I've attached the patches. The first patch should be uncontroversial; you may want to consider the others more carefully. All are against the master branch of the tor-browser-build repository, which I hope is right. I haven't tested the output, since I can't get rbm working.
@@ -204,13 +203,13 @@ if [ -L "$myname" ]; then # make that function accessible to shell scripts. # If realpath is available, use it; it Does The Right Thing.- possibly_my_real_name="`realpath "$myname" 2>/dev/null`"- if [ "$?" -eq 0 ]; then+ if myname="$(realpath "$myname" 2>/dev/null)"; then myname="$possibly_my_real_name" else # realpath is not available; hopefully readlink -f works.- myname="`readlink -f "$myname" 2>/dev/null`"- if [ "$?" -ne 0 ]; then+ if myname="$(readlink -f "$myname" 2>/dev/null)"; then+ :+ else # Ugh. complain "start-tor-browser cannot be run using a symlink on this operating system." fi
We are removing the line creating the variable possibly_my_real_name, but we are still using it with myname="$possibly_my_real_name" inside the if, so it does not look correct.
Trac: Keywords: TorBrowserTeam201910R deleted, TorBrowserTeam201910 added Status: needs_review to needs_revision
Patch 3 (Stop using a heredoc in start-tor-browser) and 5 (Put curly quotes inside single quotes, so shellcheck believes them) look good to me, so I merged them to master as commits c5f30eb31a39a61fe300389fff1ff7adecb8b163 and c14d73615ba8d9e12ae29f0168babc0ecbe55052. I also updated the subject lines to include this bug number.
Patches 2 and 4 also look good to me, but they are changing the same areas as patch 1, so I will merge them after patch 1.