Conversation
Codecov Report
@@ Coverage Diff @@
## master #963 +/- ##
==========================================
+ Coverage 86.75% 86.79% +0.04%
==========================================
Files 36 36
Lines 9112 9112
==========================================
+ Hits 7905 7909 +4
+ Misses 1207 1203 -4 |
|
This reverts only a part of #959 and we should keep #942 (comment) in mind |
|
This confirms #942 (comment) so if we bring #959 back in, we need to find a solution to call Either 1) we consider that closed means "it's over completely, nothing else to be done" and we 'clean' everything at that time or 2) we delay the memory cleaning like we do right now and we keep the 'close' and 'deep_clean' concepts separated. Right now, I would say that the architecture follows 2) |
|
I think (2) is dangerous because end users will expect (1). But it's going to be tricky to switch, so I can live with either. |
There are some features that might be needed once the plotter has been closed like |
|
Okay let's revert #959 then to be safe |
@GuillaumeFavelier, I'll leave this up to you since you've started the PR. |
This PR follows #942 (comment), let's see how the CIs react