Why doesn't this work?

perl-diddler has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Why doesn't this work? by LanX (Saint) on Apr 12, 2013 at 21:42 UTC
try `$/="\0"` unquoted `\0` is a reference to a literal number (i.e. read-only scalar) ... `DB<135> $/=\0 => \0 DB<136> ref $/ => "SCALAR" DB<137> ${$/} => 0` [download] ... and not the escape sequence for NUL byte! `DB<146> $/="\0" => "\0" DB<147> ord $/ => 0` [download] Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l] [select]
Re^2: Why doesn't this work? by perl-diddler (Chaplain) on Apr 12, 2013 at 21:57 UTC
Bingo. That was it! doh! <palmslaps self>	[reply]
Re^3: Why doesn't this work? by LanX (Saint) on Apr 12, 2013 at 22:35 UTC
maybe shorter? `DB<178> open $fh,"<",\"1\0 2\0 3" DB<179> print scalar do {local $/="\0"; () = <$fh> } 3` [download] update as one liner `perl -e '$/="\0"; print $a=()=<>'` [download] Cheers Rolf ( addicted to the Perl Programming Language)	[reply] [d/l] [select]
Re^4: Why doesn't this work? by choroba (Cardinal) on Apr 12, 2013 at 22:39 UTC
Re: Why doesn't this work? by ww (Archbishop) on Apr 12, 2013 at 21:59 UTC
You might want to read about `$/` in `perldoc perlvar`. It would also be well to recognize that settng "$/=\000" is far from the same thing as `$/=undef` or `$/="\000"`. Not an issue in your one-liners, but you should probably use any modified `$/` ( or `local $/=...` ) in as narrow a scope as possible when you change it, or make sure you reset it in any other instance (other'n your one-liners). If you didn't program your executable by toggling in binary, it wasn't really programming!	[reply] [d/l] [select]
Re^2: Why doesn't this work? by perl-diddler (Chaplain) on Apr 12, 2013 at 22:08 UTC
I forgot the 'my' in the braces, but know the point you were making -- i.e. that's why it was in braces. I wasn't trying to use slurp mode & $/=undef in the one example, -- was trying to recognize "\000" (NUL) as the line delimiter (in this case a list of NUL-terminated filenames) and use that to read the whole thing into an array and then seeing how many array members I had.	[reply]
Re^3: Why doesn't this work? by perl-diddler (Chaplain) on Apr 12, 2013 at 22:18 UTC
Make that 'my' a 'local'... geez... trying to respond too fast. Just to be sure I checked my code and only find one place where I used slurp mode that wasn't in a 1-liner, and I used local there. Glad you reminded me -- as I'd started to get it in my head that braces or my were needed... Looks like most common usage for me is shell-1-liners where Shell can't deal with NUL terminated lines, so need perl to do that.	[reply]
Re: Why doesn't this work? by Corion (Patriarch) on Apr 12, 2013 at 21:35 UTC
For me, your first example works: `C:\Dokumente und Einstellungen\corion>perl -MData::Dumper -e "{$_=[];l +ocal $/=qq(\n); @$_=<>};warn Dumper $_; my $count=@$_; print $count, +' lines';" \boot.ini $VAR1 = [ '[boot loader] ', ... ]; 5 lines` [download] while the second one does always output "1 lines". This is the special behaviour of split. To make split produce more results than needed on the left hand side, you need to give it a LIMIT. Also, you will want to call split in list context using this operator `=()=`: `> perl -e "{$/=undef, $_=<>} my $count=()=split /\n/,$_,-1; print $cou +nt" \boot.ini 6` [download]	[reply] [d/l] [select]
Re^2: Why doesn't this work? by perl-diddler (Chaplain) on Apr 12, 2013 at 21:55 UTC
Huh?... but the 2nd one worked. What did you do to your perl that it doesn't? I'm using this split: `Splits the string EXPR into a list of strings and returns the list in list context, or the size of the list in scalar context. Size of list in scalar context. && re: limit: If LIMIT is omitted (or, equivalently, zero), then it i +s usually treated as if it were instead negative but with + the exception that trailing empty fields are stripped (empt +y leading fields are always preserved);` [download] What version of split are you using?	[reply] [d/l]
Re^3: Why doesn't this work? by Corion (Patriarch) on Apr 12, 2013 at 22:05 UTC
Ah, sorry - I don't know what I did wrong while testing this. The documentation is correct and neither the limit nor the forced list context are needed: `> perl -e "{$/=undef, $_=<>} my $count=split /\n/; print $count" \boot +.ini` [download] works just as well, and just as documented.	[reply] [d/l]
Re: Why doesn't this work? by hdb (Monsignor) on Apr 12, 2013 at 21:33 UTC
This works for me: `time perl -e '$_=[];@$_=<>; my $count=@$_;print $count, " lines\n";' < + somefile.txt time perl -e '{$/=undef, $_=<>} my $count=split /\n/,$_; print $count, + " lines\n";' < somefile.txt` [download] Your second version does not when splitting on \000.	[reply] [d/l]
Re^2: Why doesn't this work? by perl-diddler (Chaplain) on Apr 12, 2013 at 22:01 UTC
The examples you list don't split on \000. They are not the examples from my code. (i.e. 2nd example was..) `time perl -e '{$/=undef, $_=<>} my $count=split /\000/,$_; print $coun +t, " lines\n";' </tmp/16` [download] Wasn't it? Or am I REALLY confused?	[reply] [d/l]
Re^3: Why doesn't this work? by hdb (Monsignor) on Apr 13, 2013 at 06:54 UTC
I cound not find even a single test file that would be more than one line under splitting on \000 so I was resorting to replacing the delimiter. So what I stated was correct but possibly irrelevant... ;)	[reply]

update