Every inhomogeneity associated with a weldment is a stress riser. Welds themselves are stress risers, especially at the toes. Porosity is a stress riser. Undercut is a stress riser. Slag inclusions are stress risers. Insufficient penetration is a stress riser. Thickness transitions are stress risers. Phase changes are stress risers.
And I do believe that more failures have been associated with thickness transitions, undercut, high entry angle weld toes, and phase changes, than arc strikes. This may be why, though I certainly don't know for sure, most of the ASME codes don't address the problem specifically. Though they do address more often thickness transitions through envelope maximums, and undercut (especially for fatigue regimes).
Not saying that arc strikes are OK, but my line of thinking may explain (IMO) some of the reticence in the codes, and my thought that we may be practicing a little bit too much handrwringing on this issue.