switch order of setting ymin and ymax

[cshanahan1]

This fixes an issue where back-to-back calls to set_limits and get_limits were producing different results for y limits. For example,

viewer.set_limits(x_min=0, x_max=20, y_min=0, y_max=20)
limits = viewer.get_limits()
(0, 20.0, 3.855, 16.145).

For whatever reason, switching the order that ymin and ymax are set fixes this issue. I'm not sure if this entirely fixes the zoom/resize issue noted in #3369, but it resolves this round tripping that should work. There is still a deeper issue going on here.

[rosteen]

I tried running this test on main and it also passes there - is there a way to make one that fails on main?

[cshanahan1]

Huh, the same test fails in a notebook on main. Maybe this requires the image to actually be rendered?

edit: The test does pass in the notebook if i don't do imviz.show()

[rosteen]

Hm, yeah, seems like it requires actual rendering of the app then. You can probably just remove the test then, no need to run it if it's not checking anything? 🤷

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.11%. Comparing base (6868bb5) to head (5445a76).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3371   +/-   ##
=======================================
  Coverage   88.11%   88.11%           
=======================================
  Files         127      127           
  Lines       19574    19574           
=======================================
  Hits        17248    17248           
  Misses       2326     2326           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rosteen]

I'm fine with merging this, but it perhaps warrants a deeper investigation at some point about why this matters.