Front page | perl.perl5.porters |
Postings from March 2014
On Mon, Mar 24, 2014 at 11:08 AM, Zefram <zefram@fysh.org> wrote:
> Ricardo Signes wrote:
>>3. "Make perlbug Unicode-aware" broke perlbug on Win32
>> https://rt.perl.org/Ticket/Display.html?id=121277
>
> I think there's a bug in the Unicode-awareness patch. $input_encoding is
> set by
>
> my $input_encoding = $::Have_get_locale_encoding ? encoding::_get_locale_encoding() : '';
>
> and then used as
>
> open(F, "<:$input_encoding", $file)
>
> $input_encoding will get values such as "UTF-8" and "ANSI_X3.4-1968"
> which are not valid layer names. It should instead be initialised as
>
> my $input_encoding = $::Have_get_locale_encoding ?
> "encoding(@{[encoding::_get_locale_encoding()]})" : '';
Thanks for spotting this. The code as-is worked fine in my testing
with the default encodings on Mac OS X, VMS, and Windows, but those
were all U.S. English environments so perhaps not representative. I
don't actually understand why encoding names don't work as layer names
and thought I had seen something indicating they do. Also, I'm not
sure whether _get_locale_encoding can ever return an empty string. If
it can, I think Zefram's suggestion would be using an empty encoding
rather than an empty layer; the latter is documented to work but I'm
not sure about the former.
> As for the CRLF issue, perlbug should probably detect whether :crlf is
> a default layer and handle it appropriately.
Does the :crlf layer even work with Unicode? Some people seem to
think it doesn't (e.g.,
<http://stackoverflow.com/questions/3320189/crlf-translation-with-unicode-in-perl>)
but I don't know enough to evaluate. If it can be made to work so
that Unicode handling is not lost, I'm ok with supporting CRLF line
endings, but it's really a bug in Notepad and not in Perl.
> Something like
>
> use PerlIO ();
> my $crlf = (grep { $_ eq "crlf" } PerlIO::get_layers(STDOUT)) ?
> ":crlf" : "";
> ...
> open(REP, ">:raw $crlf", $filename)
> ...
> open(REP, "<:raw $crlf", $filename)
>
> (The opening for input occurs in four places.)
>
> -zefram
Thread Previous
|
Thread Next