News for hackers. Stuff that matters.
Re:Setting the TZ environment variable avoids thousands of s... by unilynx on Thursday January 12, 2023 @01:09AM (#34349819)
ASCII is a subset of UTF-8. It’s time for all tools and paths to just assume UTF-8 encoding.

Most tasks assuming ASCII work fine on UTF-8 too (eg sorting). We do need to get rid of those byte order marks though, it breaks eg concatenation

Re:Setting the TZ environment variable avoids thousands of s... by mike_hock on Thursday January 12, 2023 @02:34AM (#34350328)
> Most tasks assuming ASCII work fine on UTF-8 too (eg sorting).

cliché < clichñ < clich́e

NFC vs. NFD.

Re:Setting the TZ environment variable avoids thousands of s... by yorwba on Thursday January 12, 2023 @03:28AM (#34350699)
b'cliche\xcc\x81' (cliché) < b'clichz' (clichz) < b'clich\xc3\xa9' (cliché) < b'clich\xc3\xb1' (clichñ)

Combining marks like accents are placed after the base character.

Re:Setting the TZ environment variable avoids thousands of s... by mike_hock on Friday January 13, 2023 @06:48AM (#34366546)
You're right. Sorry, I get confused by Qt's multi-decade old bug that they're not only refusing to fix but keep porting to new major versions.

I bet they have unit tests asserting that the bug is still present.

Re:Setting the TZ environment variable avoids thousands of s... by rjh29 on Thursday January 12, 2023 @04:26AM (#34351116)
Sort orders are different per locale. For example German and Swedish both have ä and ö but sort them differently relative to the other letters.
Re:Setting the TZ environment variable avoids thousands of s... by kzrdude on Thursday January 12, 2023 @06:34AM (#34352065)
The posix locale and us-english both have `B` and `a` but sort them differently. (Posix B < a, US-english a < B) :)