add cfloat header (fixes #1161)

[eddelbuettel]

As seen in #1158 and described in #1161, we are apparently missing cloat (or float.h) for numeric limits. This small commit addresses this.

Checklist

• Code compiles correctly
• R CMD check still passes all tests
• Prefereably, new tests were added which fail without the change
• Document the changes by file in ChangeLog

[codecov-commenter]

Codecov Report

Merging #1162 (7bfec31) into master (27f24f0) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1162   +/-   ##
=======================================
  Coverage   97.44%   97.44%           
=======================================
  Files          64       64           
  Lines        2783     2783           
=======================================
  Hits         2712     2712           
  Misses         71       71           

Impacted Files	Coverage Δ
inst/include/RcppCommon.h	100.00% <ø> (ø)

[eddelbuettel]

No objections by anyone so will likely merge this 'soon' ...

[kevinushey]

LGTM as well.

My only question is whether this could cause problems on "certain systems" if the user had used #include <float.h> before #include <Rcpp.h>, but I suspect it will be fine.

[eddelbuettel]

I think all 'library' header files have include guards. Having worked on #1158 I have seen quite a few 'include header sections' in the wild and I don't think there is an issue.

[eddelbuettel]

Oh, and I just ran a full rev.dep check with #include <cfloat> added. I haven't looked at in detail but it does not seem to have set the world on fire :)

[eddelbuettel]

I just committed the summary of that rev.dep run and nothing new came up.

As usual a few unrelated issues at CRAN but by and large the number of packages affected by adding the desired STRICT_R_HEADERS (cf #1158) comes down so moving in the right direction