Re: [bug-gawk] escaping open curly bracket produces unexpected warning with --posix

[arnold]

Hi.

Ed Morton <address@hidden> wrote:

> I just came across this in gawk 5.0.0 (note "4"):
>
>     1) $ echo 'foo{bar' | awk '/{/'
>     foo{bar
>
>     2) $ echo 'foo{bar' | awk --posix '/{/'
>     awk: cmd. line:1: error: Invalid preceding regular expression: /{/
>
>     3) $ echo 'foo{bar' | awk '/\{/'
>     foo{bar
>
>     4) $ echo 'foo{bar' | awk --posix '/\{/'
>     awk: cmd. line:1: warning: regexp escape sequence `\{' is not a
>     known regexp operator
>     foo{bar
>
> I understand 1, 2, and 3 but I don't understand why I get the warning 
> from 4 above. I expected `\{` to simply be treated as a literal `{` with 
> no warnings just like other RE metachars are:
>
>     $ echo 'foo[bar' | awk --posix '/\[/'
>     foo[bar
>
>     $ echo 'foo+bar' | awk --posix '/\+/'
>     foo+bar
>
> Instead I had to put it inside square brackets to get it treated as 
> literal with no warnings:
>
>     $ echo 'foo{bar' | awk --posix '/[{]/'
>     foo{bar
>
> as no amount of additional preceding backslashes would make it "work" 
> without a warning:
>
>     $ echo 'foo{bar' | awk --posix '/\{/'
>     awk: cmd. line:1: warning: regexp escape sequence `\{' is not a
>     known regexp operator
>     foo{bar
>     $ echo 'foo{bar' | awk --posix '/\\{/'
>     $ echo 'foo{bar' | awk --posix '/\\\{/'
>     awk: cmd. line:1: warning: regexp escape sequence `\{' is not a
>     known regexp operator
>     $ echo 'foo{bar' | awk --posix '/\\\\{/'
>     $
>
> Regards,
>
>      Ed.

The patch below fixes this.

Thanks for the report.

Arnold
-----------------------------------------------------------------
diff --git a/re.c b/re.c
index 174f54f8..210be068 100644
--- a/re.c
+++ b/re.c
@@ -107,7 +107,9 @@ make_regexp(const char *s, size_t len, bool ignorecase, 
bool dfa, bool canfatal)
                }
 
                const char *ok_to_escape;
-               if (do_traditional)
+               if (do_posix)
+                       ok_to_escape = "{}()|*+?.^$\\[]/-";
+               else if (do_traditional)
                        ok_to_escape = "()|*+?.^$\\[]/-";
                else
                        ok_to_escape = "<>`'BywWsS{}()|*+?.^$\\[]/-";