Scientific notation 1e-3 returns incorrect value exp(1)-3

[Former user]

There's a new bug where using `e' notation with negative powers is mishandled.

Example: (Speedcrunch 0.12) typing 1e-3 leads to Speedcrunch interpreting this as exp(1)-3 which gives answer −2.8171817154095476464e−1

On the other hand positive powers eg 1e4 are not affected.

In early versions 1e-3 would have returned 0.001 as expected.

Please fix back to usual.

[Pol Welter]

Duplicate of #734.