Re: Helpfiles again (was Re: modify functions hierarchy (was: etc.))

[Peter Stephenson <p.stephenson@xxxxxxxxxxx>]

On Mon, 25 Nov 2013 08:56:31 -0800
Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
> On Nov 25,  3:49pm, Peter Stephenson wrote:
> }
> } > (1) Completion/Unix/Command/_systemd: line 3, name of the author.
> } > The character is an 'e' with accent (in latin1 encoding, i.e., ISO8859-1).
> } 
> } Hmmm.... might be safer to fudge it as "e'".  There's not really any
> } need for non-ASCII characters in functions.  (Or simply convert to
> } UTF-8 but I think the simpler the better.)
> 
> Output of "file **/*(.) | egrep 'UTF|ISO' :
> 
> ChangeLog:                                       UTF-8 Unicode English text

I think that's OK.  If we don't need to process a file using a tool
that's worried about locales we get away with it.  I don't see any point
in forbidding UTF-8 in text files.

> Completion/BSD/Command/_portaudit:               ISO-8859 English text
> Completion/Unix/Command/_cdrdao:                 UTF-8 Unicode English text
> Completion/Unix/Command/_git:                    UTF-8 Unicode English text
> Completion/Unix/Command/_growisofs:              UTF-8 Unicode English text

Probably safest to make these ASCII; none of the non-ASCII characters
are actually needed.  The big one here is _git which has a lot of
non-ASCII quotes and "...", but they're not actually necessary, even for
readability (they're all quotes / ellipses with an ASCII lookalike), so
I don't think that's a real loss.

> Etc/ChangeLog-3.0:                               ISO-8859 English text
> LICENCE:                                         ISO-8859 English text
> Src/module.c:                                    ISO-8859 C program text
> Src/Modules/clone.c:                             ISO-8859 C program text
> Src/Modules/example.c:                           ISO-8859 C program text

Should be UTF-8 for consistency, I think.  We'd know by now if a
compiler cared about non-ASCII characters.

> Test/A05execution.ztst:                          ISO-8859 text
> Test/D02glob.ztst:                               ISO-8859 text
> Test/D07multibyte.ztst:                          UTF-8 Unicode English text
> Test/V07pcre.ztst:                               UTF-8 Unicode English text

These are all deliberate: either multibyte tests, or single bytes with
the top bit set, so I think they're OK.  Certainly they're being
explicitly tested.

Shout if you're unhappy or I'll do it tomorrow (UK time).

pws