msys2: "Norwegian Bokmål" misdisplayed in Windows installer
(outstanding part of Gna bug #21060)
The description of the "nb" language code (Norwegian Bokmål) is mangled on display, at least on my (Windows 7 British English) system. See screenshot. It looks to me like UTF-8 interpreted as ISO 8859-1 or Win1252.
bootstrap/langnames.txt (from which this string comes) is encoded in UTF-8. I'm guessing that encoding will pass straight through create-freeciv-gtk2-nsi.sh to the NSIS script, and what happens then will depend on NSIS' character set handling.
Is there some way to tell NSIS that the script it's been given is UTF-8 encoded, and if so will everything from then on Just Work?
Failing that, we could take the pragmatic step of changing the encoding of langnames.txt in svn to ISO 8859-1.
#7 Updated by Marko Lindqvist 7 months ago
- File 0016-msys2-Convert-langstat-files-to-ISO-8859-1-when-buil.patch 0016-msys2-Convert-langstat-files-to-ISO-8859-1-when-buil.patch added
Here's a patch to do installer build time conversion of the language statistics I will test when I get the opportunity. This testpatch is for S3_0 msys2 installers.
#9 Updated by Marko Lindqvist 7 months ago
- Subject changed from "Norwegian Bokmål" misdisplayed in Windows installer to msys2: "Norwegian Bokmål" misdisplayed in Windows installer
- Status changed from In Progress to Resolved
- Target version changed from 2.6.2 to 2.6.1
Splitting to two tickets, one for msys1 and one for msys2. This is mainly because fixes may land in different releases. As this ticket already has msys2 patch, making Bug #852637 the msys1 one.
It's not clear if this patch can make it to 2.6.1 release or not, but working towards it.