
     config: &'a StatusConfig,
     context: &'a Context,
 ) -> Result<Vec<Segment>, StringFormatterError> {
-    let exit_code_int: ExitCode = match exit_code.parse() {
-        Ok(i) => i,
+    // First, parse as i64 to accept both i32 or u32, then normalize to i32.
+    let exit_code_int: ExitCode = match exit_code.parse::<i64>() {
+        Ok(i) => i as ExitCode,
         Err(_) => {
             log::warn!("Error parsing exit_code string to int");
             return Ok(Vec::new());
         }
     };
 
-    let hex_status = format!("0x{:X}", exit_code_int as u32);
+    let hex_status = format!("0x{:X}", exit_code_int);
 
     let common_meaning = status_common_meaning(exit_code_int);
 

[davidkna]

In that case, I think it's better to leave int as-is for now.

[moko256]

Done, I pushed the diff to this PR. Thanks for reviewing!

[davidkna]

Please also add a note in the status module description in docs/config/README.md about this and that hex_status is unsigned. (sorry, looks like you ended up removing that).

[moko256]

Done.

[davidkna]

Thanks for handling this @moko256!