Beefy Boxes and Bandwidth Generously Provided by pair Networks
Perl: the Markov chain saw
 
PerlMonks  

RFC: Adding variable name to warning "Use of uninitialized value ??? in ..."

by LanX (Chancellor)
on Apr 15, 2017 at 01:17 UTC ( #1187988=CUFP: print w/replies, xml ) Need Help??

Just a hack as prove of concept for Re^4: determine the variable causing the error: Use of uninitialized value

Can be improved in different ways ...

Due to limitations of B::Deparse it's only working if warning happens within a subroutine.

use strict; use warnings; use Data::Dump qw/pp dd/; use B::Deparse qw/coderef2text/; use PadWalker qw/peek_my/; my %warncache; my $result; BEGIN { $SIG{__WARN__} = sub { my ($err)=@_; # ignore other warnings unless ($err =~ m/^(Use of uninitialized value) (in (\w+) .* +)$/) { warn "$err"; return; } # ignore other undef vars return if $warncache{$err}++; my ($msg_start, $msg_end, $msg_type) =($1,$2,$3); #warn "* OrigWarn:\t $err"; my ($file,$line) = (caller(0))[1,2]; my $subname = (caller(1))[3]; my ($subref,$subline) = get_subline($subname,$file,$line,$er +r); #warn "LINE:<$subline>"; my $peek_sub = PadWalker::peek_sub ($subref); my $sep = { concatenation => '\\.', printf => ',', sprintf => ',', }->{$msg_type}; my $chomp = { sprintf => '\)', }->{$msg_type}; $subline =~ s/$chomp$// if $chomp; my @split = split /\s*$sep\s*/, $subline; #dd [$subline, @split, $peek_sub]; my @undefined; for my $snippet ( @split) { while ( my ($var,$ref) = each %$peek_sub){ $var =~ s/^\%/\$/; $var =~ s/^\@/\$/; my $match="\\$var"; if ( $snippet =~ /^(.*?)($match)(.*)$/ ) { my $new="$1\$ref$3"; #warn "match VAR <$match> in $snippet as $new"; next if defined eval($new); #warn "UNDEF $snippet"; push @undefined, $snippet; } } } #dd \@undefined; # build new warning my $plural = @undefined > 1 ? "s" :""; my $new_err = "${msg_start}$plural @undefined $msg_end"; warn # ". NewWarn:\t". "$new_err\n"; $result = { oldmsg => $err, newmsg => $new_err, vars => [@undefined], line => $subline, split => [@split], peek => $peek_sub, }; }; } sub get_subline { my ($name,$file,$line,$err) =@_; #dd \@_; my $subref = \&{$name}; my $subbody = B::Deparse->new('-q','-l','-x0')->coderef2text($subr +ef); my $start = "#line \Q$line\E \"\Q$file\E\"\n"; my $end = "\n(#line|})"; #warn $subbody; #dd "match:", $subbody =~ m/($start)/; my ($subline) = $subbody =~ m/$start\s+(.*?);$end/s; return ($subref, $subline); } #warn "Version $]"; my %hash=(a=>undef,b=>[]); my $h=\%hash; my @array=({a=>undef}); my $a=\@array; while (my $case = <DATA>) { chomp $case; next unless $case; next if $case =~ /^#/; my ($name,$var) = split /\s*:\s*/,$case; warn "*** TESTING".pp [$name,$var]; no warnings 'redefine'; my $out =""; open OUT,">",\$out; my @lines = ( # one undef var qq# print OUT "$name: $var"; #, qq# printf OUT '$name %s',$var; #, qq# print OUT sprintf '$name %s',$var; #, # multiple undef vars qq# print OUT "$name: $var $var" #, qq# printf OUT '$name %s %s',$var,$var; #, qq# print OUT sprintf '$name %s %s',$var,$var; #, ); for my $line (@lines) { my $code = <<"__CODE__"; sub tst { $line }; __CODE__ eval $code; if ($@) { warn "SKIPPING TEST $@ in \n<<<$code>>>"; next; } undef $result; tst(); die "$case", pp $result if $result and not @{$result->{vars}}; #warn pp $code,$result; } #last; } exit; __DATA__ hash_ref: $h->{a} hoa_ref: $h->{b}[0] hash: $hash{a} hoa: $hash{b}[0] array: $array[1] aoh: $array[0]{a} array_ref:$a->[1] aoh_ref: $a->[0]{a} #aoh_ref: $a->[$b]{'a b'}

Output:

