> Nsis fails to create an installer file when presented a licese-data file that > is ISO-8859 encoded. In the previous version, this worked.
Advertising
> If the file is encoded as ASCII or UTF-8, this error does not occur. Looks like a "don't do that, then" kind of error. The sooner we deprecate ancient locales, the better. These files are not readable in most modern GUIs (non-UTF8 support is not even paid lip service anymore), require per-file metadata, and what you mentioned, ISO-8859, is not even Windows compatible. 8859-1 is 87% identical with CP1252, but 8859-2 and CP1250 have little overlap, etc. So if nsis 3 requires Unicode, that's a much awaited improvement, not a regression. Requiring projects to have text encoded in UTF-8 allows it being readable in all locales. I for one live in Poland, get Windows software in Polish, see Google's pages insisting on Polish despite explicitely requesting English -- yet I had three years of German in school thus I can at least roughly understand a bit of German translations you write, at least to the point of being able to see if they look plausible. Yet with ISO-8859-1, I'd get all umlauts and betas[1] corrupted. Thus, German text might be uncomfortable to read -- but I also happen to have had three years of Russian in school as well, you can guess that Russian text in an ancient locale gets a little bit more mangled than the occasional umlaut. Unicode has none of these issues. ᛗᛖᛟᚹ! [1]. You mean, ss was merely conflated with beta in some ancient charsets? -- ⢀⣴⠾⠻⢶⣦⠀ ⣾⠁⢰⠒⠀⣿⡁ ⢿⡄⠘⠷⠚⠋⠀ ... what's the frequency of that 5V DC? ⠈⠳⣄⠀⠀⠀⠀
