Scherbius's first patent of rotor system was filed already in 1918, and others were working along similar lines - the Dutchman Hugo Koch and Swede Arvid Damm filed patents for rotor and half-rotor systems on October 7 and 10, 1919. A few years later the American Edward Hebern patented his machine in the US.

The first commercial Enigmas were demonstrated at the 8th Postal Union Congress in Stockholm, Sweden in 1924 leading to the Swedes buying early Enigma C machines for evaluation purposes (ultimately deciding to buy a similar domestic machine from Damm's company). A few years later Polish Biuro Szyfrów (among others) purchased Enigma D machines on the open market, as did several other countries.

So basic knowledge of rotor-ciphers and how the older Enigmas worked was fairly widespread. But the changes made to military models, such as adding the plugboard, were not public. Nor did patent information or commercial models give any insight into the more important details of the machine's working, specifically things like the exact internal wiring of the rotors, which was the key information needed to build a functional equivalent.

The Enigma is a bit unusual in that it'd been sold without restriction. The larger, heavier and generally more secure cipher machines used for wire communications (the T52, T43 and SZ-40/42 machines) had not been sold commercially and their development was shrouded in more secrecy, and far fewer of those machines were built. (Making it all the more remarkable, really, that code breakers who'd never seen the T52 and SZ40 machines were able to construction functional equivalents of them) Which is not to diminish Rejewsky's work in any way; since we know others had tried and failed to do what he did.

(Sources: Bauer Decrypted Secrets: Methods and Maxims of Cryptology, Wik Sveriges Enigmor)