On Mon, Mar 30, 2015 at 12:00:26PM +0100, Dave Mitchell wrote:

I guess that one extra level of encoding is caused by the smoker code when
generating smoke logs, but I don't see why the 'got' message should have
an extra level of encoding on top of that, and why it's intermittent
(sometimes a mismatch between TEST and harness, and for some
configurations not at all), and why it doesn't fail for me.

The smokes seem to only fail for the permutations with LC_ALL=en_US.utf8.
I can consistently see the failures under t/harness on dromedary with

I ran a bisect as:

LC_ALL=en_US.UTF-8 PERL_UNICODE="" perl Porting/bisect.pl --start v5.21.10 --target lib/warnings.t

and it reports that the errors start at this commit:

commit 8ce2ba821761a7ada1e1def512c0374977759cf7
Author: Alex Vandiver <alex@chmrr.net>
Date: Sun Mar 22 23:08:24 2015 -0400

     Fix "...without parentheses is ambuguous" warning for UTF-8 function names

     While isWORDCHAR_lazy_if is UTF-8 aware, checking advanced byte-by-byte.
     This lead to errors of the form:

        Passing malformed UTF-8 to "XPosixWord" is deprecated
        Malformed UTF-8 character (unexpected continuation byte 0x9d, with
          no preceding start byte)
        Warning: Use of "�" without parentheses is ambiguous

     Use UTF8SKIP to advance character-by-character, not byte-by-byte.

(and by implication they are not a side effect of a later commit)

I'm not in a position to investigate further as to why, let alone provide a

Nicholas Clark

Search Discussions

Discussion Posts


Follow ups

Related Discussions

Discussion Navigation
viewthread | post
posts ‹ prev | 4 of 6 | next ›
Discussion Overview
groupperl5-porters @
postedMar 27, '15 at 8:18p
activeMar 31, '15 at 8:50a



site design / logo © 2019 Grokbase