We are having problems importing in strings of a different locale from a Unicode text file into our ism on our build machine. This import works fine on all our development machines and another build machine – but the machine we want to use is not importing in the strings properly. The build machines both just have IS 2009 standalone builder, and are Windows XP.
Each install is initially created in English and then transformed to the other language. Part of this transform imports the language strings for that locale using the Automation Interface and the script that was supplied by InstallShield. On the machines that work, the strings import fine in UTF-8. Here is an example string to be imported and how it’s displayed in the String table
However, on the machine that doesn’t work this appears to have been imported as base64 encoding – it’s clearly different to the UTF-8 encoding above
ID_STRING27103410!S`'``80#Q`&\`;```71016968304
When this is then built the text is not converted back correctly and the text appears as squares in the IDE.
This is strange as it is the exact same script that imports these text strings, the version of the Standalone Automation Interface is the same and the files are the text files are the same binary UNICODE text files – the same file imports fine on the other machines when copied over.
Does anyone know of any environmental factor or anything else that may cause this issue? This is causing us great delays and a lot of time has been spent trying to figure this out.
