tl2020 pretest begin

[David Carlisle]

On Sat, 14 Mar 2020 at 03:11, Paulo Ney de Souza <pauloney at gmail.com> wrote:

>> > But something that is signed with "Petr Ol¹ák" is very big indication that
>> > the file was written in ISO-8859-2 and saved in ISO-8859-1. In fact if you
>> > change it -- the whole file becomes correct.


No. The file is saved in ISO-8859-2 the problem is not in the file, it
is in whatever
you used to read the file which guessed the wrong encoding. (The fact
that i's not possible
to guess right every time is a problem that Unicode aims to solve but
that doesn't
mean that files in 8-bit encodings are wrong).

Being a plain text file there is no machine readable way to mark the
encoding in the file.

But it does say in the file what the encoding is:

README     ... tento soubor (formát UNIX, ISO 8859-2),


That is  "this file (UNIX format, ISO 8859-2)"  This indicates another
reason why changing the
encoding of text files needs to be passed back to the maintainers.
Changing the encoding of that
file is not a mechanical operation it requires updating the Czech text
to refer to whatever encoding
is used.

David