Em Thu, 5 Sep 2019 13:40:08 -0600 Jonathan Corbet <corbet@xxxxxxx> escreveu: > On Thu, 5 Sep 2019 16:28:10 -0300 > Mauro Carvalho Chehab <mchehab+samsung@xxxxxxxxxx> wrote: > > > I don't think we can count that python 3 uses utf-8 per default. > > > > I strongly suspect that, if one uses a Python3 version < 3.7, it will > > still default to ASCII. > > > > On a quick look, the new UTF-8 mode was added on PEP-540: > > > > https://www.python.org/dev/peps/pep-0540/ > > > > Such change happened at Python 3.7. > > That PEP is to override the locale and use utf8 unconditionally. It > says, with regard to the pre-PEP state: > > UTF-8 is also the default encoding of Python scripts, XML and JSON > file formats. > > Unicode was the reason for much of the Python 3 pain; it seems unlikely > that many installations are defaulting to ASCII anyway...? Yeah, but I remember that UTF-8 handling changed a few times during python 3 releases. I didn't really tracked what happened, as I don't usually program in Python. So, I'm actually relying on what I can find about that. Looking at Python 3.0 release[1], it says: "In many cases, but not all, the system default is UTF-8; you should never count on this default." [1] https://docs.python.org/3.0/whatsnew/3.0.html So, at least on early Python 3 releases, the default may not be UTF-8. I don't know about you, but, from time to time, people complain about UTF-8 chars when I'm handling patches (last time was on a patch series for Kernel 5.3 by a core dev in Australia, with was unable to apply a patch from me with had some UTF-8 chars). So, I'm pretty sure that some devs don't set the locale to UTF8 even those days. Thanks, Mauro
