This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Remove obsolete aliases that broke 'locale -a'
- From: Paul Eggert <eggert at cs dot ucla dot edu>
- To: Carlos O'Donell <carlos at redhat dot com>, GNU C Library <libc-alpha at sourceware dot org>
- Date: Thu, 14 May 2015 10:20:37 -0700
- Subject: Re: [PATCH] Remove obsolete aliases that broke 'locale -a'
- Authentication-results: sourceware.org; auth=none
- References: <55544398 dot 2030007 at cs dot ucla dot edu> <5554CAAC dot 20508 at redhat dot com>
On 05/14/2015 09:17 AM, Carlos O'Donell wrote:
This looks like your ChangeLog entry is not in UTF-8?
No, it's a problem in transmission. The actual patch is a mixture of
UTF-8 and Latin-1 encoding. I sent it via Thunderbird, which doesn't
know its encoding and says only "Content-Type: text/x-patch" without
specifying a character set for the attachment. The body of the
attachment, generated by Git, says "Content-Type: text/plain;
charset=UTF-8" which is not correct, as the attachment body uses some
UTF-8 and some Latin-1.
If I save the attachment into a file FOO, and edit FOO with Emacs, Emacs
ignores the charset declaration and infers Latin-1 since FOO can't
possibly be UTF-8, which botches display of the UTF-characters (the
botch you noticed). If I then type 'C-x RET r U T F - 8 RET y e s RET',
Emacs will show the patch as UTF-8 instead of Latin-1, which will fix
the botch you noticed but will mess up display of the old contents of
locale.aliases.
What a mess, huh?
Could you please convert this file to UTF-8 instead of ISO-8859?
Sure, revised patch attached. It'll have the same transmission problems
as before, and if you force Emacs to read it in UTF-8 only the old lines
will be botched.
>From 442fe29ef8cb66f8a1c012705a75b6d9a57111ce Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Thu, 14 May 2015 10:18:11 -0700
Subject: [PATCH] Remove obsolete aliases that broke 'locale -a'
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
[BZ #18412]
* intl/locale.alias: Remove obsolete aliases "bokmål" and "français"
which caused 'locale -a' to output Latin-1 data in UTF-8 locales,
breaking some applications that use 'locale -a' output.
Change the encoding of this file from Latin-1 to UTF-8
(this affects only the file's comments now).
---
ChangeLog | 7 +++++++
intl/locale.alias | 9 +++++++--
2 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index e89c14e..2346fcf 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+2015-05-14 Paul Eggert <eggert@cs.ucla.edu>
+
+ [BZ #18412]
+ * intl/locale.alias: Remove obsolete aliases "bokmål" and "français"
+ which caused 'locale -a' to output Latin-1 data in UTF-8 locales,
+ breaking some applications that use 'locale -a' output.
+
2015-05-14 Andrew Senkevich <andrew.senkevich@intel.com>
* Makeconfig (rpath-dirs, all-subdirs): Added mathvec folder.
diff --git a/intl/locale.alias b/intl/locale.alias
index ab1cb7a..ca4e777 100644
--- a/intl/locale.alias
+++ b/intl/locale.alias
@@ -24,8 +24,14 @@
# backward compatibility. Nobody should rely on the names defined here.
# Locales should always be specified by their full name.
+# Note: This file used to contain the following lines encoded in Latin-1:
+# bokmål nb_NO.ISO-8859-1
+# français fr_FR.ISO-8859-1
+# These lines were removed because they caused 'locale -a' to output
+# text encoded in Latin-1, which broke applications in UTF-8 locales. See:
+# https://sourceware.org/bugzilla/show_bug.cgi?id=18412
+
bokmal nb_NO.ISO-8859-1
-bokmål nb_NO.ISO-8859-1
catalan ca_ES.ISO-8859-1
croatian hr_HR.ISO-8859-2
czech cs_CZ.ISO-8859-2
@@ -36,7 +42,6 @@ dutch nl_NL.ISO-8859-1
eesti et_EE.ISO-8859-1
estonian et_EE.ISO-8859-1
finnish fi_FI.ISO-8859-1
-français fr_FR.ISO-8859-1
french fr_FR.ISO-8859-1
galego gl_ES.ISO-8859-1
galician gl_ES.ISO-8859-1
--
2.1.0