Re: fix quotes for @samp

[Per Bothner]

Karl Berry wrote:
> Regarding HTML quotes, ‘ and ’ are surely better in theory,
> but I seem to remember that I tried putting them in not too long ago and
> got a lot of complaints.  But I guess we can try it again, maybe things
> are better.  I see that we're doing “ and ” now, so what the
> heck.  I installed it.

Thanks.

> Actually, we could go even further.  Your patches change the output for
> @samp, but how about making `...' in straight text output
> ‘...’?

You mean if the texinfo source contains `...' we should translate it to 
‘...’?  I guess we could but it seems awfully magic.  There 
is the risk that ` and ' really are used as separate characters, rather 
than symmetric quotes.

> Regarding quotes in Info output, I really don't think changing the
> default output is worth the inevitable trouble.  I also don't think the
> output should depend on the locale of the person who runs makeinfo; Info
> output gets shipped around, installed on different computers, etc.

I agree - but see below.

> Instead, seems to me that the right thing to do is make the output
> respect @documentencoding, and support utf-8 there.  

But according to the texinfo manual, @documentencoding specifies the 
encoding of the *input* texinfo document.  What does that have to do 
with the whether to use Unicode in the into *output* - which as you said 
is shipped all over the place.

> and support utf-8 there.  Of course this is
> not exactly trivial, but it is the path we have started down.  E.g.,
> makeinfo --enable-encoding on a document with @documentencoding
> ISO-8859-1 will produce 8-bit Info files.

This seems wrong, unless info-reading programs can determine which 
encoding is used.  This is worse than depending on the locale of the 
person running makeinfo.  Depending on the encoding specified in the 
texinfo document for the default encoding can never be the right thing.

> Regarding quotes in general, I expect you will be unhappy to hear that
> after the latest discussion on gnu-prog-discuss (a couple months ago by
> now), I expended a fair amount of effort discussing it with rms, Bruno,
> Paul Eggert, and others, and the consensus was that we had to stick with
> `...' for default (C locale) output, that nothing else was widely
> supported enough

Plain '...' is widely supported, and correct, though non-optimal.

> + In the C locale, GNU programs should stick to plain ASCII for quotation
> + characters in messages to users: preferably 0x60 (@samp{`}) for left
> + quotes and 0x27 (@samp{'}) for right quotes.  It is ok, but not
> + required, to use locale-specific quotes in other locales.

This is, I believe, wrong.  It should be "preferably 0x27 (@samp{'}) for 
both left and right quotes".  See the link in my original message 
(http://www.cl.cam.ac.uk/~mgk25/ucs/quotes.html).  Gcc now emits 'xxx' 
in the C locale.
--
        --Per Bothner
address@hidden   http://per.bothner.com/