Nickolay Olshevsky

This is something that would always happen as time flies: expiring keys, updated CI images and runners, changed timestamps and so on. So I don't see a cure from such...

> maybe just `memset(&s, 0, sizeof some_struct_t);` ? That will not work within the migration to C++, when, for instance, list variables will become a std::vector or std::map.

Please not yet, if possible - I have some postponed code which will definitely conflict with those changes (see https://github.com/ni4/rnp/pull/4)

@antonsviridenko Yep, those are created automatically once version tag is pushed to the GitHub. On the releases page we can edit tag to fill it up with data, attach files...

> @ni4 a couple of things which I thought of doing here in addition to this bare minimum are armoring support in flags and a meaningful test. What else would...

@andrey-utkin Let me re-check it first. It could be related to already fixed issue in RNP v0.16.0 (which didn't get into the Thunderbird release yet, as far as I know).

Thunderbird 91.6.1 uses RNP v0.15.2, and, actually, that should fix the issue I suspected (fixed via PR #1531 and included into the v0.15.2 release). Also RNP marks that uid as...

@kaie Do you have any suggestions? As I checked, RNP v0.15.2 marks all uids from the aforementioned key as valid.

Most of CLI tests were written to check compatibility with GnuPG on initial steps, processing large amounts of random data/generating a lot of keys and so on. Right now we...

- PR #1600 reduces CLI tests run time by dropping redundant checks - Interesting observation: rnp_tests with OpenSSL backend runs much more faster, for instance `rnp_tests.test_dsa_roundtrip` runs 173 sec with...