On Fri, Oct 29, 2021 at 05:20:23PM +0200, Remi Tricot-Le Breton wrote: > > > Rémi, am I missing something or is it just that this code snippet indeed > > > has a bug that was not spotted by the regtests (which I'm fine with, > > > they're regression tests, not unit tests seeking 100% coverage) ? > > > > You are not missing anything. This is indeed broken, because no reg-test > > covers calling the `http_auth_bearer` fetch with a string parameter. > > > > Thus the tests always go into the 'else' part where `get_http_auth()` > > handles the parsing. > > > > The correct fix would then of course be that the parsing logic is moved > > into a dedicated function (ideally based on the ist API) and not copied > > into several different functions. > > > > I retract my patch. > > > > Remi: Shall I file an issue to track the duplicated parsing logic or > > will you handle this based on this ML thread? > > Willy and I talked about this together but I forgot to add the list when I > first replied to him. Some patches are waiting for a review already, > including yours.
And I've now applied your v3, which passes the regtests correctly this time, thanks! Willy