Felix Lechner <felix.lech...@lease-up.com> writes: > On Mon, Apr 5, 2021 at 1:10 PM Russ Allbery <r...@debian.org> wrote:
>> See also Debian Policy 8.4, which explicitly requires this: > Thanks for throwing that into the mix. Obviously, I accept policy and > will implement it, but the approach reminds me of the Goldsboro > switch. [1] For what it's worth, I didn't mean that to be any sort of demand or question about whether you accepted Policy, just a note for awareness since it's in a section of Policy where one might not think to look for this sort of requirement. The Policy discussion is at <https://bugs.debian.org/232448> if you're curious about the background. I think everyone agrees this is a somewhat unexpected corner of Ada and isn't how one would normally expect a compiler on a Linux system to work, although it's apparently related to the implementation strategy for an Ada standard requirement. My opinion at the time (which I think I didn't note on the bug) was that it was unlikely that GNAT was going to change such long-standing behavior however odd it might seem to folks who don't work with Ada regularly, it created interoperability problems if not followed, and the scope was limited to the folks who were packaging for Ada (who were already all intending to follow this policy), so it seemed easiest for Policy to just document it. It does unfortunately make the Lintian logic more complex because it introduces a special case. -- Russ Allbery (r...@debian.org) <https://www.eyrie.org/~eagle/>