Apply suggestions from code review

Co-authored-by: Oscar Dowson <odow@users.noreply.github.com>

Author: odow

Project.toml

@@ -28,7 +28,7 @@ MathOptInterfaceLDLFactorizationsExt = "LDLFactorizations"
 
 [compat]
 BenchmarkTools = "1"
-CliqueTrees = "1.17"
+CliqueTrees = "1.13"
 CodecBzip2 = "0.6, 0.7, 0.8"
 CodecZlib = "0.6, 0.7"
 ForwardDiff = "1"

src/Utilities/sparse_square_root.jl

@@ -44,15 +44,15 @@ If unsuccessful, this function returns `nothing`.
 
 ## Extensions
 
-By default, this function attempts to use a `LinearAlgebra.cholesky`. If that
+By default, this function attempts to use a Cholesky decomposition. If that
 fails, it may optionally use various extension packages.
 
 These extension packages must be loaded before calling `compute_sparse_sqrt`.
 
 The extensions currently supported are:
 
  * The LDL routine in `LDLFactorizations.jl`
- * The pivoted cholesky in `CliqueTrees.jl`
+ * The pivoted Cholesky in `CliqueTrees.jl`
 """
 function compute_sparse_sqrt(Q::AbstractMatrix)
     # There's a big try-catch here because Cholesky can fail even if