Here is another thing worth reporting. I found that passphrase-repeat is entirely ignored, regardless of the value set.

Do you configured gpg so that you did not get a passphrase confirmation?

Something related seems to still be happening in 2.2.1. make test passes, but here on macOS 10.12.6., my ~/Library/Logs/DiagnosticReports is full of crash reports for scdaemon. As far as I can tell from the timestamps, it looks like scdaemon is *still* getting called by the gpg test suite, even though I built gpg with the flag --disable-scdaemon.

This bug is still present in 2.1.23.

I've found that I can get the test to succeed if I drop --disable-scdaemon from my configure flags. I'm far from qualified to diagnose this, but I suspect that the tests have a bug in which they still try to test the scdaemon despite the presence of --disable-scdaemon in the configure flags.

The "Suppress error for card availability check" (https://dev.gnupg.org/rGa8dd96826f8484c0ae93c954035b95c2a75c80f2) won't cleanly apply to 2.1.22, so my build was without it.

Not only can I not list any issues, I can't do anything at all when I go to "dev.mutt.org". All I see is: "You Shall Not Pass: Restricted Dashboard
You do not have permission to view this object.
Users with the "Can View" capability:
Members of the project "g10code" can take this action."

This is still a problem on OS X 10.11.5. OS X's System Integrity Protection
"feature" is causing that test failure. If S.I.P is disabled there's no problem.

A similar-looking test failure happens in perl
(https://rt.perl.org/Public/Bug/Display.html?id=126706). Perhaps the diagnosis is
the same here.

Also, the OSX community would much appreciate it if you guys would cut a release
containing this fix as soon as possible. We aren't any less likely to misenter our own
passwords than anyone else :)

Neal: It works!

However you should be aware that I had to manually --disable-pinentry-emacs, or else
I got this:

-----8<-----------------------------
Making all in emacs
gcc -DHAVE_CONFIG_H -I. -I.. -I/usr/local/Cellar/gnoopeegee/1.8.0/include -
I/usr/local/Cellar/gnoopeegee/1.8.0/include -I../pinentry -Wall -g -O2 -Wall -
Wcast-align -Wshadow -Wstrict-prototypes -Wformat -Wno-format-y2k -Wformat-security
-W -Wno-sign-compare -Wno-missing-field-initializers -Wdeclaration-after-statement -
Wno-pointer-sign -Wpointer-arith -MT pinentry-emacs.o -MD -MP -MF .deps/pinentry-
emacs.Tpo -c -o pinentry-emacs.o pinentry-emacs.c
mv -f .deps/pinentry-emacs.Tpo .deps/pinentry-emacs.Po
make[2]: * No rule to make target ../assuan/libassuan.a', needed by pinentry-
emacs'. Stop.
make[1]: * [all-recursive] Error 1
make: *** [all] Error 2
------------8<----------------------------

I myself don't use emacs, but probably there are some OSX users who'd want pinentry-
emacs.

Can you still try building with the supplied patch and sending me the scancodes

for delete.

Sorry, I don't understand. I did build with the supplied patch, as I reported below,
and I included the output of /tmp/pinentry-curses-output.txt as instructed.

I attempted to bisect, but when I when back to version 0.8.1, I could not build. I get
stuff I don't understand and don't know how to fix, like:

config.status: error: cannot find input file: `gtk/Makefile.in'

And other stuff that seems related to older checkouts needing different versions of
autoconf/automake. I simply don't have the expertise to deal with that, or the time to
curate multiple versions of autotools. So I can't git bisect.

I did however find that the oldest release version of pinentry that I could build on my
system was 0.7.1, and it also had broken backspace/delete/whatever.

I also tried on OSX 10.5.8, same thing.

I suspect pinentry was always broken on OSX.

It sounds like you are missing some build dependencies. Perhaps something

related to iconv?

Well, I don't know. I'm not a programmer. I am not missing any deps when I build the
released versions of pinentry. I don't know what's different about the git repo
versions.

I made sure gettext was in my path and tried again, having applied your patch. It
compiled.

I then followed your original instructions. I typed:

onetwo<the-key-called-Delete-on-Mac>

and hit OK.

The contents of pinentry-curses-output.txt was:
6f
6e
65
74
77
6f
7f

Needless to say, "backspacing" still failed.

The delete key never did anything, because the cursor is always at the end of

the line. (Delete deletes the character in front of the cursor.) Perhaps you
mean the backspace key.

I mean the key which has the word "delete" on it, as seen in the gallery here:
https://www.apple.com/keyboard/. It is in the same spot, and is labeled "delete",
on all Mac keyboards, not just the model shown in that gallery. They key in
question is the one that Mac users press to simultaneously back up the cursor and
delete whatever char was in the arrive-at position.

Probably now someone will tell me that the key is improperly labeled by Apple :)

The delete key never did anything, because the cursor is always at the end of

the line. (Delete deletes the character in front of the cursor.) Perhaps you
mean the backspace key.

I had to clone the git repo first. Once there, I applied the patch and I ran
autogen.sh. I can't get the resulting configure script to work. Both "./configure"
and "./configure --enable-maintainer-mode" result in the following:

---8<-----------------------------------
checking for ncursesw... no
checking for ncurses... no
checking for initscr in -lncursesw... no
checking for initscr in -lncurses... yes
checking for ncurses include dir... none
./configure: line 8466: syntax error near unexpected token `iconv'
./configure: line 8466: ` AC_LIB_LINKFLAGS_BODY(iconv)'
---8<-----------------------------------