CI badge, correct Julia requirement to 1.6 (#9)

* CI badge, test Julia 1.3

* correct julia requirement

.github/workflows/CI.yml

@@ -13,7 +13,7 @@ jobs:
       fail-fast: false
       matrix:
         version:
-          - '1.6'
+          - '1.6' # LTS
           - '1'
 #          - 'nightly'
         os:

Project.toml

@@ -6,12 +6,12 @@ version = "1.2.0"
 xsum_jll = "e979a739-315a-50ee-b437-b496a9503be1"
 
 [compat]
-julia = "1.3"
+julia = "1.6"
 xsum_jll = "2"
 
 [extras]
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 
 [targets]
-test = ["Test", "Random"]
+test = ["Test", "Random"]

README.md

@@ -1,4 +1,5 @@
 # Xsum: Exactly rounded floating-point sums in Julia
+[![CI](https://github.com/JuliaMath/Xsum.jl/workflows/CI/badge.svg)](https://github.com/JuliaMath/Xsum.jl/actions?query=workflow%3ACI)
 
 The Xsum package is a Julia wrapper around Radford Neal's [xsum package](https://gitlab.com/radfordneal/xsum)
 for exactly rounded double-precision floating-point summation.  The [xsum algorithm](https://arxiv.org/abs/1505.05571) takes `n` double precision (`Float64` or smaller) floating-point values as input and computes the "exactly rounded sum" — equivalent to summing the values in *infinite* precision and rounding the result to the nearest `Float64` value.
@@ -27,4 +28,4 @@ so they are less efficient), or negate one in-place with `Xsum.negate!(s)`.
 
 For example, if you wanted to compute an exactly rounded sum of a large vector `x` in parallel, you could call `accumulate!(XAccumulator(), xslice)` on a sequence of *slices*
 (portions) of `x` in parallel, and then combine the sub-accumulators to obtain the
-final sum.
+final sum.