In a school-based setting, most examiners break up testing. I have tested adolescents for 3.5 hours straight, but, even for them, I usually try to break it up into 1.5 hour sessions. I don't think I've ever tested an 8-year-old for a session longer than 2 hours, some of which would have been relatively light activities, like an initial interview, or drawing tasks.
OTOH, it is not uncommon for clinic-based examiners to sit young children for 3.5 hour sessions, probably because they have less access to them.
The WJ is a very lengthy test, especially if you do the whole thing, and especially for high-functioning children. Typically, I give breaks, and monitor attention and effort closely, especially after the first hour.
So it's not an unheard of session length for an 8-year-old, but it's not optimal either, and many examiners would have split it into two or more sessions. Sometimes the examiner isn't someone who is building-based, though, and there is pressure to get the whole thing done in one day, to avoid a second site visit. Also, teachers usually hate having students pulled, which may influence the decision to split up testing. (Once you have the child out of the classroom, you want to make the best of it, in case you don't get another opportunity.) None of these realities justify creating a situation where the results are less than optimal performance, but they are part of the pragmatic details examiners deal with.
