tl2020 pretest begin

[Zdenek Wagner]

...
>
> I hesitate to reply to this, partly because I'm still on 2019,
> partly because most people on this list understand things which I
> don't, and partly because I've no idea how Paulo is looking at these
> files which do look to be borken.
>
> I'm well aware that running 'file' (and even diff, if it decides
> that the first few hundred lines are 8859-1 or ASCII, e.g. when
> diffing firefox versions, can produce rubbish), so a couple of
> questions :
>
> For Paulo - how are you looking at these files to see this rubbish ?
>
I can answer containing the Czech text from CWEB. Paolo's locales are
set to ISO-8859-1 but the Czech text is apparently either in
ISO-8859-2 or in CP-1250. If Paolo's locales were UTF-8, the result
would be different. Being native Czech I can recognize it.

> For anyone, perhaps after Paulo has responded - is there a
> "recommended" way to look at the files, so that the desired encoding
> will be obeyed ?  And if so, do the files need some markings to say
> "use this encoding" ?  When unicode works (and I think I somewhat
> understand Norbert's comment about what people use in Japan) it is
> at least semi-legible (although in Han probably at least ugly, and
> maybe not really legible) whereas the examples above are just
> garbage.
>
Z. Wagner