[pcre-dev] [Bug 2305] \N{U+NNNN} folding regression with 10.…

Top Page
Delete this message
Author: admin
Date:  
To: pcre-dev
Subject: [pcre-dev] [Bug 2305] \N{U+NNNN} folding regression with 10.32-RC1
https://bugs.exim.org/show_bug.cgi?id=2305

Philip Hazel <ph10@???> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|NEW                         |RESOLVED


--- Comment #2 from Philip Hazel <ph10@???> ---
Nothing has changed here. Unless you set the UTF option, which implies Unicode,
/i is implemented for codepoints less than 256 with a lookup table that is
created from tolower(). In the default C locale this won't find another case
for non-ASCII characters. However, in (e.g.) a French locale, it does. \N{U+41}
is just a synonym for \x{41} in PCRE2. This is what the current 10.31 release
does:

PCRE2 version 10.31 2018-02-12
/(?i:\x{41}\x{c1})/
    a\x{e1}
No match


/(?i:\x{41}\x{c1})/utf
    a\x{e1}
 0: a\x{e1}


/(?i:\x{41}\x{c1})/locale=fr_FR
    a\x{e1}
 0: a�


The same is true in 10.32-RC1 using the \N notation. I wonder why this is
causing a regression, because it shouldn't have worked before.

--
You are receiving this mail because:
You are on the CC list for the bug.