[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freebsd-hackers
Subject:    Re: anybody from the XFree86 team here ?
From:       "Kaleb S. KEITHLEY" <kaleb () ics ! com>
Date:       1998-09-30 21:32:48
[Download RAW message or body]

sbabkin@dcn.att.com wrote:
> 
> I've tried to
> contact the former X consortium but got no answer. 

Not surprising, given that the X Consortium has been out of business
since the end of 1996. The Open Group took over starting January 1st,
1997.

> I still think that the matter is fairly important
> for X11 i18n, so this is my next try :-)
> 
> I'm working on conversion of TTF fonts into Type1 format. 

I have to wonder why you want to go this way, when at this point in time
just about everyone is using the new TrueType fontserver. And before too
long XFree86 will release servers that can use TTF fonts directly.

> Actually, all
> I need is to get some russian scalable fonts :-)
> 
> The Type1 font library in X11 is INTENTIONALLY BROKEN.

I agree, it sounds like it may be broken. I question your motives for
yelling about it being intentionally broken -- it almost certainly
wasn't intentional, at least not in the nefarious way that your shouting
about it might tend to imply.

> It ignores the encoding specification in the font itself and forces all
> the
> fonts to use Latin1 encoding. I think, I know the reason: there exist
> broken font editors (like Fontografer) that specify Adobe Standard
> encoding but actually use names from the Adobe Latin1 encoding.
> So, the library parses the encoding table well but then it just throws
> out the learned encoding table and uses Latin1.

Sounds like a reasonable analysis. Out of curiosity, what does Adobe's
Type1 specification say, if anything, about it?

> And guess what ? Some of the characters used in the russian fonts
> do not have any names in Adobe Latin1 encoding. 

Amazing. I'd never have guessed. ;-)

> This problem can be fixed by a two-line fix that I propose: let's
> force only Adobe Standard encoded fonts to use Adobe Latin1
> encoding and leave alone the fonts with explicitly defined
> encoding table.
> 
> Of course, I have fixed my local copy of X sources but it does
> not enjoy me at all to do that with each next installation.
> 
> Can it please be fixed in the main X repository ?

Get a bug-report template (the file named "bug-report") from your copy
of the X11R6.x sources, or from the XFree86 sources, or from
ftp://ftp.x.org/pub/R6.4/xc/bug-report (when it's back up). Fill it out
and email it to xbugs@opengroup.org.

> Also, if somebody needs, I have written also the keyboard layouts
> for XFree86 for all 4 russian encodings.

What is it about the character encoding that changes the keyboard
layout? (Hint: Nothing! That the lower eight bits of the keysym aren't
correct for KOI8-R or CP866 is irrelevent.)

What was wrong with the Russian keyboard layout(s) that were already in
both the sample implementation releases and in XFree86's releases?

If there are Russian keyboards whose layout doesn't match any of the
layouts included in either of the releases, then it's worth submitting a
separate bug-report for those.

> And, finally, pushing my luck :-)  : I would like very much to see
> the characters with codes 0x80-0x9F defined as printable in
> xterm. 

In CP866, yes; but in the official GOST KOI-8 and ISO8859-5 they're not
printable (and yes, I know that nobody in Russia uses ISO8859-5).

> Of course, I do that with each copy of xterm I work with,
> but I would like to see that in the standard distribution. Provided
> that they were defined to be ignored in X11R6.1 and earlier,
> that should not hurt anybody. The positive sides of redefinition
> are:
> 1. These codes are used for uppercase russian letters in CP866
> encoding

The sample implementation does not support the old MS-DOS CP866 locale.
To my knowledge XFree86 does not either. They only support ISO8859-5 and
KOI8-R. To properly support a CP866 locale you'll need to do a lot more
than just hack xterm -- Xlib will need a CP866 XLC_LOCALE database, and
Xlib itself will need to be fixed to map the X keysyms to CP866 the same
way it does for KOI8-R. 

XFree86 may decide to add CP866 support -- that's their choice. I do
question the utility of putting it into the sample implementation, which
already has support for the two standard Cyrillic encodings.

My suggestion: If you've got a lot of documents in CP866 -- recode them
(using GNU recode) into KOI8-R or ISO8859-5.

-- 
Kaleb S. KEITHLEY

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message

[prev in list] [next in list] [prev in thread] [next in thread]