Re: Atomic operations in perl and Tk::IO

[hiseldl]

Update: <homer_simpson_voice>DOH! cloooosures</homer_simpson_voice>
-- I'll leave the text here so that no else who reads this makes the same mistake. :-)

In your call to component_window(), you assign $pcrh and $pid a la     my ($prch,$pid) = @_; now, I don't see a shift nor a pop nor any other @_ manipulations, which makes your call to tail_lines() incorrect:

sub component_window {
    my ($prch,$pid) = @_;
    ...
    my $tail = Tk::IO->new( 
       -linecommand => sub { tail_lines($prch, $_[0]) });
    ...
}

sub tail_lines {
    my ($proc,$text) = @_;
    ...
}

because $_[0] is the same as $prch when you call tail_lines($prch, $_[0]), and this seems to be not what you want.

You should be able to either pass the third element directly or get the third element in a var at the top of your sub (i.e.

sub component_window {
    my ($prch,$pid,$text)=@_;
    ...

and pass that...

    ...
    my $tail = Tk::IO->new( 
       -linecommand => sub { tail_lines($prch, $text) });
    ...
}

--
hiseldl
What time is it? It's Camel Time!

[rinceWind]

because $_[0] is the same as $prch when you call tail_lines($prch, $_[0]), and this seems to be not what you want.

Sorry, hiseldl, you are not correct. The call to tail_lines is in an anonymous sub, which is a closure, which is passed a line of text by the Tk::IO mechanism.

See Tk::IO for more details

[hiseldl]

I've been writing Tk scripts for over a year now, and I temporarily forgot about closures as callbacks, oops, sorry. I'll leave the post as it is to maintain continuity of the thread and as a reminder to other's using Tk to watch for closures in the callbacks. :-)

Cheers!

--
hiseldl
What time is it? It's Camel Time!