Name "main::OUT" used only once: possible typo at /home/lanx/pm/warn_u +ndef.pl line 141. *** TESTING["hash_ref", "\$h->{a}"] at /home/lanx/pm/warn_undef.pl lin +e 138, <DATA> line 2. Use of uninitialized value $$h{'a'} in concatenation (.) or string at +(eval 10) line 2, <DATA> line 2. Use of uninitialized value $$h{'a'} in printf at (eval 12) line 2, <DA +TA> line 2. Use of uninitialized value $$h{'a'} in sprintf at (eval 14) line 2, <D +ATA> line 2. Use of uninitialized values $$h{'a'} $$h{'a'} in concatenation (.) or +string at (eval 16) line 2, <DATA> line 2. Use of uninitialized values $$h{'a'} $$h{'a'} in printf at (eval 19) l +ine 2, <DATA> line 2. Use of uninitialized values $$h{'a'} $$h{'a'} in sprintf at (eval 22) +line 2, <DATA> line 2. *** TESTING["hoa_ref", "\$h->{b}[0]"] at /home/lanx/pm/warn_undef.pl l +ine 138, <DATA> line 3. Use of uninitialized value $$h{'b'}[0] in concatenation (.) or string +at (eval 25) line 2, <DATA> line 3. Use of uninitialized value $$h{'b'}[0] in printf at (eval 27) line 2, +<DATA> line 3. Use of uninitialized value $$h{'b'}[0] in sprintf at (eval 29) line 2, + <DATA> line 3. Use of uninitialized values $$h{'b'}[0] $$h{'b'}[0] in concatenation ( +.) or string at (eval 31) line 2, <DATA> line 3. Use of uninitialized values $$h{'b'}[0] $$h{'b'}[0] in printf at (eval + 34) line 2, <DATA> line 3. Use of uninitialized values $$h{'b'}[0] $$h{'b'}[0] in sprintf at (eva +l 37) line 2, <DATA> line 3. *** TESTING["hash", "\$hash{a}"] at /home/lanx/pm/warn_undef.pl line 1 +38, <DATA> line 5. Use of uninitialized value $hash{"a"} in concatenation (.) or string a +t (eval 40) line 2, <DATA> line 5. Use of uninitialized value $hash{"a"} in printf at (eval 41) line 2, < +DATA> line 5. Use of uninitialized value $hash{"a"} in sprintf at (eval 42) line 2, +<DATA> line 5. Use of uninitialized value $hash{"a"} in concatenation (.) or string a +t (eval 43) line 2, <DATA> line 5. Use of uninitialized value $hash{"a"} in concatenation (.) or string a +t (eval 43) line 2, <DATA> line 5. Use of uninitialized value $hash{"a"} in printf at (eval 44) line 2, < +DATA> line 5. Use of uninitialized value $hash{"a"} in printf at (eval 44) line 2, < +DATA> line 5. Use of uninitialized value $hash{"a"} in sprintf at (eval 45) line 2, +<DATA> line 5. Use of uninitialized value $hash{"a"} in sprintf at (eval 45) line 2, +<DATA> line 5. *** TESTING["hoa", "\$hash{b}[0]"] at /home/lanx/pm/warn_undef.pl line + 138, <DATA> line 6. Use of uninitialized value $hash{'b'}[0] in concatenation (.) or strin +g at (eval 46) line 2, <DATA> line 6. Use of uninitialized value $hash{'b'}[0] in printf at (eval 48) line 2 +, <DATA> line 6. Use of uninitialized value $hash{'b'}[0] in sprintf at (eval 50) line +2, <DATA> line 6. Use of uninitialized values $hash{'b'}[0] $hash{'b'}[0] in concatenati +on (.) or string at (eval 52) line 2, <DATA> line 6. Use of uninitialized values $hash{'b'}[0] $hash{'b'}[0] in printf at ( +eval 55) line 2, <DATA> line 6. Use of uninitialized values $hash{'b'}[0] $hash{'b'}[0] in sprintf at +(eval 58) line 2, <DATA> line 6. *** TESTING["array", "\$array[1]"] at /home/lanx/pm/warn_undef.pl line + 138, <DATA> line 8. Use of uninitialized value $array[1] in concatenation (.) or string at + (eval 61) line 2, <DATA> line 8. Use of uninitialized value $array[1] in printf at (eval 62) line 2, <D +ATA> line 8. Use of uninitialized value $array[1] in sprintf at (eval 64) line 2, < +DATA> line 8. Use of uninitialized value $array[1] in concatenation (.) or string at + (eval 65) line 2, <DATA> line 8. Use of uninitialized values $array[1] $array[1] in concatenation (.) o +r string at (eval 65) line 2, <DATA> line 8. Use of uninitialized values $array[1] $array[1] in printf at (eval 68) + line 2, <DATA> line 8. Use of uninitialized values $array[1] $array[1] in sprintf at (eval 71 +) line 2, <DATA> line 8. *** TESTING["aoh", "\$array[0]{a}"] at /home/lanx/pm/warn_undef.pl lin +e 138, <DATA> line 9. Use of uninitialized value $array[0]{'a'} in concatenation (.) or stri +ng at (eval 74) line 2, <DATA> line 9. Use of uninitialized value $array[0]{'a'} in printf at (eval 76) line +2, <DATA> line 9. Use of uninitialized value $array[0]{'a'} in sprintf at (eval 78) line + 2, <DATA> line 9. Use of uninitialized values $array[0]{'a'} $array[0]{'a'} in concatena +tion (.) or string at (eval 80) line 2, <DATA> line 9. Use of uninitialized values $array[0]{'a'} $array[0]{'a'} in printf at + (eval 83) line 2, <DATA> line 9. Use of uninitialized values $array[0]{'a'} $array[0]{'a'} in sprintf a +t (eval 86) line 2, <DATA> line 9. *** TESTING["array_ref", "\$a->[1]"] at /home/lanx/pm/warn_undef.pl li +ne 138, <DATA> line 11. Use of uninitialized value $$a[1] in concatenation (.) or string at (e +val 89) line 2, <DATA> line 11. Use of uninitialized value $$a[1] in printf at (eval 91) line 2, <DATA +> line 11. Use of uninitialized value $$a[1] in sprintf at (eval 93) line 2, <DAT +A> line 11. Use of uninitialized values $$a[1] $$a[1] in concatenation (.) or stri +ng at (eval 95) line 2, <DATA> line 11. Use of uninitialized values $$a[1] $$a[1] in printf at (eval 98) line +2, <DATA> line 11. Use of uninitialized values $$a[1] $$a[1] in sprintf at (eval 101) lin +e 2, <DATA> line 11. *** TESTING["aoh_ref", "\$a->[0]{a}"] at /home/lanx/pm/warn_undef.pl l +ine 138, <DATA> line 12. Use of uninitialized value $$a[0]{'a'} in concatenation (.) or string +at (eval 104) line 2, <DATA> line 12. Use of uninitialized value $$a[0]{'a'} in printf at (eval 106) line 2, + <DATA> line 12. Use of uninitialized value $$a[0]{'a'} in sprintf at (eval 108) line 2 +, <DATA> line 12. Use of uninitialized values $$a[0]{'a'} $$a[0]{'a'} in concatenation ( +.) or string at (eval 110) line 2, <DATA> line 12. Use of uninitialized values $$a[0]{'a'} $$a[0]{'a'} in printf at (eval + 113) line 2, <DATA> line 12. Use of uninitialized values $$a[0]{'a'} $$a[0]{'a'} in sprintf at (eva +l 116) line 2, <DATA> line 12.

