mkin/tests, branch v0.9.50.3

Fix issues for release

2020-10-08T07:31:35Z

mkinfit: Ignore unused components of state.ini

2020-07-17T14:27:46Z

Improve tests

2020-07-15T12:16:29Z

- Reduce significant digits for the objective function output in mkinfit(..., quiet = FALSE) as R and R-devel gave different output on my system - Add makefile target 'devtest' for testing with R-devel, in order to fix problems showing up with R-devel on Travis

Improve handling of (partially) failing fits

2020-07-15T10:30:39Z

Test plots on r-devel only

2020-06-15T08:27:19Z

following the arguments of Xavier Robin https://github.com/r-lib/vdiffr/issues/86#issuecomment-636447231

Fix a bug, update for current vdiffr

2020-06-15T07:05:27Z

Vignette FOCUS_L failed as I had introduced a bug in the handling of warnings. Current vdiffr only runs visual tests if R < 4.1.0, skipping r-devel for now, see https://github.com/r-lib/vdiffr/commit/630a29d013361fd63fea242f531e2db6aef37919

Improve handling of warnings, reorganize tests

2020-05-29T14:05:11Z

Warn if standardized residuals are unlikely normal

2020-05-29T13:03:04Z

This revealed a bug in the data returned in mkinfit$data in the case of the d_3 algorithm, which also affected the residual plot - the data from the direct fitting was not returned even if this was the better method.

use_of_ff = "max" for shorthand models like "SFO"

2020-05-28T10:24:32Z

Use all cores per default, confint tolerance

2020-05-26T16:52:01Z

Also, use more intelligent starting values for the variance of the random effects for saemix. While this does not appear to speed up the convergence, it shows where this variance is greatly reduced by using mixed-effects models as opposed to the separate independent fits.