A quick analysis of the first position shows that
0...e5 was playable in that there was not a fundamentally better choice. Stockfish would prefer
0...Nbd7 but the resulting score is only .2 points better than
0...e5. Those fractions add up after a while, but if one is looking to uncramp one's game, it could be a bargain. 7 hours of grinding shows that neither side gets much of an advantage after
0...e5 1. O-O. The score settled on about .75 in White's favor.
Here's the continuation at 15 minutes per move. I think this shows that 1... e5 is playable and leads to a sharp game. White retains a slight advantage the whole time.
[fen "rn1qkb1r/pp3pp1/2p2np1/4p3/3P4/6P1/PPP1NPBP/R1BQK2R w KQkq - 0 1"]
2.O-O Nbd7 3.b4 Be7 4.Rb1 O-O 5.c4 exd4 6.Nxd4 Nb6 7.b5 c5 8.Nb3 Qxd1 9.Rxd1 Nxc4 10.Bxb7 Rad8 11.Re1 Bd6 *
Here's the second position. I didn't see ALNS's message until this morning, so I thought I might as well post the results. There is a deviation in the moves, probably the result of me using the computer last evening while the program chugged in the background.
[fen "r2qkb1r/pp1n1pp1/2p2np1/4p3/3P4/6P1/PPP1NPBP/R1BQ1RK1 w KQkq - 0 1"]
1.b4 Be7 2.Rb1 O-O 3.c4 exd4 4.Nxd4 Nb6 5.c5 Nbd5 6.Qb3 a5 7.a3 Qd7 8.Bb2 Nc7 9.Rbd1 Nfd5 10.Nc2 Qg4 11.Bf3 Qf5 12.Rfe1 Rfe8 13.Be4 Qg4 14.Qc4 Qh5 *
The machine tallies this position as anywhere from .55 to .70 (through the next 16 moves!) depending on the exact variation being calculated. That is, there is no appreciable change in the score in the entire variation. It would seem
0... e5 is a viable move.