And the oddity from Friday becomes even more with a fix that comes by. Given I've been dealing with, RegExp, "expected", I said till evening. But an independent test script shows things are coming up from elsewhere.
If you last evening's post, the matchers returned a false negative every alternate run. Pulled the flow out to a script out of the environment only to find it is consistent. Add to that the fact that the run returns a consistent response when gets loaded with each request.
As such, with this run, out of all options but two. Instead of going the regex route, used a simpler [at first glance] indexOf and reversed the order. Switched a RegExp().exec(email) to a email.indexOf() to check presence. This IS odd how the RegExp engine works.
Yet another funny observation while at it. Did a process.hrtime based profiling of the two flows. Contrary to assumption, the RegExp took 18.6ms against 0.268ms with the indexOf version. Mind had a minor blowup to say the least.
Now that the issue gets closed, this, to the skeptical self means something is a major miss with the request loop. Not much bandwidth at hand to be honest so as much as I don't wish to, have to log and continue for now. Not to forget, praying against hopes this doesn't screw concurrency at some point.