Re: [Groff] comments

[Werner LEMBERG]

> > While I fully agree that groff's source code files should have
> > much more comments, I'm not really happy with the layout you
> > provide in your patch.  James Clark's code has a certain
> > compactness which I would like to retain
>
> Personal preferences about style obviously differ. Thank you for
> having adjusted it to your preferences.

I hope you can live with it! :-)

> > I also prefer having no variable names in declarations.
>
> Then you need to use 'arg1', 'arg2', 'arg3' to refer to the
> arguments...

Yes.  Normally, this isn't a big problem IMHO.  Functions with more
than, say, three arguments are rare.

> Well, I used "XXX" only when talking about strings, such as ",c".
> And for `XXX' I apologize:

No need to apologize!  I'm a strange, old-fashioned guy doing weird
things for no good reasons...

> in all systems I use nowadays this looks asymmetric and therefore
> ugly, therefore normally I don't even contemplate to use this style.

Hehe, I'm using the marvellous ETL bitmap fonts at size 12x24 within
Emacs.  Here, the glyphs for ` and ' are looking fine for quoting.

> I used the term "parallelogram" to care about slanted glyphs. For
> example, while the backslash glyph in a font may have a width of 7
> pixels, in a slanted version of the same font the width - defined as
> the horizontal dimension of the bounding box - will be only 5 pixels
> or so. Since the get_width() function doesn't take into account the
> slant, it will return 7 pixels.

Well, the artificial slanting done with \S is *extremely* rare.  For
fonts with natural slanting, the bounding box has already been
adjusted to the correct width within the font definition file (while
importing the metrics).

> Here's an additional comment patch, to make get_height() and
> get_depth() clearer to those who already know what "ascent" and
> "descent" is.

Applied, thanks.


    Werner