FOREMOST: Applicants should enter the actual date the PR left Canada, and the actual date the PR entered Canada (date that a border official actually authorizes entry into Canada, which if that is after midnight can mean a day later than when the plane landed on the tarmac). ALL exit and entry dates, ACCURATELY.
I concur in observation that IRCC processing agents readily understand passport stamps and what they mean. In particular, there is no reason to worry about how IRCC will interpret stamps made by other countries which are dated between dates the applicant reported exiting Canada and the date the applicant reports next entering Canada.
Indeed, such stamps as you describe are often ERRONEOUSLY relied on, by applicants, when they are filling in the dates in the presence calculation. This is a common error. No big deal so long as the applicant has a good buffer. But it is always best to be precisely ACCURATE.
An entry stamp into another country NEVER documents the date the PR exited Canada. It only documents a date the PR was for sure in that other country. Note, for example, PRs leaving Canada on an overnight Trans-Pacific flight often do NOT actually arrive at their destination until TWO calendar days later.
In other instances, PRs sometimes have holdovers in countries where their passport is not stamped, so the first date of entry stamped in the passport can be multiple days later than the date the PR actually left Canada.
The applicant's obligation is to accurately report the date the applicant actually left Canada. ALL dates, all dates ACCURATELY.
Also note that for many PRs applying for citizenship there may be several trips abroad for which there are NO stamps at all in the passport.
IRCC figures it out. It mostly looks at passport stamps to see if there is something indicating an inconsistency or incongruity. Minor mistakes are, ordinarily, no big deal. So an applicant who mistakenly relied on the date of entry into another country to report the date of exit, again as long as there is enough of a buffer over the minimum that should NOT cause a problem . . . but of course, again, the more accurate the better.
