Bash范围的非数字或字母

Question

I played around with bash {..} constructs today. I knew

{a..z}

would generate all letters,

{0..9}

digits etc. (numbers in general obviously), but By mistake I got

{Z..a}

yielding:

Z [  ] ^ _ ` a

The characters in between "Z" (90) and "a" (97) are the ASCII 91-96. The astute reader will notice there is a character missing - "\\", 92. I'm guessing because of it's special nature. Is this expected behavior as output? Specifically, I'm guessing the \\ is being used to escape the space in front of it after substitution, but @John1024 notes that:

echo {Z..a}a

will complain on missing backticks, while the previous version (no a ) does not. How exactly is substitution working? Is there a bug?

Second, I guessed the range operator is cooler than I thought and can do any range of ASCII characters I choose, but {[.._} for example fails. Am I missing something to make this work or is this just a curiosity? Are there any more ranges besides letters/digits I can use? and if not, why not do nothing (fail, echo as is) for 'jumping' from caps to lower?

Answer 1

The \\ is being generated; however, it subsequently appears to be treated as escaping the following space. Compare:

$ printf '%s\n' 'Z' '[' ']' '^' '_' '`' 'a'
Z
[
]
^
_
`
a
$ printf '%s\n' {Z..a}
Z
[

]
^
_
`
a

The extra blank line following the [ is the space escaped by the backslash generated by {Z..a} .

Answer 2

A special variable obase can be used with bc to print almost any character range(s):

for n in {91..95}; do printf "\x$(echo "obase=16; $n" | bc)"; done

Result:

[\]^_

↳ https://www.gnu.org/software/bc/manual/html_mono/bc.html#TOC6