Zsh Mailing List Archive
Messages sorted by:
Reverse Date,
Date,
Thread,
Author
Odd behavior with various (q) array modifiers and non-printable characters (backspace, newline)
- X-seq: zsh-users 26917
- From: Zach Riggle <zachriggle@xxxxxxxxx>
- To: Zsh Users <zsh-users@xxxxxxx>
- Subject: Odd behavior with various (q) array modifiers and non-printable characters (backspace, newline)
- Date: Wed, 11 Aug 2021 15:29:07 -0500
- Archived-at: <https://zsh.org/users/26917>
- List-id: <zsh-users.zsh.org>
Hello all!  It's me again, hopefully with a useful question.
TLDR: (q-) behaves unexpectedly for \010 aka "\b", and there are other
corner cases that may be unintended.
I love the (q-) modifier for array expansion, since it allows printing
out e.g. a command and then running it.
    run() {
        echo '$' "${(q-)@}"
        "$@"
    }
The docs (https://zsh.sourceforge.io/Doc/Release/Expansion.html,
search for "q-" to jump to the right area) state:
> If a q- is given (only a single q may appear), a minimal form of single quoting is used that only quotes the string if needed to protect special characters. Typically this form gives the most readable output.
> If a q+ is given, an extended form of minimal quoting is used that causes unprintable characters to be rendered using $’...’. This quoting is similar to that used by the output of values by the typeset family of commands.
However, it appears that half of the q modifiers handle 0x08 / \010 /
"\b" aka "backspace" in an unexpected manner.  Notably, (q) is
incorrect, (qq) (q+) and (q-) embed a literal backspace, which causes
unexpected output (it's there if you pipe to e.g. xxd, but it should
be escaped).
    $ mysed=( gsed "s/\ba\b/x/" )
    $ echo "aaaa a aaa" | "${mysed[@]}"
    aaaa x aaa
    $ echo ${(q)mysed}
    gsed s/\ba\b/x/
    $ echo ${(qq)mysed}
    'gsed' 's/x/'
    $ echo ${(qqq)mysed}
    "gsed" "s/\ba\b/x/"
    $ echo ${(qqqq)mysed}
    $'gsed' $'s/\ba\b/x/'
    $ echo ${(q+)mysed}
    gsed 's/x/'
    $ echo ${(q-)mysed}
    gsed 's/x/'
This issue also appears with newlines, but for these one could argue
it's entirely legitimate to have a *literal* newline that breaks the
quoted statement onto multiple lines.  Once again, the same culprits
are (qq), (q-), and (q+).  Interestingly, bare (q) is actually
incorrect.
    $ newline=( aa "b\nb" cc )
    $ echo ${(q)newline}
    aa b\nb cc
    $ echo ${(qq)newline}
    'aa' 'b
    b' 'cc'
    $ echo ${(qqq)newline}
    "aa" "b\nb" "cc"
    $ echo ${(qqqq)newline}
    $'aa' $'b\nb' $'cc'
    $ echo ${(q-)newline}
    aa 'b
    b' cc
    $ echo ${(q+)newline}
    aa 'b
    b' cc
IMHO both newline and backspace should be escaped at LEAST for (q-)
and (q+), and newline behavior for (q) is incorrect.  What's your
opinion?
Zach Riggle
Messages sorted by:
Reverse Date,
Date,
Thread,
Author