Why do these two command variants, which are similar, give different outputs?

Question

Command 1:

$ rm hello.txt 2>/dev/null || { echo “Couldn’t delete hello.txt” }
"Couldn't delete hello.txt"

Command 2

$ rm hello.txt 2>/dev/null || { echo 'Couldn’t delete hello.txt' }
Couldn't delete hello.txt

---

Note: hello.txt doesn't exist in the current directory.

Answer 1

The first command

rm hello.txt 2>/dev/null || { echo “Couldn’t delete hello.txt” }

contains the characters <U+201C>(LEFT DOUBLE QUOTATION MARK), <U+2019> (RIGHT SINGLE QUOTATION MARK) and <U+201D> (RIGHT DOUBLE QUOTATION MARK) which are in no way special to the shell and are output as such.

The second command

rm hello.txt 2>/dev/null || { echo 'Couldn’t delete hello.txt' }

contains a string enclosed in single quotes; the character between n and t is again <U+2019>, which is not special to the shell.

Answer 2

The pair “” is called smart quotes. Similarly, your ’ in Couldn’t is also a smart quote.

Those are unlike normal quotes in the sense that they're curly to differentiate the begin and end of the quoting part. They have no meaning in most (if not all) programming/scripting languages I know and are just the same as other Unicode characters.

So your first echo doesn't actually have any quote and the command will receive those words as separate arguments. You can check that easily

$ for s in “Couldn’t delete hello.txt”; do echo $s; done
“Couldn’t
delete
hello.txt”

$ for s in 'Couldn’t delete hello.txt'; do echo $s; done
Couldn’t delete hello.txt

In fact your output is wrong (did you really copy paste it?) because there's no way smart quotes become normal quotes after echo. The output should still be smart quote as in my example above

As you can see, the second case is a single string containing a Unicode character outside the ASCII range, so the quotes will be stripped by the shell

See What are curly quotes and can I use them in my code?