switch_map followed by box_it fails to satisfy CoreObservable - #280
Conversation
Co-authored-by: Shahar "Dawn" Or <mightyiampresence@gmail.com>
CoreObservable
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #280 +/- ##
==========================================
+ Coverage 82.42% 82.58% +0.16%
==========================================
Files 84 84
Lines 11756 11929 +173
==========================================
+ Hits 9690 9852 +162
- Misses 2066 2077 +11 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
Preserve InnerObs on SwitchMapOuterObserver with PhantomData so boxed switch_map keeps the downstream observer relation. Also clean up the reproduction test formatting and unused closure parameters.
|
Thanks for the focused reproduction. The root cause is that I pushed a fix that makes I also kept the reproduction coverage for Verified with: |
|
Hey, @M-Adoo. Thank you so much! |
Hey, this is not a concrete contribution, but an issue reproduction.
I don't suppose you need me to open a corresponding issue, right?
Anyway, @shivaraj-bh and I tried to think how we may be holding it wrong and couldn't.
We added a control group to check our sanity. And variations on what we actually originally hit.
Then we compared between the
FlatMapandSwitchMaptypes and what caught our eye is the sheersize of the
SwitchMapCoreObservableimplementation. I remarked "oh, wow. I wouldn't be surprised if an issue was overlooked there".Anyway, I hope this report is actually useful and suffices for you experts to identify the issue or let us know how we're holding it wrong.
Thank you.