Description

Before this PR, you can access rendered error values that are raised in a try/catch block by accessing the rendered element of the catch error value:

$ try { ls nonexist.txt } catch {|e| print "my cool error:" $e.rendered }
my cool error:
nu::shell::directory_not_found

  × Directory not found
  help: /home/rose/nonexist.txt does not exist

However, the rendered errors don't include the labels present in the real rendered error, which would look like this:

$ ls nonexist.txt
Error: nu::shell::directory_not_found

  × Directory not found
   ╭─[entry #46:1:4]
 1 │ ls nonexist.txt
   ·    ──────┬─────
   ·          ╰── directory not found
   ╰────
  help: /home/rose/nonexist.txt does not exist

After this PR, the rendered error includes the labels:

$ try { ls nonexist.txt } catch {|e| print "my cool error:" $e.rendered }
my cool error:
Error: nu::shell::directory_not_found

  × Directory not found
   ╭─[entry #4:1:10]
 1 │ try { ls nonexist.txt } catch {|e| print "my cool error:" $e.rendered }
   ·          ──────┬─────
   ·                ╰── directory not found
   ╰────
  help: /home/rose/nonexist.txt does not exist

This change is accomplished by using the standard error formatting code to render an error. This respects the error theme as before without any extra scaffolding, but it means that e.g., the terminal size is also respected. I think this is fine because the way the error is rendered already changed based on config, and I think that a "rendered" error should give back exactly what would be shown to the user anyway.

@fdncred, let me know if you have any concerns with the way this is handled since you were the one who implemented this feature in the first place.

User-Facing Changes

The rendered element of the try/catch error record now includes labels in the error output.

Tests + Formatting

• 🟢 toolkit fmt
• 🟢 toolkit clippy
• 🟢 toolkit test
• 🟢 toolkit test stdlib

After Submitting

N/A