Questions[charset](server)

I'm troubleshooting a smtp issue involving character-set encoding and it's extremely difficult to involve the end user. I have a packet capture of an incoming smtp session that results in problematic behavior but if I try to copy/paste the mime source from wireshark, the issue doesn't occur.

I'm reasonably certain that the difference in my testing is that I'm copying the strings in the packet capture from the Wireshark "follow tcp stream" output and that's not necessarily the bit-for-bit accurate copy of the data as it arrived on the wire.

So my question is this: how can I get an EXACT copy of the bit-for-bit SMTP data and send it again for reproduction purposes? Is there a way to "replay" a transmission?

What is the right way to enable correct charset headers in NGINX?

I'm analyzing my website with Google Page Speed. It says that I should specify the charset of HTML files in HTTP-headers. What is the right way to do this?

I already tried to set charset utf-8; in the server {} declaration of my NGINX configuration file, but it hasn't got any effect.

My server responds with the following header:

Connection: keep-alive
Date: Fri, 16 Sep 2011 12:43:24 GMT
Last-Modified: Fri, 02 Sep 2011 15:13:17 GMT
Server: nginx/0.7.67

Thank you.

Here's the output of locale:

LANG=zh_CN.GBK
LC_CTYPE="zh_CN.GBK"
LC_NUMERIC="zh_CN.GBK"
LC_TIME="zh_CN.GBK"
LC_COLLATE="zh_CN.GBK"
LC_MONETARY="zh_CN.GBK"
LC_MESSAGES="zh_CN.GBK"
LC_PAPER="zh_CN.GBK"
LC_NAME="zh_CN.GBK"
LC_ADDRESS="zh_CN.GBK"
LC_TELEPHONE="zh_CN.GBK"
LC_MEASUREMENT="zh_CN.GBK"
LC_IDENTIFICATION="zh_CN.GBK"
LC_ALL=

How can I change all of them to UTF8? How can I make the locale setting persistent in CentOS 5.5?

I want to backup a mediawiki database stored in a MySQL server 5.1.36 using mysqldump.

Most of the wiki articles are written in spanish and a don't want to mess up with it by creating the dump with the wrong character set.

mysql> status
--------------
...
Current database:       wikidb
Current user:           root@localhost
...
Server version:         5.1.36-community-log MySQL Community Server (GPL)
....
Server characterset:    latin1
Db     characterset:    utf8
Client characterset:    latin1
Conn.  characterset:    latin1
...

Using the following command:

mysql> show create table text;

I see that the table create statement set the charset to binary:

CREATE TABLE `text` (
  `old_id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `old_text` mediumblob NOT NULL,
  `old_flags` tinyblob NOT NULL,
  PRIMARY KEY (`old_id`)
) ENGINE=InnoDB AUTO_INCREMENT=317 DEFAULT CHARSET=binary MAX_ROWS=10000000 AVG_ROW_LENGTH=10240

How should I use mysqldump to properly generate a backup for that database?

After some problems with UTF8 and none-UTF8 strings, we're standardising on UTF8. One thing I need to do is check that everything is in UTF8 in the MySQL database? What do I need to check?

• Server default characterset
• Default character set of each database
• Does each text column have a character set? How do I check for that?

I was thinking of putting all this in a nagios plugin to check that everything is in UTF8. Suggestions?