What version of gpg-agent and scdaemon are you running?

$ gpg-agent --version
gpg-agent (GnuPG) 2.0.0
...
$ scdaemon --version
scdaemon (GnuPG) 2.0.0
...

Moritz Schulte created a test case for this or a similar problem:

#!/bin/bash

SOCK=$(echo "SCD GETINFO socket_name" | gpg-connect-agent | awk '{ print $2 }')
echo "scdaemon socket is $SOCK"

echo "first run..."

gpg-connect-agent -S $SOCK<<EOF
/definqfile POPUPKEYPADPROMPT /dev/null
/definqfile DISMISSKEYPADPROMPT /dev/null
SERIALNO
SERIALNO
LEARN --force
EOF

echo -e "\n OKAY, now reinsert the smartcard, wait two seconds and press
enter \n"
read a

echo "second run..."

gpg-connect-agent -S $SOCK<<EOF
/definqfile POPUPKEYPADPROMPT /dev/null
/definqfile DISMISSKEYPADPROMPT /dev/null
SERIALNO
SERIALNO
LEARN --force
EOF

Just as a quick follow-up to this script I wrote:

[...]

The output I get on every (approx.) second run is:

moritz@kiste:~/Arbeit/g10/test-cases$ ./scdaemon-bug.sh
scdaemon socket is /tmp/gpg-7DrMzg/S.scdaemon
first run...
S SERIALNO D2760001240101000001000000CF0000 0
OK
S SERIALNO D2760001240101000001000000CF0000 0
OK
S SERIALNO D2760001240101000001000000CF0000 0
S APPTYPE OPENPGP
S EXTCAP gc=1+ki=1+fc=1+pd=0
S DISP-NAME Schulte<<Moritz
S DISP-LANG de
S DISP-SEX 1
S LOGIN-DATA moritz%0A%14F=3%18
S KEY-FPR 1 7F31D6DDEB0EA3E5D6DAB5BC0A7F4D909EC83163
S KEY-FPR 2 DEDB4462D3D618676D4EE78B8EFEA57E4F850554
S KEY-FPR 3 E98CFF3B4601466A79B430327EED887CB2F9878A
S CHV-STATUS +1+254+254+254+3+3+3
S SIG-COUNTER 33
S KEYPAIRINFO BA3F59E10685CCEC226449613E92EE7613551B4F OPENPGP.1
S KEYPAIRINFO FE373824F9A6962CE250CB8DDCE97C2660CC3E44 OPENPGP.2
S KEYPAIRINFO A2F466EB816996967D9CB967DC3F86E2C6E0BE67 OPENPGP.3
OK

• OKAY, now reinsert the smartcard, wait two seconds and press enter **

second run...
S SERIALNO D2760001240101000001000000CF0000 0
OK
S SERIALNO D2760001240101000001000000CF0000 0
OK
S SERIALNO D2760001240101000001000000CF0000 0
S APPTYPE OPENPGP
S EXTCAP gc=1+ki=1+fc=1+pd=0
S LOGIN-DATA moritz%0A%14F=3%18
OK
moritz@kiste:~/Arbeit/g10/test-cases$

And that's wrong. Since, in my understanding the second block of status
information should be identical to the first one. But it's stripped.

I cannot reproduce this bug EVERY time i run the script. It's a timing
problem.

I had problems debugging this; in the end I suspected the bug in the
application caching code. IIRC, when i disabled the cache THIS bug was gone.

mo

Moritz, this should be fixed in the current SVN of 2.0.10. Would you mind to
test it?

I cannot reproduce this problem anymore. Neither with the test case
script, nor during ssh authentication with several card-reinsertions.

Thanks as lot!
mo