Bad character encoding in local documentation

[FYHenry]

Bug summary

• Character encoding error int the HTML table elements of the documents.
• Substitution of any non ASCII character by ? character.

Variables

• WEB_BROWSER can be a Web browser command like firefox or lynx,
• DOC_ROOT is the system directory like /usr/share/doc/weechat-doc/html/ in Debian distribution,
• DOC_PAGE_TITLE is the title part in the file name like weechat_plugin_api in weechat_plugin_api.fr.html,
• DOC_LANG is the two-characters language code like fr in weechat_plugin_api.fr.html.

Steps to reproduce

1. Run from the shell ${WEB_BROWSER} file://${DOC_ROOT}/${DOC_PAGE_TITLE}.${DOC_LANG}.html
2. Match some word patterns containing ? in the table HTML elements.
3. Repeat the two first operations replacing the Web browser by a text editor like nano to match in the HTML source code.

Current behavior

Some word patterns like the regular expression \W[a-zA-Z0-9]+\?[a-zA-Z0-9]+\W are matched in HTML table cells from the translated documents.

Expected behavior

Some word patterns like the regular expression \W[a-zA-Z0-9]+\D[a-zA-Z0-9]+\W are matched in HTML table cells from the translated documents : the non ASCII characters can appear in the table cells.

Suggested solutions

Modify documents processing for its local version.

Additional information

LANG = fr_FR.UTF-8

---

• WeeChat version: 4.0.2
• OS, distribution and version: Debian 11 GNU/Linux Bullseye (x86_64)
• Terminal: Gnome Terminal
• Terminal multiplexer (screen/tmux/…/none): None

[flashcode]

The fix made is just a warning about missing locales when doc is built, with a fallback to English locale if a locale is missing.

This will just use English, but the real fix for French docs is to install all locales before building them.