In perl.git, the branch smoke-me/khw-locale has been created
<http://perl5.git.perl.org/perl.git/commitdiff/b23bc406f40dd3f45a13570e2455a56a42d67c2e?hp=0000000000000000000000000000000000000000>
at b23bc406f40dd3f45a13570e2455a56a42d67c2e (commit)
- Log -----------------------------------------------------------------
commit b23bc406f40dd3f45a13570e2455a56a42d67c2e
Author: Karl Williamson <[email protected]>
Date: Mon Nov 28 14:16:18 2016 -0700
lib/locale.t: Don't assume NUL is a control
A test that assumed NUL would be considered a control fails in locales
where it isn't considered a control.
This was found on FREEBSD, where the locale "hi_IN.ISCII-DEV" is
defective (based on the Wikipedia article on ISCII
https://en.wikipedia.org/wiki/Indian_Script_Code_for_Information_Interchange).
Only the code points 0x09-0x0D are considered controls in this
implementation, whereas the article says ISCII is a superset of ASCII,
so should have 33 controls, not just 5. (Unrelated to this ticket, but
another apparent defect I saw is that this implementation defines 0x91,
but the article says that code point is unassigned.)
M lib/locale.t
commit b8e10a6491e032eadbbd080cf32a9165c31a806e
Author: Karl Williamson <[email protected]>
Date: Mon Nov 28 09:09:23 2016 -0700
PATCH: [perl #129953] lib/locale.t failures on FREEBSD
Because collation is very complicated, perl has opted to use the one
furnished in libc. And because Perl strings are different from C
strings in that the former allows embedded NULs, something needs to be
done when collating strings that contain NULs. What happens is that
when calculating the collation, each embedded NUL is replaced by the
lowest collating control character in the range 1-255. If no suitable
control character is found, the lowest collating code point is used.
(The reason controls are preferred is that collation is context
sensitive, with adjacent characters potentially affecting the collation
order, and controls are much less likely to have any such interactions.)
This bug occurred if the character replacing the NUL is UTF-8 variant.
Basically it got doubly encoded as UTF-8. On most systems the NUL
replacement is \x01, which is invariant under UTF-8, but in freebsd, the
replacement is \x80, which is UTF-8 variant. The solution is to store
the replacement as a code point, and encode it as UTF-8 only if
necessary, once. This actually simplifies the code
M intrpvar.h
M locale.c
commit f0876567ed46fde697ca938cb3d5af5713367f0c
Author: Karl Williamson <[email protected]>
Date: Sun Nov 27 10:25:49 2016 -0700
locale.c: Add some comments
M locale.c
commit 4295948b20cfa72a0fac03168bcedcecdd072bc0
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 22:04:13 2016 -0600
more debugging
M locale.c
M util.c
commit 530338919e1a133bf728962a79d974fe4b6df2a4
Author: Karl Williamson <[email protected]>
Date: Sun Nov 27 10:39:12 2016 -0700
Revert "lib/locale.t: Debugging"
This reverts commit
M lib/locale.t
commit 652c728b6a161f72305e1c1b99507941f14bb4da
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:32:26 2016 -0600
lib/locale.t: Debugging
M lib/locale.t
commit 355dcde35e8ab2eee0e04f41bb9a6b2cc221b687
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:31:30 2016 -0600
t/loc_tools.pl: Make a copy of inputs for bisect
M t/loc_tools.pl
commit c64af7639fe68976128ad6014c2434d521d7cfba
Author: Karl Williamson <[email protected]>
Date: Sun Nov 27 10:42:09 2016 -0700
Revert "loc_tools.pl: Debug statements"
M t/loc_tools.pl
commit b78f29addd4d1a02f55688a68dc47b3c15512ef6
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:31:05 2016 -0600
loc_tools.pl: Debug statements
M t/loc_tools.pl
commit 2701dd94526c169022a4ede9685c384fb5e3d64d
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:29:49 2016 -0600
locale.c for bisect
M locale.c
commit 124e2745c4ac3c5e35a3e44188961307717f7a23
Author: Karl Williamson <[email protected]>
Date: Sat Oct 29 12:24:21 2016 -0600
smoke
M lib/locale.t
commit 6420296d7c96d0c1d34ffc176deb82b562b10e40
Author: Karl Williamson <[email protected]>
Date: Tue Aug 9 22:36:37 2016 -0600
kid51 debug statements
M locale.c
commit 10539bd68049681f6a7a4c9fdb8d5dc810942e22
Author: Karl Williamson <[email protected]>
Date: Mon Nov 28 09:07:52 2016 -0700
Revert "hints/freebsd.sh: Don't use libc strlcpy"
M hints/freebsd.sh
commit 0e336102cdffa60c6e7a96f6bef671566c06b256
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:34:17 2016 -0600
hints/freebsd.sh: Don't use libc strlcpy
M hints/freebsd.sh
commit ec71087f0c6dd5ba19b9d2954bd752cb7865500a
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:26:14 2016 -0600
Revert "XXX wait until can write diag msg showing how to override: perl.c:
Add error test for locale"
M perl.c
M pod/perldiag.pod
commit bdcf6ac5f3a0d906db99b443318b5711fbd0c147
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:26:01 2016 -0600
Revert "p2008"
M locale.c
commit 5d2f37b8632b34bf9867079cbee4e137e9dc269e
Author: Karl Williamson <[email protected]>
Date: Sun Oct 30 21:25:36 2016 -0600
Revert "Move bulk of POSIX::setlocale to locale.c"
M embed.fnc
M ext/POSIX/POSIX.xs
M locale.c
M proto.h
commit 3494aea92d2fd32736c875f660ac18133745ba4d
Author: Karl Williamson <[email protected]>
Date: Sat May 28 21:47:30 2016 -0600
Move bulk of POSIX::setlocale to locale.c
This is the first step in making an XS callable setlocale() which
is generalized to work with POSIX 2008 locale operations while retaining
essentially the same API.
M embed.fnc
M ext/POSIX/POSIX.xs
M locale.c
M proto.h
commit 9c38c0365d938475ddf8e29300c6d3501f5eeadf
Author: Karl Williamson <[email protected]>
Date: Thu Jul 21 22:05:56 2016 -0600
p2008
M locale.c
commit 3de795297945ab794741a485ecd355c89617e1ff
Author: Karl Williamson <[email protected]>
Date: Wed Aug 10 12:08:01 2016 -0600
XXX wait until can write diag msg showing how to override: perl.c: Add
error test for locale
M perl.c
M pod/perldiag.pod
-----------------------------------------------------------------------
--
Perl5 Master Repository
