Thanks for the detailed analysis; we will check to tomorrow why this was changed.

Using clang for Windows is not tested or suggested. Thus low priority.

Now also done for libksba.

Note: With rK3e029a4ed0059116febe05924a14009ca622e3c5 autogen.sh was extended with targets to run converage scripts. However, there is no real documentation and it is questionable whether autogen.sh is the right place - after all the coverage stuff needs more support in a project which can't be easily done in a general way in autogen.sh. These chnages to autogen.sh have meanwhile be replaced by standard autogen.sh.

Yes, please.

Different pinentries provide different options. The curses pinentry does not have that external password manager thingy. Mixing GUI and tty use seems to be a rare case.

Didn't you had this problem with 1.10.x ? I can't remember that we changed anything at theat code for years. The only recent change was rC656ca459e which should not have changed anything. Maybe the problem is elsewhere.

Frankly, having environ declared in unistd.h is a glibc convenience and other systems likey don't have it. easy to fix, though.

Please create a bug report instead of pointing to a specific commit. Sure the commit is helpful but w/o a bug report we can't track a bug. I'll do it for you this time ;-)

It is trivial append a bogus signature and would thuns inhibit to check the expected signature.

Note that the origin stored for the key is for example required if a key is updated by fingerprint. In that case we don't known from which user ID to take the origin.

Done for all branches,