Many thanks for your helpful comments. After all them I mostly agree with rhm, who summarizes what I have learnt. Not opening 5♦ was not an agreement constraint. It would have not broken our agreements, I just wanted to keep 3NT on the picture, but I think partner's hand requirements for 3NT are not easier than for 5♦. 6♦ is quite improbable, and probably 5♦ is the long term winner. Once I opened 1♦, my second voice was pass...!, trying not to push them to game, but there was nothing that could avoid it now... My partner only cards where (sadly) QJ♣, exactly 54=108752=6=QJ952, so 5♦ would probably be -2 (not easy to find the club ruff), and the best rebid seeing all cards, but I don't think it is worth the voice anyway. 5♠ would probably make so no big loss, but I wanted to check if my reasoning was very wrong