make the saveat and tstop handling match OrdinaryDiffEq more closely

[oscardssmith]

and remove bug where we would pretend to hit tstops when not actually doing so. Specifically, the check of while integrator.tdir * (integrator.t - first(integrator.opts.tstops)) < -1e6eps() meant that tstops would be ignored if the solver happened to be close to the time that the solver happened to step. I'm not 100% sure that the issue the comment was warning about is gone (partly because the comment appears to have been cut off in the middle), but I suspect that the if integrator.t == integrator.tprev check that I added a few months ago may have fixed it.

[ChrisRackauckas]

Tests really don't seem to like this 😓

[oscardssmith]

yeah. I was missing a few signs. This will hopefully work now.

[codecov]

Codecov Report

Attention: 3 lines in your changes are missing coverage. Please review.

Comparison is base (81fb59a) 81.80% compared to head (7d21bfb) 81.66%.
Report is 4 commits behind head on master.

Files	Patch %	Lines
src/common_interface/integrator_utils.jl	33.33%	2 Missing ⚠️
src/common_interface/solve.jl	94.73%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #436      +/-   ##
==========================================
- Coverage   81.80%   81.66%   -0.14%     
==========================================
  Files          11       11              
  Lines        1440     1440              
==========================================
- Hits         1178     1176       -2     
- Misses        262      264       +2     

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