Content search fails for Katakana words containing dakuten/handakuten

Distribution

Mint 22.3

Package version

6.6.3

Frequency

Always

Bug description

When using Nemo's built-in content search, words written in Katakana that contain dakuten or handakuten are not found.

Examples:

Works:

• オイル

Fails:

• ランプ
• ピザ
• ブドウ
• ガラス

The same files can be searched successfully with grep, so the file contents are correct.

Searching by filename works correctly.
The problem only affects searching inside file contents.

Steps to reproduce

1. Create a UTF-8 text file containing the word "ランプ".
2. Open the folder in Nemo.
3. Click the search button.
4. Select content search.
5. Search for "ランプ".

The search returns "No files found".

Searching for "オイル" in another text file works correctly.

Expected behavior

Nemo should find UTF-8 text containing Katakana words regardless of whether they include dakuten or handakuten.

The search results should match grep.

Additional information

Environment:

• Linux Mint 22.3 (Zena)
• Nemo 6.6.3+zena
• Locale: ja_JP.UTF-8

The same search works correctly on another machine:

• Linux Mint 22.2 (Zara)
• Nemo 6.4.5+xia

So this appears to be a regression introduced between Nemo 6.4.5 and 6.6.3.

Terminal output from starting Nemo:

Nemo-WARNING **: Action '90_new-launcher.nemo_action' is missing dependency: cinnamon-desktop-editor

I do not think this warning is related to the search issue.

Start by reproducing the issue with Nemo's content search using UTF-8 text containing ランプ, ピザ, ブドウ, or ガラス, and compare the result with grep and Nemo 6.4.5. Trace the content-search path across the regression range; done means Katakana words with dakuten or handakuten are found while filename search remains working.