CSDT export: multi-byte characters converted into question marks (Doc ID 1448713.1)

Last updated on AUGUST 22, 2019

Applies to:

Symptoms

When the CSDT exports sites on Linux, it converts UTF-8 multi-byte characters to "?" when it generates the asset file names.

This causes two problems:

1. "?" is an illegal character on Windows, making it impossible to check out those sites from the source repository on Windows.

2. If assets refer to other assets by name (specifically, assets that are linked to image assets), the file name is encoded as UTF-8 in the dependent file, but the file names of the image assets have the converted "?" character.
During import, the CSDT fails to find the dependencies.

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.