restore best-fit parameter values for high accuracy values of constrained values

[newville]

Description

This makes sure that the best-fit (or current) values are restored to their original values in Parameters.create_uvars.

As shown in #905, values for constraint expressions may be off slightly (around uncertainty or machine precision) without this fix.

Type of Changes

• Bug fix
• New feature
• Refactoring / maintenance
• Documentation / examples

Tested on

Python: 3.10.10 | packaged by conda-forge | (main, Mar 24 2023, 20:17:34) [Clang 14.0.6 ]

lmfit: 1.2.2.post0+g8b2038a.d20230727, scipy: 1.11.1, numpy: 1.25.1, asteval: 0.9.31, uncertainties: 3.1.7

Verification

Have you

• included docstrings that follow PEP 257?

• referenced existing Issue and/or provided relevant link to mailing list?

• verified that existing tests pass locally?

• verified that the documentation builds locally?

• squashed/minimized your commits and written descriptive commit messages?

• added or updated existing tests to cover the changes?
• updated the documentation and/or added an entry to the release notes (doc/whatsnew.rst)?
• added an example?

[codecov]

Codecov Report

Merging #907 (2a70f77) into master (8b2038a) will increase coverage by 0.62%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #907      +/-   ##
==========================================
+ Coverage   92.60%   93.23%   +0.62%     
==========================================
  Files          10       10              
  Lines        3693     3696       +3     
==========================================
+ Hits         3420     3446      +26     
+ Misses        273      250      -23     

Files Changed	Coverage Δ
lmfit/parameter.py	98.48% <100.00%> (+<0.01%)	⬆️

... and 4 files with indirect coverage changes

[newville]

I think this is uncontroversial, so I'll just merge it.