[l2h] \htmlsetstyle / css file question

[Dan Scholnik]

This is something I thought I had figured out a while back, then put
aside, and now I can't remember what I figured out.  I'm trying to add
the line

P	      {  line-height : 2.5ex  }

to the .css file generated by latex2html (and we'll put aside that
fact that I don't have a clue what the P stands for) by using the
html.sty command

\htmlsetstyle[P]{}{ line-height=2.5ex }

What I get is

DIV.P	      {  line-height : 2.5ex  }

Now, I don't really know Perl, but I traced through the latex2html
source (99.2beta6) anyway, and found this block of code:

    print "\n *** Adding document-specific styles *** ";
    while (($env,$style) = each %env_style) {
        if ($env =~ /\./) {
            $env =~ s/\.$//;
            print STYLESHEET "$env\t\t{ $style }\n";
        } elsif ($env =~ /inline|^(text|math)?((tt|rm|sf)(family)?|(up|it|sl|sc)(shape)?|(bf|md)(series)?|normal(font)?)$/) {
            print STYLESHEET "SPAN.$env\t\t{ $style }\n";
        } elsif ($env =~ /\./) {
            print STYLESHEET "$env\t\t{ $style }\n";
        } elsif ($env =~ /^(preform|\w*[Vv]erbatim(star)?)$/) {
            print STYLESHEET "PRE.$env\t\t{ $style }\n";
        } elsif ($env =~ /figure|table|tabular|equation|$array_env_rx/) {
            print STYLESHEET "TABLE.$env\t\t{ $style }\n";
        } else {
            print STYLESHEET "DIV.$env\t\t{ $style }\n";
        }
    }

This is where the DIV comes from.  I'm a bit confused by this, since the
third "if" test is the same as the first.  Further, the first case
removes a trailing '.', but the sub process_htmlstyles already does that,
via a fix that was discussed on the list 5/19/99.  Now if that fix
was not in place, then the above latex code would result in $env="P.",
which would be caught by the first line, have it's '.' stripped, and
I'd get what I want.  But $env="P" doesn't match any of the cases,
and the default DIV.P results.

I can see several ways to change the code to get what I want, but I'd
like to know if I'm missing something fundamental here, and if there
is a "right" way to fix it.

thanks,
Dan Scholnik