"Make Intron" sometimes creates intron with non-canonical splice sites

[GreggHelt2]

"Make Intron" is supposed to take as a starting point the genome coord clicked on in an exon, then search around that point for the closest canonical splice sites, and effectively create an intron by splitting the exon and truncate the two new exons to the canonical splice sites. But when zoomed in to the base level, sometimes makes an intron with non-canonical splice sites instead.

[GreggHelt2]

This is not tied to zoom level. It happens when, within the selected exon, there is no canonical splice site 5' and/or 3' of the point clicked. In which case the server, instead of giving up entirely, makes a 2bp intron with non-canonical splice site(s) instead. So not really a bug, but still confusing for users.
Also, while checking out this issue, found something that I feel is a bug. When trying to create an intron within an exon, if the nearest splice site is right at the edge of the exon, can end up creating a zero-length exon. See attached image.