Re: [bug-gawk] Is \x24 the literal dollar character?

[arnold]

Hi.

Thanks for the report. I will investigate.  I'm on vacation, so it
may take a while before I can get back to you.

Arnold

"Kozics Peter (FM)" <address@hidden> wrote:

> Dear,
>
>
> (1)
> this matches:
> $ gawk '/\$Id: .*\$/ {print}' <<< '$Id: rcsid$'
> $Id: rcsid$
>
> (2)
> I expected that this would match as well, but it didn't:
> $ gawk '/\x24Id: .*\x24/ {print}' <<< '$Id: rcsid$'
>
> The expectation was based on gawk manual section 3.2: \x24 should be
> the literal dollar character, not the dollar metacharacter.
>
> (3)
> Now, let's go on, this does not match either:
> $ gawk '/\\x24Id: .*\\x24/ {print}' <<< '$Id: rcsid$'
>
> (4)
> And this one still not:
> $ gawk '/\\\x24Id: .*\\\x24/ {print}' <<< '$Id: rcsid$'
>
> (5)
> At long last, this matches again:
> $ gawk '/\x5c\x24Id: .*\x5c\x24/ {print}' <<< '$Id: rcsid$'
> $Id: rcsid$
>
> which looks to me awkward and quite counterintuitive.  
>
> -------------
> The problem with (1) is that when the regexp is in a file under RCS
> control, RCS will destroy the regexp upon checkout by performing a
> keyword substitution. So the straightforward and seemingly manual-
> compliant solution would be (2), which is unfortunately not.
>
> I wonder if I found a gawk bug or a flaw in the regexp / literal / meta
> concept or a vague place in the gawk manual. Or just misunderstood
> something?
>
> -------------
> OS:
> $ uname -a
> Linux gygv 5.2.18-200.fc30.x86_64 #1 SMP Tue Oct 1 13:14:07 UTC 2019
> x86_64 x86_64 x86_64 GNU/Linux
>
> gawk:
> $ gawk --version
> GNU Awk 4.2.1, API: 2.0 (GNU MPFR 3.1.6-p2, GNU MP 6.1.2)
> Copyright (C) 1989, 1991-2018 Free Software Foundation.
>
> gawk manual:
> This is Edition 4.2 of GAWK: Effective AWK Programming
>
>
> yours
> KP
>