[tldoc] TL21 guide typos

[Karl Berry]

Takuto spotted some remaining typos in the new texlive-en, diff
below. Thanks Takuto. Sorry about this. No urgent need to import these
fixes, but if you want to ...

Goal is to have everything in by Tues (Mar 22) ... --thanks, karl.

--- texlive-en.tex	(revision 58555)
+++ texlive-en.tex	(working copy)
@@ -339,7 +339,7 @@ The default layout is:
       \begin{ttdescription}
         \item [i386-linux] \GNU/Linux binaries (32-bit)
         \item [...]
-        \item [x86\_64-darwin] \MacOSX\ binaries
+        \item [universal-darwin] \MacOSX\ binaries
         \item [x86\_64-linux] \GNU/Linux binaries (64-bit)
         \item [win32] Windows binaries
       \end{ttdescription}
@@ -2756,7 +2756,7 @@ Ken Brown (\pkgname{i386-cygwin}, \pkgname{x86\_64
 Simon Dales (\pkgname{armhf-linux}),
 Johannes Hielscher (\pkgname{aarch64-linux}),
 Akira Kakuto (\pkgname{win32}),
-Dick Koch (\pkgname{x86\_64-darwin}),
+Dick Koch (\pkgname{universal-darwin}),
 Mojca Miklavec (\pkgname{amd64-freebsd},
                 \pkgname{i386-freebsd},
                 \pkgname{x86\_64-darwinlegacy},
@@ -3846,7 +3846,7 @@ invisible to users, but a notable change in distri
 General: \begin{itemize}
 \item Donald Knuth's changes for his 2021 tuneup of \TeX\ and Metafont are
 incorporated
-(\url{https://tug.org/TUGboat/tb42-1/tb130knuth-tuneup21.pdf}. They are
+(\url{https://tug.org/TUGboat/tb42-1/tb130knuth-tuneup21.pdf}). They are
 also available on CTAN as the \code{knuth-dist} and \code{knuth-local}
 packages. As expected, the fixes are for obscure cases and do not affect
 any behavior in practice.
@@ -3876,7 +3876,7 @@ generalized variant of the new \cs{tracingstacklev
 
 MetaPost: \begin{itemize*}
 \item |SOURCE_DATE_EPOCH| environment variable support for reproducible output.
-\item Avoid wrong final \texttt{\%} in mpto.
+\item Avoid wrong final \texttt{\%} in \texttt{mpto}.
 \item Document \texttt{-T} option, other fixes to manual.
 \item Value of \texttt{epsilon} changed in binary and decimal modes, so
 |mp_solve_rising_cubic| works as expected.