I noticed that the Japanese text in the Windows installers for the latest releases (v3.0.14, v4.0.7, and v5.0.4) is completely corrupted (showing Mojibake), as seen in the attached screenshot.
Background:
I previously contributed the Japanese translation for these installers. When I built the installer locally using Inno Setup Unicode version on a Japanese Windows environment, the text displayed correctly.
Root Cause Analysis:
At that time, I provided the .isl files encoded in Shift-JIS (CP932). While this works on Japanese machines, it seems the official build pipeline is interpreting these files using a different local code page (likely CP1252), causing the encoding mismatch.
Proposed Fix:
To ensure the installer displays correctly regardless of the build environment's locale, the Japanese .isl files should be converted to UTF-8 with BOM.
Inno Setup Unicode version will correctly recognize the UTF-8 BOM and handle the characters properly, preventing this regression in future releases.
I am happy to provide the updated files encoded in UTF-8 with BOM if needed. Please let me know how you would like to proceed.
firebird_japanese_installer_mojibake.zip
I noticed that the Japanese text in the Windows installers for the latest releases (v3.0.14, v4.0.7, and v5.0.4) is completely corrupted (showing Mojibake), as seen in the attached screenshot.
Background:
I previously contributed the Japanese translation for these installers. When I built the installer locally using Inno Setup Unicode version on a Japanese Windows environment, the text displayed correctly.
Root Cause Analysis:
At that time, I provided the .isl files encoded in Shift-JIS (CP932). While this works on Japanese machines, it seems the official build pipeline is interpreting these files using a different local code page (likely CP1252), causing the encoding mismatch.
Proposed Fix:
To ensure the installer displays correctly regardless of the build environment's locale, the Japanese .isl files should be converted to UTF-8 with BOM.
Inno Setup Unicode version will correctly recognize the UTF-8 BOM and handle the characters properly, preventing this regression in future releases.
I am happy to provide the updated files encoded in UTF-8 with BOM if needed. Please let me know how you would like to proceed.
firebird_japanese_installer_mojibake.zip