For a CW-complex, there's the cellular boundary formula that $$ d_n(e^n_\alpha)=\sum_\beta d_{\alpha\beta}e^{n-1}_\beta $$ where the coefficients $d_{\alpha\beta}$ are the degrees of the map $$ S^{n-1}\to X_{n-1}\to S^{n-1} $$ where the first map is the attaching map of $e^n_\alpha$ to the $n-1$-skeleton, and the second map is collapsing the $X_{n-1}$ except for $e^{n-1}_\beta$.
What if you had a CW-complex consisting of a $0$-cell, with two $1$-cells attaching to it to make a figure-eight, and then attached a $2$-cell to one of the one cells to get a figure-eight with a filled in loop.
If I wanted to compute the coefficient in $d_2(e^2)$ of $e^1_\gamma$ where $e^1_\gamma$ is the $1$-cell I filled in, If I follow the composition above, the second map would collapse the unfilled loop to a point, but then it seems like I end up with $D^2$, not $S^1$. Am I doing something wrong?
You're left with $S^1$. You've already mapped the boundary of the $2$-cell to itself, and the other $1$-cell is collapsed to the $0$-cell.