On Jul 7, 2017, at 1:08 AM, Pascal Quantin <pascal.quantin@xxxxxxxxx> wrote:
> g_ascii_tolower
Note: in a given locale, "i" isn't necessarily the lower-case version of "I".
In fact, it's definitely *not* the lower-case version of "I" in Turkish. This caused crashes at one point:
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=2010
g_ascii_tolower() gives the same answer in all locales. If you're dealing with a configuration file, then, unless the configuration file's language is based on Turkish (or Azerbaijani, Kazakh with the Roman alphabet, Tatar with the Roman alphabet, or Crimean Tatar, if the Wikipedia page
https://en.wikipedia.org/wiki/Dotted_and_dotless_I
is correct), that answer is the one you want. (And if it *is* based on one of those languages, you would *still* want the same answer in all locales, as there's no guarantee that the file would only be read in a Turkish, Azerbaijani, Kazakh, Tatar, or Crimean Tatar locale, so you'd have to map "I" to "ı" and "İ" to "i".)