http://qs321.pair.com?node_id=1225867


in reply to Re^7: B::Xref buggy?
in thread B::Xref buggy?

> Concise doesn't track line numbers,

I'm confused ...did you look at the output I provided?

This looks pretty much like the lexical order in the source which is needed to track variables and subs.

and the line numbers are available: 

# 6:     print "@row";                                              #:
+line 6
-                    <;> ex-nextstate(main 7 tst_b_xref2.pl:6) v:*,&,x
+*,x&,x$,$

[download]

The -src option dumps the original source line just before every nextstate OP and the line number is available in tst_b_xref2.pl:6

Could you please provide an example code where this behaviour is broken in the B::Concise output and information is lost?

Probably I'm unaware of some Perl temporal magic ? ? ?

Cheers Rolf
(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery FootballPerl is like chess, only without the dice

Replies are listed 'Best First'.
Re^9: B::Xref buggy?
by dave_the_m (Monsignor) on Nov 15, 2018 at 17:25 UTC
    This looks pretty much like the lexical order in the source which is needed to track variables and subs
    I think that is the heart of your misapprehension. Concise doesn't output ops in source lexical order, apart from happenstance. As a trivial counterexample, this source: 
    unless ($var1) {
    $var2++;
}
else {
    $var3++;
}

    [download]
    gives: 
    perl -MO=Concise,-exec /tmp/p | grep var

3  <#> gvsv[*var1] s
7      <#> gvsv[*var3] s
d  <#> gvsv[*var2] s

    [download]
    Note that var3 comes before var2.

    Dave.

[reply]
[d/l]
[select]
      Thanks, Dave.

      I was aware that OPs do kind of gotos to next-OPs (the ->label at the end of each line).

      But it didn't seem logical that the -src option could be offered if B::Concise wasn't respecting the order.*

      The solution is that the source lines they are not in order, but the line numbers are still reliable:

      D:\tmp>type tst_b_xref3.pl
unless ($var1) {
    $var2++;
}
else {
    $var3++;
}

      [download]
      (Win Activestate 5.24) 
      D:\tmp>perl -MO=Concise,-src tst_b_xref3.pl|find "var"
tst_b_xref3.pl syntax OK
# 1: unless ($var1) {
3              <#> gvsv[*var1] s ->4
# 5:     $var3++;
7                    <#> gvsv[*var3] s ->8
# 2:     $var2++;
d                    <#> gvsv[*var2] s ->e

      [download]

      So strictly speaking it's still possible to use B::Concise to emulate B::Xref on line number level.

      Cheers Rolf
      (addicted to the Perl Programming Language :)
      Wikisyntax for the Monastery FootballPerl is like chess, only without the dice

      *) Interestingly this also depends on the Perl version. Here the vars appear in order: (Ubuntu 5.18)

      $ perl -MO=Concise,-src tst_b_xref3.pl|grep var
tst_b_xref3.pl syntax OK
# 1: unless ($var1) {
3                 <#> gvsv[*var1] s ->4
# 2:     $var2++;
8                    <#> gvsv[*var2] s ->9
# 5:     $var3++;
e                    <#> gvsv[*var3] s ->f

      [download]

      update

      and that's the reason why, 5.24 is optimizing the negation away by swapping the branches: 

      D:\tmp>perl -MO=Deparse tst_b_xref3.pl
tst_b_xref3.pl syntax OK
if ($var1) {
    ++$var3;
}
else {
    ++$var2;
}

      [download]
[reply]
[d/l]
[select]
        The solution is that the source lines they are not in order, but the line numbers are still reliable:
        (Cough) 
        $ cat /tmp/p
for (
    $var1++;
    $var2++;
    $var3++
) {
    $var4++;
}


$ perl -MO=Concise,-src /tmp/p | grep var
/tmp/p syntax OK
3           <#> gvsv[*var1] s ->4
f                    <#> gvsv[*var2] s ->g
# 6:     $var4++;
9                          <#> gvsv[*var4] s ->a
c                       <#> gvsv[*var3] s ->d

        [download]

        Dave.

[reply]
[d/l]

In Section Seekers of Perl Wisdom