Your periodic reminder that #microsoft is not a competent or serious company:

blog.orange.tw/posts/2025-01-w…

(@bagder is remarkably restrained in the quoted responses. I would have gone off the rails, but do not need to, because we decided on day one that #varnishcache would not run on Windows).

If you think that is not bad enough, read the Cyber Safety Review Board's report about the Microsoft Exchange clowncar:

cisa.gov/sites/default/files/2…

in reply to daniel:// stenberg://

It is absolutely, entirely and 100% a Microsoft problem.

And it is a prime example of "Marcitecture":

"Whatever you do with this Unicode thing, DO NOT UPSET PROGRAMMERS so much the might look at other platforms! All existing code SHALL JUST WORK!!1!"

But what really amazes me, is how and why anybody still takes them seriously about /anything/ ?!

in reply to daniel:// stenberg://

If only the transition to native Unicode wouldn't be a gargantuan undertaking. The world isn't UTF-16, so these interfaces need to be identified (with their encoding/codepage) and a conversions put in place. Windows UTF-8 support could be a solution, but it's stuck in beta, with issues. Switching the official curl build to Unicode is an option, but there are unresolved issues, e.g. string encoding in the libcurl API is undefined. I wish this would be easier.
This entry was edited (3 months ago)
in reply to daniel:// stenberg://

Proposing this patch to re-enable Unicode in curl official Windows builds. Caveats described in 2021 remain, but perhaps now the balance tilts towards giving it another go: github.com/curl/curl-for-win/c…