Re: interpreting a strongly typed regexp constant variable as literal regular expression?

[arnold]

Hi.

I've been offline for a few days.

As explained in c.l.a. and here, you've misunderstood how strongly
typed regexp constants work.  They do not indirect to pull in values
of other variables.  (I have to admit, I don't understand how you
came to that conclusion from the gawk doc. I rechecked it this morning
and it doesn't say anything like that.)

Your suggestions are interesting, but involve way too much implementation
complexity for the value gained.  There are other features in gawk
already that will let you do similar things:

        a = @/^a/
        b = "a"
        foo ~ SYMTAB[b]

Also, as pointed out, the values in case labels must be constants,
they can't be variables, so to do the kind of checks you're looking
for you must use a sequence of if-else statements.

Thanks,

Arnold

Craig <sargonemail@gmail.com> wrote:

> Ok,
>
> Suggested awk feature/addition:
>
> So, currently, in order to have @/val/ in example 2 be used/evaluated
> as a regular expression statement like /^a/ in example 1; the switch
> block containing the strongly typed regular expression has to be dumped
> into a temporary awk script.  This temporary awk script is then
> called/executed to produced the desired result.
>
> I'd be nice to just be able to just add @ before the switch keyword,
> ( aka @switch) to indicate to awk that at runtime the strongly typed
> regular expression(s) in switch block should be updated with constants
> before awk executes the @switch block.
>
> A general approach to getting around having to generate/shell out to
> a temporary awk script in order to replace the typed regular expressions
> in a switch case with constants would be to add a way to indicate to awk
> that at run-time, there are strongly typed regular expressions in
> the a switch block that need to be evaluated before the switch block
> is run.
>
> 1) switch block specific method
> Add @switch to indicate need to fill-in/evaluate the the strongly typed
> regular expression(s) in switch case lines with constants before switch
> block is entered. Only strongly typed constants that typeof() returned
> regexp would be changed as if the regexpr had a / at the start & end.
> aka "case @foo :" foo changed to a string with / before and after,
> so  "case /foo_string/"
>
> 2) Genaral method, not specific to switch block
> A @{ ... @} block around the switch block would tell awk to first
> re-evaluate the strongly typed regular expression constant variable(s)
> contained within the @{ ... @} block before executing the awk code
> within the @{ ... @} block.  The @{ ... @} block is a general approach
> to make things easier to gawk & understand what's going on.
> It would allow any strongly typed regular expressions in a awk block
> statement to be evaluated at-run before the block within the @{...@}
> is evaluated.
> Note:  This would not parse/evaluate sub-blocks within the @{ ... #} block
> example uses:
>
>  * "case @/foo/ :" would be interpreted as if "case /foo/ : was the line
>
>  * @func_name(val) would result in the value returned by the function
>     specified by @func_name
>  * function() could be called too if results of function() assigned to a
>    regexpr variable & function returned a value that could be used in switch
>    case line
>
> Thanks,
> Craig
>
>
>
>
> On Sat, May 30, 2020 at 11:28 AM Craig <sargonemail@gmail.com> wrote:
>
> > Ok,
> >
> > This was originally asked in comp.lang.awk and was asked to be
> > sent to the bug list.
> > ( https://groups.google.com/forum/#!topic/comp.lang.awk/DjvDi3E1-XU )
> >
> > Ok, as of awk 4.2, regular expression constants are a first-class
> > data type using the syntax @/foo/.
> > Per  Edition 5.1 of "GAWK: Effective AWK Programming: A User’s
> > Guide for GNU Awk", Section 6.1.2.2 "Strongly Typed Regexp Constants",
> > /@var/ should be useable in the case part of a switch statement.
> > Basic, working script examples discussed provided at the end of
> > this message.
> >
> > What is the syntax to get 'case @/var/ :' in script 2
> > to be interpreted as an actual case constant regular expression
> > and not a string?
> > e.g In script 2, the line 'case @/var/ :' needs to be evaluated as
> > if the line was script 1's line, 'case /var/ :' and not as if the
> > line was script 3's line, 'case "@/var/".
> >
> > For the given examples,  Script 1 generates the desired output.
> > The desired output for script 2 should be the same the output
> > generated by script 1.
> > As provided, in script 2, the 'Yes" is printed as the 4th record
> > instead of 2nd record because @/var/ isn't interpreted as an
> > actual regular expression, but as a string (as in script 3's
> > example).
> >
> > typeof() shows the script 2 variable as being type regexp.
> >
> > Thanks,
> > Craig
> >
> > ===== script 1
> > awk '{
> > switch($0) {
> >    case /^a/ :
> >       print "Yes"
> >    break
> >    default :
> >       print "NR: " NR
> >    break
> >    }
> > }' << END
> > g
> > a
> > c
> > @/a/
> > END
> >
> > ==== script 1 output:
> > NR: 1
> > Yes
> > NR: 3
> > NR: 4
> >
> > ===== script 2
> > awk '{ var = @/^a/
> > switch($0) {
> >    case @/var/ :
> >       print "Yes"
> >    break
> >    default :
> >       print "NR: " NR
> >    break
> >    }
> > }' << END
> > g
> > a
> > c
> > @/var/
> > END
> >
> > ==== script 2 output
> > NR: 1
> > NR: 2
> > NR: 3
> > Yes
> >
> > ===== script 3
> > awk '{ var = @/^a/
> > switch($0) {
> >    case "@/var/" :
> >       print "Yes"
> >    break
> >    default :
> >       print "NR: " NR
> >    break
> >    }
> > }' << END
> > g
> > a
> > c
> > @/var/
> > END
> >
> > ==== script 3 output
> > NR: 1
> > NR: 2
> > NR: 3
> > Yes
> >
> >