Add numerous \DeclareFontSeriesChangeRule entries

base/ltfssaxes.dtx

@@ -132,8 +132,9 @@
 % The rules set up use explicit series values not \cs{..default}
 % indirections; my current feeling is that this is in fact better.
 %
-% With 9 weights and 9 width classes this table is getting a bit large
-% in the end (324 entries) but on the other hand it doesn't change and
+% With 9 weight values (from \texttt{ul} to \texttt{ub}) and 9 width values
+% (from \texttt{uc} to \texttt{ux}), this table is getting a bit large in
+% the end (1345 entries), but on the other hand it doesn't change and
 % accessing speed and it is fast this way.
 %
 % We could alternatively split the axis and maintain weight and width
@@ -147,30 +148,40 @@
 %
 %    We make \texttt{m} to reset both weight and width (as this is
 %    how it always worked). To reset just the width \texttt{?m} is
-%    provided and to reset just the weight \texttt{m?}.
+%    provided and to reset just the weight \texttt{m?} is provided.
 %
 % \item
 %
-%   We do support ``\texttt{m}\textit{width}'' and
+%    We do support ``\texttt{m}\textit{width}'' and
 %    ``\textit{weight}\texttt{m}'', e.g., \texttt{mec} to mean ``go to
 %    medium weight and extra-condensed width''. At the end of the
 %    process we automatically drop any leftover \texttt{m} in the
 %    series name (unless it is just a single \texttt{m}).
 %
-%
 % \item
 %
 %    If there is no table entry then the target series is used
-%    unconditionally. This means that any request to set both weight
-%    and width (e.g. \texttt{bx} or \texttt{ulc}) needs no table
-%    entries.
-%    For that reason there are no entries which have a weight+width as
-%    request (i.e., second argument).
+%    unconditionally. This means that we usually do not need entries
+%    where the second and third arguments are identical (unless we want
+%    to use the fourth argument to specify an alternative target series).
+%    In particular, this means:
+%    \begin{itemize}
+%    \item
+%
+%      Any request for \texttt{m} needs no entry, i.e. there are no entries
+%      which have \texttt{m} as second argument.
 %
-%    In particular this is also true for cases involving \texttt{m},
-%    e.g., \texttt{bm} (bold medium width) which automatically gets
-%    reduced result in \texttt{b} or \texttt{mc} (medium weight
-%    condensed) which becomes \texttt{c} as a result.
+%    \item
+%
+%       Any request to set both weight and width (e.g. \texttt{bx} or
+%       \texttt{ulc}) needs no entry. For that reason, there are no entries
+%       which have a weight+width as second argument (except for a few
+%       cases where the fourth argument is used to specify an alternative
+%       target series). In particular this is also true for cases involving
+%       \texttt{m}, e.g. \texttt{bm} (bold medium width) which
+%       automatically gets reduced to \texttt{b} or \texttt{mc} (medium
+%       weight condensed) which becomes \texttt{c} as a result.
+%    \end{itemize}
 %
 % \item
 %
@@ -185,16 +196,6 @@
 %    shape to \texttt{n} and then find that \texttt{bc/n} doesn't
 %    exist either and thus ending up with \texttt{m/n}.
 %
-% \item
-%
-%    Also: while I did set up all nine standard weight values from
-%    \texttt{ul} to \texttt{ub} I only bothered to provide entries for
-%    \texttt{ec},  \texttt{sc}, \texttt{c} and \texttt{x}, because other levels of
-%    compression/expansion are not in any real fonts that I know.
-%
-%    Could and perhaps should be eventually extended to cover the
-%    whole set.
-%
 % \end{itemize}
 %
 %    \begin{macrocode}