The part of the change submitted in D129224 that enables the folding of strtol calls with a nonnull endptr fails to reflect in the pointer's offset the length of the leading part of the consumed string preceding the first converted digit. The tests for this part of the change were inadvertently disabled so the bug has gone unnoticed both during testing and in review. This change fixes the problem and adjusts the existing tests to let them exercise it.
Since this bug is apparently causing problems downstream I'll plan on committing the fix today and address any subsequent review comments afterwards.
Please use size_t here. (Not that it's really likely to matter here, but for the sake of clarity.)