Advanced science. Smarter decisions. Superior monetization.

Upgrading character encoding

Audience

This section is only relevant to administrators who are upgrading from OpenX 2.4 or earlier. If you are installing OpenX, or you are performing a different upgrade, you do not need to read this.

Additionally, if all manager accounts (formerly known as admin/agency accounts) in your OpenX installation have been using the interface in English (or another language with a pure ASCII character set) then this change will not affect you.

The problem with encoding

In OpenX 2.4 and earlier, language packs were provided for a number of languages. Each language pack contained a different character encoding, and set the user's browser to use that encoding (in the case of 2.0.11 and lower some languages did not set the encoding to use, and relied on the browser to set the encoding).

The encoding used by the browser was used as the encoding for the data stored in user-entered fields in the database

In OpenX 2.6 and later we have changed the encoding of these language packs to UTF-8. This was required by (among other things) the change to users and accounts; you might now have the situation where items could be edited by different users who might have different selected languages. Without switching to UTF-8 encoding, any data created by one user and then modified by another with different language settings would have become corrupted.

All OpenX language packs have been converted to UTF-8 encoding, so it shouldn't matter if multiple users (with different languages) edit each others' data.

What this means

For the majority of users the upgrade and data conversion should be completely automatic.

In cases where automatic detection is not possible we do not modify any user-data. However, we have provided a tool which can be used to change the encoding of user data after the upgrade.

Do I need to do anything?

Probably not. The upgrader should have converted any applicable data in your database into UTF-8 encoding.

Appendix

The upgrade system will apply the following conversion rules based on the manager's selected language:

ACCOUNT'S LANGUAGE ENCODING CONVERTED FROM
Chinese (big5) big5
Czech iso-8859-2
French iso-8859-15
Hebrew windows-1255
Hungarian iso-8859-2
Korean EUC-KR
Polish iso-8859-2
Portuguese iso-8859-15
Russian (cp1251) windows-1251
Russian (koi8r) koi8-r
2.0 upgrades only
Chinese (gb2312) gb2312