GitLab

Not used for anything yet.

It works, but just making sure it works even with wildcards.

When they are regenerated by make (not during initial configure run),
they were not set executable, which was useless. This way they are set
executable twice during the initial configure-make-make check, but at
last it works.

The folder contains both generated and directly written files, so we
need to handle them separately, because of when we have different
builddir than srcdir.

after it returns false.
okayed on jabber, and I don't think we need a separate ticket/branch for
this as it's just a small documentation update.

Instead of EXTRA_DIST, this allows make to regenerate them on .in
change.

Just to bring in changes and solve conflicts, there was some refactoring
elsewhere.

Conflicts:
	src/lib/datasrc/database.cc
	src/lib/datasrc/datasrc_messages.mes
	src/lib/datasrc/tests/database_unittest.cc

Conflicts:
	src/lib/datasrc/database.cc
	src/lib/datasrc/database.h
	src/lib/datasrc/sqlite3_accessor.cc
	src/lib/datasrc/sqlite3_accessor.h
	src/lib/datasrc/tests/database_unittest.cc
	src/lib/datasrc/tests/sqlite3_accessor_unittest.cc

mention "in cache" (similar to related log messages)
(yes, the ID already mentions it)
okayed on jabber

Noticed on the distcheck builder.
I didn't test the distcheck, just the dist target.