[Groff] Re: problem with -Thtml and groff_char.man

[Gaius Mulley]

Hi Werner,

> PS: Where can I get a *complete* list of character entities for
>    various HTML versions?

html2.0   http://www.w3.org/MarkUp/html-spec/html-spec_13.html#SEC13
html3.2   http://www.w3.org/TR/REC-html32
html4.0   http://www.w3.org/TR/html4/sgml/entities.html

>  . The first line of the character table (starting with `!') is in
>    the same line as the table header.

a bug - I will look into this. Incidently you mentioned image clipping
problems some time ago. Was this incorrect right clipping or vertical
alignment problems? Or both :-)

>  . I think there is no ``' character entity in HTML -- my
>    browser doesn't display it, and I couldn't find a reference.

hmm yes you're right it doesn't exist - and it should be removed..

>  . It is not clear to me what the following comment and translations
>    in html.tmac are really doing:

>      .\" Now set any characters defined in devps/S but not in
>      .\" devhtml/S to nul -- these are generated by eqn but not used
>      .\" by grohtml.  grops generated images during the alternative
>      .\" pass.
>      .if !c\[*y] .tr \[*y]
>      .if !c\[fa] .tr \[fa]
>      ...

>    This causes a lot of empty entries in groff_char.man which
>    shouldn't be there at all.  It looks like that you tried to fix a
>    bug which introduced another one.

yes I think this may have occurred. I meant to nullify many of the
characters used when constructing equations and generating html-4.0
text.  For example when processing Lester's papers gtroff would give
many undefined character warnings when it was performing the html pass
(not so for the PS pass obviously).  Perhaps this was the wrong
solution - maybe it would be cleaner to check the suppression \O value
- and if we are suppressing output then we should also ignore
undefined characters? I hope this doesn't have any nasty knock on
effects though?

> Finally, I ask you to have a look at ec.tmac: I suggest that we use
> this method to introduce support for various HTML versions (but no
> styles); the standard character entities should be HTML 4.0 (or newer
> if this exists); specifying, say, `-mhtml32' selects character entity
> support for HTML 3.2, etc.

yes will do this (ec.tmac) is new to me!

Gaius