A correction to that last post. While 19th century piano tuners thought they were tuning to equal temperament, they actually weren't, and the place where it went wrong was precisely at the "fifth" between G# and Eb, where the "circle" was supposed to close.

Willis Miller's dissertation

The guitar might have been more accurately ET at the time, but was a pretty rare instrument then and couldn't have been in tune with the piano.

He also points out that Rameau changed his mind and started advocating an inaccurate and unimplementable version of equal temperament after writing the "Treatise on Harmony". It was no surprise that this later work was completely ignored.

This sounds nice, doesn't it?

Chopin in unequal temperament