LYNX-DEV A suggestion for ™

[Alan J. Flavell]

I see that in LYCharUtils.c the entity ™ is converted 
into ® and then re-scanned.

The LYCharSets.c file also refers repeatedly to the trademark (TM)
entity as if it were identical to the registered trademark, circle-R.

But, circle-R and TM are different glyphs, and ™ is the numerical
representation of TM, not of circle-R.  As soon as ™ has been
turned into ® , "information" has been irrevocably lost! 

It's already possible to tailor the LYCharSets.c file to render the
™ entity as desired - by a TM glyph in codes where it's
available, or by an approximation such as (tm) or (TM).  But ™
cannot presently be handled in this way, it always comes out the
same as ® comes out.

I would like to make the following suggestion.

In LYCharUtils, ™ should be turned into ™ - this is just the
same length, so it fits the stated constraint that the string should not
grow in length.  Then it can be re-scanned, and substitued just as
™ gets substituted, for whatever is configured in LYCharSets. 

As a result, any tailoring of the LYCharSets.c file would take effect
equally for ™ as for ™ , which seems logical to me. 

My only remaining problem was that when I tried fixing this in
LYCharUtils.c, I couldn't quite understand the present code, and
I wasn't able to get it working correctly.  Could I perhaps interest
someone who understands the code better to finish this job off, or
to tell me why it cannot be done for some reason that I've overlooked?

Thanks a lot.  (I'm not on the lynx-dev list, so an email copy of any
discussion would be much appreciated) 


;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;