G'day Rolf,

That feature was introduced in v5.34: "perl5340delta: Blanks freely allowed within but adjacent to curly braces".

My tests:

$ alias perle
alias perle='perl -Mstrict -Mwarnings -Mautodie=:all -MCarp::Always -E
+'

$ perlbrew switch perl-5.32.0

$ perle 'say $]; say join ",",map {ord} split //, $_ for "\x{ 41 }", "
+\x{41}", "\x{263A}", "\x{ 263A }"; say for "\x{41}", "\x{ 41 }"'
Non-hex character ' ' terminates \x early.  Resolved as "\x{00}" at -e
+ line 1.
Non-hex character ' ' terminates \x early.  Resolved as "\x{00}" at -e
+ line 1.
Non-hex character ' ' terminates \x early.  Resolved as "\x{00}" at -e
+ line 1.
5.032000
0
65
9786
0
A


$ perlbrew switch perl-5.34.0

$ perle 'say $]; say join ",",map {ord} split //, $_ for "\x{ 41 }", "
+\x{41}", "\x{263A}", "\x{ 263A }"; say for "\x{41}", "\x{ 41 }"'
5.034000
65
65
9786
9786
A
A
[download]

It's not mentioned in perlop, so a doco omission but not a bug in Perl itself. Perhaps raise a doco bug report.

Hi Ken,

> hat feature was introduced in v5.34

Thanks, I'm feeling stupid now.

I'm having the newest available Strawberry Perl 5.32 installed and didn't think about the perldoc being even newer.

> It's not mentioned in perlop, so a doco omission

I searched for "version" in perlop and there are many mentions of differences to older Perl.

Not sure if this is always done, but I certainly would have preferred seeing a "(since v5.34)" somewhere.

And I found the description rather cryptic too.

• but shows optional blanks inside and adjoining the braces
shouldn't it rather be "ignores" or "tolerates blanks"?

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

As usual in these cases, Syntax::Construct qw{ blanks-in-curlies }.
map{substr$_->[0],$_->[1]||0,1}[\*||{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^*ARGV,3]

shouldn't it rather be "ignores" or "tolerates" whitespace?

My suggested phrasing would be on "but allows° optional blanks between the braces and the hex value"

(And whether I said "blanks" or "spaces" or "horizontal whitespace" or "whitespace" would probably depend on whether tests show that it allows only space, or any horizontal space, or any horizontal or vertical space. Like you, I only have Strawberry so limited to 5.32 and cannot test the feature.)

update: °: Rereading the original, I think they intended that sentence to mean the equivalent of "This second example of the syntax means the same as the example above, but shows the optional blanks that are allowed..."

'I searched for "version" ...'

On the basis that version notes often look like "introduced in 5.nn", I searched for "34": 8 instances found; none relate to this issue.

'... but shows ... shouldn't it rather be "ignores" or "tolerates" whitespace?'

I believe that was intended to contrast with the previous line (\x{263A} vs. \x{ 263A }). Depending on device, browser and font, the difference between \x{263A} and \x{ 263A } may not be all that obvious.

Regardless, a ninth note (e.g. [9] Introduced in v5.34.0.) would add clarity.

— Ken

Na Rolf, I had this same failure but didn't post it, because I had *a lot* of failures and first assume that I'm doing something backwards. I can get there partially now, but I can't clear the warning:

$ perl -de1
...
  DB<1> $str = "\x{ 41 }"                                             
+          

  DB<2> p $str                                                        
+          

  DB<3> $str1 = "\x{41}"                                              
+          

  DB<5> p $str1                                                       
+          
A
  DB<6> $str2 = "\x{263A}"                                            
+          

  DB<7> p $str2                                                       
+          
Wide character in print at (eval 30)[/usr/share/perl/5.28/perl5db.pl:7
+38] line 2.
...
&#9786;
  DB<8> binmode STDOUT, ':utf8';                                      
+          

  DB<9> p $str2   
Wide character in print at (eval 30)[/usr/share/perl/5.28/perl5db.pl:7
+38] line 2.                                                          
+    
...
&#9786;
[download]

I tried everything I could throw at it. Am I on STDOUT whilst invoking perl -de1 ? What am I missing?

> What am I missing?

the interactive debugger is redirecting STDOUT with every eval. You need to do it inline

  DB<3> binmode STDOUT, ':utf8'; print "\x{263A}"
\u263A
[download]

of course your terminal° must also be capable to display smileys :)

I think if you want to change STDOUT globally for the debugger, some of the unicode flags in perlrun should do. (untested)

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

°) in my case emacs inside the console

"I think if you want to change STDOUT globally for the debugger, some of the unicode flags in perlrun should do. (untested)"

I tried all of these:

$ perl -C -de1
...

$ perl -de1
...
  DB<1> use open OUT => qw{:encoding(UTF-8) :std}
...


$ perl -Mopen=OUT,:encoding\(UTF-8\),:std -de1
...
[download]

All gave the "Wide character ..." warning. All still printed "☺". There's probably other avenues of investigation for a global change: I didn't try anything further.

— Ken

the interactive debugger is redirecting STDOUT with every eval. You need to do it inline

Aha...

  DB<1> binmode STDOUT, ':utf8'; print "\x{263A}"                               
☺
  DB<2>   

Endlich erlungen. Vielen Dank....