summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--docs/Samba-Guide/SBE-UpgradingSamba.xml38
1 files changed, 38 insertions, 0 deletions
diff --git a/docs/Samba-Guide/SBE-UpgradingSamba.xml b/docs/Samba-Guide/SBE-UpgradingSamba.xml
index 0100a34aa5..5f9fca2460 100644
--- a/docs/Samba-Guide/SBE-UpgradingSamba.xml
+++ b/docs/Samba-Guide/SBE-UpgradingSamba.xml
@@ -462,6 +462,7 @@ Paths:
</para>
<para>
+ <indexterm><primary></primary></indexterm>
It is important that both the &smb.conf; file and the <filename>secrets.tdb</filename> should
be backed up before attempting any upgrade. The <filename>secrets.tdb</filename> file is version
encoded and therefore a newer version may not work with an older version of Samba. A backup
@@ -470,6 +471,43 @@ Paths:
</sect3>
+ <sect3>
+ <title>International Language Support</title>
+
+ <para>
+ <indexterm><primary>unicode</primary></indexterm>
+ <indexterm><primary>character set</primary></indexterm>
+ <indexterm><primary>codepage</primary></indexterm>
+ <indexterm><primary>internationalization</primary></indexterm>
+ Samba-2.x had not support for Unicode, instead all national language character set support in file names
+ was done using particular locale codepage mapping techniques. Samba-3 supports Unicode in file names, thus
+ providing true internationalization support.
+ </para>
+
+ <para>
+ <indexterm><primary>8-bit</primary></indexterm>
+ Non-English users whose national language character set has special characters and who upgrade naively will
+ find that many files that have the special chracters in the file name will see them garbled and jumbled up.
+ This typically happens with umlauts and accents because these characters were particular to the codepage
+ that was in use with Samba-2.x using an 8-bit encoding scheme.
+ </para>
+
+ <para>
+ <indexterm><primary>UTF-8</primary></indexterm>
+ Files that are created with Samba-3 will use UTF-8 encoding. Should the file system ever end up with a
+ mix of codepage (unix charset) encoded file names and UTF-8 encoded file names, the mess will take some
+ effort to set straight.
+ </para>
+
+ <para>
+ <indexterm><primary>convmv</primary></indexterm>
+ A very helpful tool is available from Bjorn Jacke's <ulink url="http://j3e.de/linux/convmv/">convmv</ulink>
+ work. Convmv is a tool that can be used to convert file and directory names from one encoding method to
+ another. The most common use for this tool is to convert locale encoded files to UTF-8 Unicode encoding.
+ </para>
+
+ </sect3>
+
</sect2>
</sect1>