Printf: avoid converting to Float64 or BigFloat

[jmkuhn]

Fixes #178

Currently Printf is converting DecimalFloatingPoint values to Float64 which causes issues such as

julia> @sprintf("%.36f", d64"0.1")
"0.100000000000000005551115123125782702"

julia> @sprintf("%e", floatmax(Dec64))
"Inf"

With this PR

julia> @sprintf("%.36f", d64"0.1")
"0.100000000000000000000000000000000000"

julia> @sprintf("%e", floatmax(Dec64))
"1.000000e+385"

Currently %a/%A formatting isn't implemented. I'll do that in a separate PR.

@quinnj

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.81%. Comparing base (87b39dd) to head (19cf5e0).
Report is 3 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master     #187       +/-   ##
===========================================
+ Coverage   80.45%   95.81%   +15.35%     
===========================================
  Files           1        2        +1     
  Lines         435      597      +162     
===========================================
+ Hits          350      572      +222     
+ Misses         85       25       -60     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stevengj]

LGTM. Merge when ready.

[jmkuhn]

Looks like I no longer have write access to merge. It is ready to merge.

[stevengj]

I added you to the repo so that you should be able to merge in the future.