numbering of sub-lemmas

[thosgood]

it seems like the numbering breaks down when there are lemmas inside other lemmas. for example, II.8.3.3 in the PDF (and the original EGA) is listed as II.8.3.5 in the web version, because there are two mini-lemmas inside the proof of II.8.3.2

[ryankeleti]

Nice catch, I'll look into it