On Sun Jun 29 03:11:12 2014, cfaerber wrote:
The interesting part is that the tests are the tests provided with
Unicode 7.0.0.
Yes, and my comments in rejecting this ticket were based on ignorance. I'm sorry. Though it still should have been rejected, as it still doesn't appear to me to be a core Perl bug. I looked in more detail at the first error in the CPAN report. It is this:

# Failed test 'to_ascii('ß。𐋳ⴌ\u1DD8') throws error P1 V6 [data/IdnaTest.txt:992]'
# at t/uts46_to_ascii-trans.t line 765.
# got: 'ss.xn--weg506dvy5n'
# expected: undef

I looked at that line in the .t file, and it is this:

𐋳ⴌ\x{1DD8}", %p)}, undef, "to_ascii\(\'ß\。𐋳ⴌ\\u1DD8\'\)\ throws\ error\ P1\ V6\ \[data\/IdnaTest\.txt\:992\]") or ($@ and diag($@));

(Most likely you will not have the fonts to display all of this correctly. the one character I don't have in my fonts is U+102F3, COPTIC EPACT NUMBER ONE HUNDRED, newly encoded in Unicode 7.0. I got that far before, and just assumed the test was supposed to return undef because the code point had not been encoded before, but now is. But I was wrong. There is another reason it is supposed to be undef.

Line 992 from the Unicode 7.0 IdnaTest.txt file is this:
B; ß。𐋳ⴌ\u1DD8; [P1 V6]; [P1 V6] # ß.𐋳ⴌᷘ

(BTW, thanks for cross referencing the line number of the Unicode file in the .t test; it made this a lot easier.)

The brackets indicate that this is supposed to fail, and the codes within the brackets indicate why. I started to follow why it should fail, but it wasn't obvious without more digging than I had time for. So perhaps the test from Unicode is wrong, or the module is buggy. I see that the .t correctly gets failure with the preceeding, similar tests,

Karl Williamson

via perlbug: queue: perl5 status: rejected

Search Discussions

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupperl5-porters @
postedJul 1, '14 at 7:23p
activeJul 1, '14 at 7:23p

1 user in discussion

Karl Williamson via RT: 1 post



site design / logo © 2021 Grokbase