Errata for "Equidistribution on the modular surface and L-functions",
https://users.renyi.hu/~gharcos/heegner.pdf

The following correction was kindly communicated to me by Alexandru Popa.

In the case of d>0, formula (20) requires an additional sign condition
due to the assumption on infinity types in the beginning of Section 4 in [19]
(appearing later as Assumption 4.2.5).

If J denotes the class of (sqrt{d}) in H_d, then the condition is that
psi(J) should be +1 (resp. -1) when g is even (resp. odd). If this condition
fails, then the left hand side of (20) vanishes automatically, while the right 
hand side of (20) can be non-vanishing.

Note that this correction has no effect on equidistribution. That is, (19)
remains true as a consequence of the quoted subconvexity estimates.