version 1.0.1
- fixed issue where in some cases rule-violating imputations were done (GH issue #5)
- fixed Rd bug (thanks to Kurt Hornik for pointing out the changes)
- impute_lr now passes 'eps' argument to lintools::subst_value. This resolves
  creation of numerically infeasible systems of linear equations in certain
  cases where a small substitution value (e.g. 1E-12) is multiplied with a
  large coefficient (e.g. 1E5).

version 1.0.0
- In 'impute_lr' there is now an option to choose which sub-methods to use.
- Internal: 'impute_implied' now uses lintools::compact to detect
  inequality pairs that together yield an equality.
- If Fourier-Motzkin imputation fails (usually because of a lack of memory), a
  message is emitted.
- Imputation by pseudoinverse is more robust against edge cases where
  there are no equalities.

version 0.1.4
- package now uses 'tinytest' for testing
- Registered native routines as required by CRAN

version 0.1.3
- fix: 'correct_typos' crashed in presence of certain inequality rules
- fix: 'impute_lr' gave faulty results when using 'var_group' validation 
       rules (thanks to GH user smartie5)
- fix: 'impute_lr' gave faulty results in cases where data contradicts rules 
       (issue #2). Thanks to GH user smartie5
  

version 0.1.2
- bugfix in range-imputation (crashed when both upper and lower limit are -Inf)
- updated a test, and switched off for now to accomodate for updating 'lintools'
- registered native routines as now required by CRAN

version 0.1.1
- bugfix in zero-imputation

version 0.1.0
- initial release