9.2. Compatibility Issues

As of version 2.1.0, Campsite uses UTF-8 encoding the older database and templates must be converted; the database conversion is more complex and error prone while the templates conversion requires only the issuing of one shell command.

Conversion from any character set to UTF-8 can be achieved by using standard iconv tool; for usage type “man iconv”.

Converting the database can be done in three steps:

1. dumping the old database into different files

2. converting the files from their character sets to UTF-8

3. reconstructing the database from the files

Important!!! The database upgrade must take place before Campsite application upgrade. If the application upgrade already took place, “ArticleTopics” and “Topics” tables must be added to the list of tables that need not be converted to UTF.

Open a shell session and type all the following commands described below at the command prompt. Shell commands are written in italic style.

1. Dumping the old database into different files

• Dump tables structure only:

  mysqldump ... -d [database] > tables.sql

where [database] should be replaced by the name of the Campsite database and “...” with other needed parameters (user, host etc.)

• Dump data for tables that need not be converted to UTF:

  mysqldump ... -t --tables [database] ArticleIndex AutoId Events \

  Images KeywordIndex SubsByIP SubsDefTime SubsSections Subscriptions \

  UserPerm UserTypes Users > data_nolang.sql

Don't forget to add “ArticleTopics” and “Topics” to the list of tables if Campsite upgrade took place already.

• Dump data for tables that should be converted to UTF

  Group languages having the same character set (see Languages in admin interface and read the following note below). For every character set used repeat the following steps:

1. select all the languages in the character set and compose the where clause:

   "IdLanguage = [l1] or IdLanguage = [l2] ..."

   where [l1], [l2] are the language identifiers. In order to read the language identifiers and character sets log in to Campsite database and run the following sql: “select Name, Id, CodePage from Languages”.

2. mysqldump ... -t --tables --where="IdLanguage = [l1] or IdLanguage = [l2]..." \

   [database] Articles Classes Countries Dictionary Errors Issues \

   KeywordClasses Sections TimeUnits > data_[charsetn]_1.sql

3. mysqldump ... [database] -t --tables \

   --where="IdDefaultLanguage = [l1] or IdDefaultLanguage = [l2]..." \

   Publications > data_[charsetn]_2.sql

4. mysqldump ... -t --tables --where="Id = [l1] or Id = [l2]..." [database] \

   Languages > data_[charsetn]_3.sql

5. make a list of all X tables (tables whose names start with X) and dump their data:

   mysqldump ... -t --tables --where="IdLanguage = [l1] or IdLanguage = [l2]..." \

   [database] [list_of_X_tables] > data_[charsetn]_4.sql

Replace [l1], [l2] with language identifiers, [database] with Campsite database name, “...” in mysqldump command with other needed parameters (user, host etc.), [charsetn] with the name of the character set (ISO-8859-1, ISO-8859-2 etc.), [list_of_X_tables] with the list of tables whose names start with X; these are article content tables.

Note: In order to simplify the conversion all not used languages except English may be deleted. If only one language is used and it's character set is ISO-8859-1 there is no need to group languages by character set: just delete unused languages and perform the steps b through d described above and remove “--where” clause from commands.

Select the character set for Log and dump Log table data:

mysqldump ... -t --tables [database] Log > data_log.sql

Selecting the character set for Log table data: decide which language you use most often when logging in to Campsite admin interface and read the character set of the language from Languages.

2. Convert files to UTF:

   Convert tables.sql:

   iconv -f ISO_8859-1 -t UTF8 tables.sql > tables_utf.sql

Warning!!! data_nolang.sql must not be converted to UTF8.

For every character set used repeat the following steps:

1. iconv -f [charsetn] -t UTF8 data_[charsetn]_1.sql > data_[charsetn]_1_utf.sql

2. iconv -f [charsetn] -t UTF8 data_[charsetn]_2.sql > data_[charsetn]_2_utf.sql

3. iconv -f [charsetn] -t UTF8 data_[charsetn]_3.sql > data_[charsetn]_3_utf.sql

4. iconv -f [charsetn] -t UTF8 data_[charsetn]_4.sql > data_[charsetn]_4_utf.sql

Convert Log table data:

iconv -f [log_charset] -t UTF8 data_log.sql > data_log_utf.sql

where [log_charset] is the character set selected for Log table.

3. Restore the database:

   mysql ... [database] < tables_utf.sql

   mysql ... [database] < data_nolang.sql

For every character set used repeat the following steps:

mysql ... [database] < data_[charsetn]_1_utf.sql

mysql ... [database] < data_[charsetn]_2_utf.sql

mysql ... [database] < data_[charsetn]_3_utf.sql

mysql ... [database] < data_[charsetn]_4_utf.sql

Restore Log table data:

mysql ... [database] < data_log_utf.sql

Converting the templates to UTF-8: issue the following commands for every template in the /look/ directory:

iconv -f [charset] -t UTF-8 [template_name] > [template_name].utf8

mv -f [template_name].utf8 [template_name]

where [charset] must be replaced by the template character set and [template_name] by the template being processed.

Updating templates structure: Some attributes of List Article statement are no longer supported in version 2.1, so all templates using those attributes must be updated.

Attributes no longer supported:

- IsOn|IsNotOn FrontPage
- IsOn|IsNotOn Section

Replacing unsupported attributes:

- "IsOn FrontPage" replaced by: "OnFrontPage is on"
- "IsNotOn FrontPage" replaced by: "OnFrontPage is off"
- "IsOn Section" replaced by: "OnSection is on"
- "IsNotOn Section" replaced by: "OnSection is off"

For details see List of Articles.

9.1. Upgrading from 1.x or 2.0	9.3. New Template Language Features
	Last update: 2005-10-12 07:59:51