numerical problems on latest travis run

[christophsax]

Error: all.equal(R, old$R, tol = 1e-06) are not all TRUE

perhaps a new R version? need to check

0.09schecking tests ... ERROR
Running the tests in ‘tests/test-all.R’ failed.
Last 13 lines of output:
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  'denton-cholette' removes the transient movement at the beginning of the series and is preferable to the original 'denton' method in most cases.
  Error: all.equal(R, old$R, tol = 1e-06) are not all TRUE
  Execution halted checking R code for possible problems ... NOTE
plot.td: no visible global function definition for ‘par’
plot.td: no visible global function definition for ‘grid’
plot.td: no visible global function definition for ‘mtext’
SubAggregation: no visible global function definition for ‘tail’
Undefined global functions or variables:
  grid mtext par tail
Consider adding
  importFrom("graphics", "grid", "mtext", "par")
  importFrom("utils", "tail")
to your NAMESPACE file. 

[christophsax]

Deviations of this size are fine, increasing tolerance to 1e-5

  > all.equal(R, old$R, tol = 1e-6)
  [1] "Component “y2q”: Component “lit_log”: Component “rss”: Mean relative difference: 1.063779e-06"         
  [2] "Component “y2q”: Component “lit_log”: Component “s_2”: Mean relative difference: 1.063779e-06"         
  [3] "Component “y2q”: Component “lit_log”: Component “s_2_gls”: Mean relative difference: 1.063779e-06"     
  [4] "Component “y2q”: Component “lit_log”: Component “tss”: Mean relative difference: 1.063782e-06"         
  [5] "Component “y2q”: Component “lit_log”: Component “rho”: Mean relative difference: 1.064845e-06"         
  [6] "Component “q2m”: Component “lit_rss”: Component “coefficients”: Mean relative difference: 1.384461e-06"
  [7] "Component “q2m”: Component “lit_rss”: Component “se”: Mean relative difference: 7.106941e-06"          
  [8] "Component “q2m”: Component “lit_rss”: Component “residuals”: Mean relative difference: 7.660566e-06"   
  [9] "Component “q2m”: Component “lit_rss”: Component “rho”: Mean relative difference: 3.165895e-06"  

  >   all.equal(R$q2m$lit_rss$rho, old$R$q2m$lit_rss$rho, tol = 1e-6)
  [1] "Mean relative difference: 3.165895e-06"

[christophsax]

which fixes travis...