Fabian Stelzer wrote:
> On 09.02.2022 11:20, Todd Zullinger wrote:
>> Interesting. I do have a patch in my gnupg-2.3 series to
>> reload the gpg agent after changing the trustlist, as the
>> changes were not picked up prior to that. In my case, I was
>> running the tests in an environment where gpg could not
>> prompt me. (It also seems like we should try harder to have
>> the test suite reject such prompts).
>>
>
> Yes, gpg-agent in general can be problematic for the tests. I'm not familiar
> enough with gpg but I don't know if we can get by without it?
With modern gnupg, the secret keyring access is handled by
gpg-agent. So it's no longer optional, which is mildly
unfortunate for automated tests..
>> diff --git a/t/lib-gpg.sh b/t/lib-gpg.sh
>> index 6bc083ca77..38e2c0f4fb 100644
>> --- a/t/lib-gpg.sh
>> +++ b/t/lib-gpg.sh
>> @@ -75,6 +75,7 @@ test_lazy_prereq GPGSM '
>> gpgsm --homedir "${GNUPGHOME}" -K --with-colons |
>> awk -F ":" "/^fpr:/ {printf \"%s S relax\\n\", \$10}" \
>> >"${GNUPGHOME}/trustlist.txt" &&
>> + (gpgconf --reload all || : ) &&
>>
>> echo hello | gpgsm --homedir "${GNUPGHOME}" >/dev/null \
>> -u committer@xxxxxxxxxxx -o /dev/null --sign -
>>
>> --- 8< ---
>
> This patch fixes it for me.
Excellent.
> I have prepared the patch with the simple strstr() matching I can post in a
> bit. I would add your two gpg test lib patches to it if thats ok?
Absolutely. Thank you for working on this and pulling it
together.
Cheers,
--
Todd
