Re: is the name of the file being interpreted available?

[Ed Morton]

Thanks Andy, I figured it had to be SOMEWHERE. Regarding the subtleties 
you mention - just like $0 in shell is the script being interpreted even 
if the code within it calls functions defined elsewhere, I'm just 
looking for that same functionality in awk so my  real complexity is:

a) searching `PROCINFO["argv"]` for a string that starts with `-f`, in 
which case the rest of the string is the script file name, or a string 
that IS `-f`, in which case the next index contains the script file 
name, and
b) figuring out how to handle the case where 2 script file names are 
passed as arguments.

"a" is easy:

   BEGIN {
        for ( i=1; i in PROCINFO["argv"]; i++ ) {
            if ( !isarray(i) ) {
                arg = PROCINFO["argv"][i]
                if ( arg == "-f" ) {
                    scriptFile = PROCINFO["argv"][i+1]
                }
                else if ( sub(/^-f/,"",arg) ) {
                    scriptFile = arg
                }
            }
        }
        printf "scriptFile: <%s>\n", scriptFile
   }

but "b" isn't so obvious.

I don't really have a use for this right now, I'm just wondering how I'd 
handle it if I ever wanted to do something like write a script that 
behaves differently based on the name of the file or directory that 
contains it.

    Ed.

On 6/19/2022 11:09 AM, Andrew J. Schorr wrote:
> On Sun, Jun 19, 2022 at 10:28:32AM -0500, Ed Morton wrote:
> > I just realised that while shell has $0 to hold the name of the file
> > containing the script being interpreted, I don't know if that same
> > information is available in any builtin variable in awk. Obviously I
> > could write:
> >
> >      awk -v script='/foo/bar/script.awk' -f '/foo/bar/script.awk' input
> >
> > or similar but is there any way to get that '/foo/bar/script.awk'
> > info inside the awk script without any manual intervention so I can
> > just write:
> >
> >      BEGIN { "we are running", foo }
> >
> > for some value of `foo` inside the script to output:
> >
> >      we are running /foo/bar/script.awk
> I don't think that's currently exposed, although you can see the full
> commandline in the PROCINFO[argv] array.
>
> https://www.gnu.org/software/gawk/manual/html_node/Auto_002dset.html
>
> 'The PROCINFO["argv"] array contains all of the command-line arguments (after
> glob expansion and redirection processing on platforms where that must be done
> manually by the program) with subscripts ranging from 0 through argc - 1. For
> example, PROCINFO["argv"][0] will contain the name by which gawk was invoked. '
>
> You can test for yourself by dumping out the contents of the SYMTAB array to
> see everything that is defined.  For example:
>
> bash-4.2$ cat /tmp/test.gawk
> function printvar(what, val,   i) {
>     if (isarray(val)) {
>        for (i in val)
>           printvar((what "[" i "]"), val[i])
>     }
>     else
>        printf "%s = [%s]\n", what, val
> }
>
> BEGIN {
>     for (i in SYMTAB) {
>        printvar(i, SYMTAB[i])
>     }
> }
>
> bash-4.2$ gawk -f /tmp/test.gawk | grep test.gawk
> PROCINFO[argv][2] = [/tmp/test.gawk]
>
> There are some subtleties here. When you say "the script being interpreted",
> how would you handle situations where awk source file A contains function X
> that calls a function Y located in source file B? If one is executing function
> Y called from X, is the current source file B or A? And so on.
>
> Regards,
> Andy