Cheers Rolf
(addicted to the Perl Programming Language and ☆☆☆☆ :)
Je suis Charlie!

Replies are listed 'Best First'.
Re: RFC: Adding variable name to warning "Use of uninitialized value ??? in ..."
by dave_the_m (Prior) on Apr 15, 2017 at 14:04 UTC
    My implementation in the perl core was a compromise between complexity, performance and convenience. It's currently about 750 lines of src code. It could certainly be expanded to extract the variable name under more circumstances, but I had to draw a line somewhere.

    Note also that it's quite hard to avoid false positives. For example, for an empty array @a, in $a[0] + $a[1] at the time of execution of the addition, both args are the same special address &PL_sv_undef, so its not possible to determine whether a particular undef value came from the LHS or the RHS.

    Dave.

      Hi Dave

      Thanks for the feedback :)

      > My implementation in the perl core was a compromise between complexity, performance and convenience. 

      I know, and this is in no way a criticism of your implementation! I'm aware of the complexity.

      I want to open an opportunity to better inspect problems in pure Perl.

      At work we are using many here docs and help in faster identifying root problems would be appreciated.

      See it as a potential add on module like diagnostics and not as a feature request for core.

      > It's currently about 750 lines of src code. It could certainly be expanded to extract the variable name under more circumstances, but I had to draw a line somewhere.

      That's why I'm delegating the hard part to B::Deparse

      > Note also that it's quite hard to avoid false positives

      That's why I'm reporting all potential culprits with undef values in the first warning and ignoring all other warnings from the same line.

      I don't want to risk that a true positive gets lost because of a false positive.

      It's up to the user to find out which are real.

      Cheers Rolf
      (addicted to the Perl Programming Language and ☆☆☆☆ :)
      Je suis Charlie!

Re: RFC: Adding variable name to warning "Use of uninitialized value ??? in ..."
by LanX (Chancellor) on Apr 15, 2017 at 14:58 UTC
    Brainstorm:

    To Do
    • cover more cases , not only concatenation , printf and sprintf , like addition and join . (...where do I find a complete list?)
    • test false positives
    • make SIG handler collaborating with older hooks, like for diagnostics
    • test package variables
    • test global variables ,symbols, barewords
    • cover slices like @a[0..2]
    • test with different Perl versions in perlbrew
    • handle call from filescope and eval
    * 17.4.2017
    • handle multiple statements in same line
    Nice to Have
    • report original source line too
    • patch B::Deparse instead of parsing string output
    • add hooks for further customization by user
    • handle tied variables

    Cheers Rolf
    (addicted to the Perl Programming Language and ☆☆☆☆ :)
    Je suis Charlie!

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: CUFP [id://1187988]
Approved by Athanasius
Front-paged by kcott
help
Chatterbox?
and the leaves swirl about...

How do I use this? | Other CB clients
Other Users?
Others cooling their heels in the Monastery: (3)
As of 2017-07-28 01:24 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?
    I came, I saw, I ...
























    Results (424 votes). Check out past polls.