Re: mdocml: Accomodate for \f(Cx formatting.

Re: mdocml: Accomodate for \f(Cx formatting.

[Ingo Schwarze]

Hi,

the meaning of the "C*" font family is not "constant width" in general.
Even in groff, its meaning is output-device dependent, and only a few
devices support it: mainly -Tps and -Thtml, but not -Tascii or -Tdvi.

Even in groff -Tps, its meaning isn't "constant width", but rather
"Adobe Courier".  There are many other -Tps families we do not
support (A, BM, H, HN, N, P, T).  Even when aiming for groff -Tps C*
support only, this patch is incomplete, \f[CBI] is missing.

Hence, i mildly disagree with this patch - but merged it anyway
because it doesn't do that much harm.  I do think, however, that
any font escape except \fR \fI \fB \f1 \f2 \f3 \fP merits a warning
that it's utterly non-portable both regarding different roff
implementations and even regarding different output devices.

I don't consider this an urgent matter at all, but thought
i'd bring this up in case anybody has more thoughts on the
subject...

Yours,
  Ingo


kristaps@mdocml.bsd.lv wrote on Sun, Nov 06, 2011 at 09:43:14AM -0500:

> Log Message:
> -----------
> Accomodate for \f(Cx formatting.  Noted by Andreas Vogele, thanks!
> 
> Modified Files:
> --------------
>     mdocml:
>         mandoc.c
>         roff.7
> 
> Revision Data
> -------------
> Index: mandoc.c
> ===================================================================
> RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/mandoc.c,v
> retrieving revision 1.60
> retrieving revision 1.61
> diff -Lmandoc.c -Lmandoc.c -u -p -r1.60 -r1.61
> --- mandoc.c
> +++ mandoc.c
> @@ -369,8 +369,15 @@ out:
>  
>  	switch (gly) {
>  	case (ESCAPE_FONT):
> -		if (1 != rlim)
> +		/*
> +		 * Pretend that the constant-width font modes are the
> +		 * same as the regular font modes.
> +		 */
> +		if (2 == rlim && 'C' == *rstart)
> +			rstart++;
> +		else if (1 != rlim)
>  			break;
> +
>  		switch (*rstart) {
>  		case ('3'):
>  			/* FALLTHROUGH */
> Index: roff.7
> ===================================================================
> RCS file: /usr/vhosts/mdocml.bsd.lv/cvs/mdocml/roff.7,v
> retrieving revision 1.33
> retrieving revision 1.34
> diff -Lroff.7 -Lroff.7 -u -p -r1.33 -r1.34
> --- roff.7
> +++ roff.7
> @@ -143,6 +143,8 @@ escape followed by an indicator: B (bold
>  (revert to previous mode).
>  A numerical representation 3, 2, or 1 (bold, italic, and regular,
>  respectively) may be used instead.
> +The indicator or numerical representative may be preceded by C
> +(constant-width), which is ignored.
>  .Pp
>  Examples:
>  .Bl -tag -width Ds -offset indent -compact
--
 To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv

Re: mdocml: Accomodate for \f(Cx formatting.

[Kristaps Dzonsons]

On 12/11/2011 23:40, Ingo Schwarze wrote:
> Hi,
>
> the meaning of the "C*" font family is not "constant width" in general.
> Even in groff, its meaning is output-device dependent, and only a few
> devices support it: mainly -Tps and -Thtml, but not -Tascii or -Tdvi.
>
> Even in groff -Tps, its meaning isn't "constant width", but rather
> "Adobe Courier".  There are many other -Tps families we do not
> support (A, BM, H, HN, N, P, T).  Even when aiming for groff -Tps C*
> support only, this patch is incomplete, \f[CBI] is missing.
>
> Hence, i mildly disagree with this patch - but merged it anyway
> because it doesn't do that much harm.  I do think, however, that
> any font escape except \fR \fI \fB \f1 \f2 \f3 \fP merits a warning
> that it's utterly non-portable both regarding different roff
> implementations and even regarding different output devices.
>
> I don't consider this an urgent matter at all, but thought
> i'd bring this up in case anybody has more thoughts on the
> subject...

Ingo,

Hm, you have a good point.  Maybe mention should be omitted from the 
manual to avoid encouraging this bad behaviour?

And while yes, in principle warning about bad formatting tokens is a 
good idea -- there just so many of them!  We should maybe warn when 
people are doing the correct thing. ;)

Thoughts?  (About the manual entry, of course...)

Kristaps
--
 To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv

Re: mdocml: Accomodate for \f(Cx formatting.

[Ingo Schwarze]

Hi Kristaps,

Kristaps Dzonsons wrote on Sun, Nov 13, 2011 at 01:14:34AM +0100:
> On 12/11/2011 23:40, Ingo Schwarze wrote:

>> the meaning of the "C*" font family is not "constant width" in general.
>> Even in groff, its meaning is output-device dependent, and only a few
>> devices support it: mainly -Tps and -Thtml, but not -Tascii or -Tdvi.
>>
>> Even in groff -Tps, its meaning isn't "constant width", but rather
>> "Adobe Courier".  There are many other -Tps families we do not
>> support (A, BM, H, HN, N, P, T).  Even when aiming for groff -Tps C*
>> support only, this patch is incomplete, \f[CBI] is missing.
>>
>> Hence, i mildly disagree with this patch - but merged it anyway
>> because it doesn't do that much harm.  I do think, however, that
>> any font escape except \fR \fI \fB \f1 \f2 \f3 \fP merits a warning
>> that it's utterly non-portable both regarding different roff
>> implementations and even regarding different output devices.
>>
>> I don't consider this an urgent matter at all, but thought
>> i'd bring this up in case anybody has more thoughts on the
>> subject...

> Hm, you have a good point.  Maybe mention should be omitted
> from the manual to avoid encouraging this bad behaviour?

Yes, i think that is a good idea.

We might accept \f(.[BIR123] and drop *any* middle character
from it, in case people use \f(TB or \f(PI, but that's very
low priority.  Then again, such hacks will never be complete,
and just dropping such escapes is fine with me, too.

> And while yes, in principle warning about bad formatting tokens is a
> good idea -- there just so many of them!

Well, at some point i'll probably look into that, it can't be
that hard.  But it's not high priority either.

Yours,
  Ingo
--
 To unsubscribe send an email to tech+unsubscribe@mdocml.bsd.lv