Refatoring of all functions that print to stdout/err so they accept a generic type that implements the Display trait

[MitchellBerend]

The following locations are potential candidates. These should not have a big impact on the rest of the code base, but they allow all the errors and other enums to have the same implementation strategy.

I would love some feedback on this.

• src/infra/err.rs
  

  tool-sync/src/infra/err.rs

  Lines 3 to 13 in 65b53bd

  	/// Print an error message and exit with code 1
  	pub fn abort_with(err_msg: &str) -> ! {
  	eprintln!(
  	r#"Aborting 'tool-sync' with error:
  	* {}"#,
  	err_msg
  	);
  	process::exit(1);
  	}

• src/infra/err.rs
  

  tool-sync/src/infra/err.rs

  Lines 15 to 31 in 65b53bd

  	/// Print an error message, suggesting opening an issue and exit with code 1
  	pub fn abort_suggest_issue(err_msg: &str) -> ! {
  	eprintln!(
  	r#"Aborting 'tool-sync' with error:
  	* {}
  	Please, open an issue in the 'tool-sync' repository and provide as many
  	details as possible to diagnose the problem if you want to get help with
  	this issue:
  	* https://github.com/chshersh/tool-sync/issues/new"#,
  	err_msg
  	);
  	process::exit(1);
  	}

• src/infra/err.rs
  

  tool-sync/src/infra/err.rs

  Lines 33 to 37 in 65b53bd

  	/// Print just the message and exit
  	pub fn abort(err_msg: &str) -> ! {
  	eprintln!("{}", err_msg);
  	process::exit(1);
  	}

• src/sync/prefetch.rs
  

  tool-sync/src/sync/prefetch.rs

  Lines 39 to 42 in 65b53bd

  	fn expected_err_msg(&self, tool_name: &str, msg: &str) {
  	let tool = format!("{}", style(tool_name).cyan().bold());
  	self.pb.println(format!("{} {} {}", ERROR, tool, msg))
  	}

• src/sync/prefetch.rs
  

  tool-sync/src/sync/prefetch.rs

  Lines 44 to 64 in 65b53bd

  	fn unexpected_err_msg(&self, tool_name: &str, msg: &str) {
  	let tool = format!("{}", style(tool_name).cyan().bold());
  	let err_msg = format!(
  	r#"{emoji} {tool} {msg}
  	If you think you see this error by a 'tool-sync' mistake,
  	don't hesitate to open an issue:
  	* https://github.com/chshersh/tool-sync/issues/new"#,
  	emoji = ERROR,
  	tool = tool,
  	msg = msg,
  	);
  	self.pb.println(err_msg);
  	}
  	fn finish(&self) {
  	self.pb.finish()
  	}
  	}

• src/sync/progress.rs
  

  tool-sync/src/sync/progress.rs

  Lines 87 to 93 in 65b53bd

  	pub fn failure(&self, pb: ProgressBar, tool_name: &str, tag: &str, err_msg: String) {
  	pb.set_prefix(self.fmt_prefix(FAILURE, tool_name, tag));
  	let failure_msg = format!("{}", style(err_msg).red());
  	pb.set_message(failure_msg);
  	pb.finish();
  	}

[chshersh]

@MitchellBerend Thanks a lot for collecting all the usages of errors!

I marked this issue as blocked by #88 as the Display refactoring needs to happen first. Using Display instead of custom display() is definitely a welcome change 👏🏻

For this particular issue, it's a bit hard to evaluate whether it'll simplify the code or make it less maintainable. In theory, specifying the err_msg argument as some polymorphic type <Msg: Display> could open a road to future refactoring where we use some specialised trait for errors. E.g. std::error::Error or something from eyre.

But in any case, if refactoring proposed by this issue makes function signatures more generic and reduces usages of to_string() and display() on call sites, I'm happy to apply this refactoring 🆗

[MitchellBerend]

std::error::Error does not have the Display trait implemented and rust requires either the trait or the struct/enum to be defined in the crate itself, so I'm not sure if we can use that with the proposed change. Im not sure how eyre gets around this though.

[MitchellBerend]

Should we add a blocked label btw? That seems like a good idea so others don't waste time on issues that can't actually be solved yet.

[chshersh]

The blocked label sounds like a good idea! I've created it and added it to a blocked issue.

However, this issue is no longer blocked as #88 actually was already done by @zixuanzhang-x 😅

There's no way to specify what issue is blocking inside the label text. So I propose to put the "Blocked by #NUMBER" string in the beginning of every issue that is blocked by something else.