I analyzed this issue. To edit the text of a cell, we place a textbox control over its text and activate it. To finish the editing (one of the several ways), you can click anywhere outside of this textbox control. This allows the user to finish editing by clicking outside of the cell text, which is good in the case when the cell contains other parts - such as cell image. When the user clicks the cell image, the editing is finished. Clicking the cell's horizontal grid line at the bottom is also the case.
However, according to another iGrid rule related to the editing process, the grid should activate editing if the selected cell is clicked. That explains why the cell is put into edit mode again. But as you wrote, this is not a correct behavior and may cause side effects and problems.
I agree with you. But looking at all this, we should pose the following question: what actions and related events should be done/raised in this scenario? What is your opinion?