[bug-gawk] Issue #Q5: regexp: gsub: symbol `&' in replacement parameter

[Denis Shirokov]

Hi!

Another strange regexp issue:

Q5:
If character `&' is present in parameter `replacement' of gsub
function, then this character will be replaced by string that actually
matches to regular expression:

>gawk --version
GNU Awk 4.0.1
Copyright (C) 1989, 1991-2012 Free Software Foundation.

>type Q5.gwk

BEGIN{
        a="abcdef"
        print "`" a "'"
        gsub(/cd/,"C&D&Z",a)
        print "`" a "'" }

>gawk -f Q5.gwk
`abcdef'
`abCcdDcdZef'


------------------------------------------------------
(file Q5.gwk attached)

As you can see - string `cd' matched by regular expression /cd/ was
been replaced by `CcdDcdZ' - not by `C&D&Z'.
It looks like this symbol has some special meaning in replacement string.

If this is `by design' then what else characters have special
meanings? Is there a way to work in GSUB with parts of matched strings
like in gensub function?

Q5.gwk
Description: Binary data