Re: [newbie] Problems with character output.

• Previous message: Michele Dondi: "Re: subtracting strings"
• In reply to: Reven: "[newbie] Problems with character output."
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Sat, 9 Oct 2004 13:10:28 +0100

On Sat, 9 Oct 2004, Reven wrote:

> I've installed ActivePerl on win XP and I'm having some problems.
> I've tried documentation at activestate but found nothing on this
> topic.

Your problem is in using a command window (which by default is
effectively providing an MS-DOS environment).

[...]
> $u = "á"; # The value of this string is an "a" acute
[...]
> gives no result. Instead of an "a tilde"

I think you mean "a-acute" (in iso-8859-1 or windows-1252 coding, that
would be 0xE1)

> I get a Greek Beta.

Here's a clue. Visit
http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/PC/
and inspect the CP437 (USA national DOS codepage) or CP850
(multinational DOS codepage) tables, where you will find that 0xE1
represents "small letter sharp-s" (the German double-s character),
which has fooled you by looking rather like a Greek beta.

> I've also tried "iso-8859-1" and "windows-1252", but to the same
> effect. I'm quite lost. This is just a wild guess: could there be
> any problem with the console itself?

In a sense, yes. This is a longstanding misunderstanding, which
Microsoft have not put much effort into documenting for the end user:
right from the start of MS Windows, the DOS command window has
implemented the MS-DOS character "code pages", which pre-date current
8-bit character coding conventions (such as iso-8859-x and
windows-125y for various x and y).

> Could I be *so* lucky to find a bug?

At best it could be described as "documented as broken", but the
documentation is very hard to find if you don't know what you're
looking for. I haven't studied this issue specifically in XP, but I
first met it in Win95, and again later (and somewhat differently) in
Win/NT4.

You may be able to fool it by changing your DOS window font from its
initial setting (does yours say "Raster Fonts", as my Win2000 system
is doing?) to e.g "Lucida Console". However, doing that globally
might have some unpleasant effects on any software which was actually
designed to run under DOS (for example, DOS box-drawing characters
will come out funky).

There may be some useful terms in this posting that you can use to
Google for other answers related to this issue. Good luck.

• Previous message: Michele Dondi: "Re: subtracting strings"
• In reply to: Reven: "[newbie] Problems with character output."
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]