# This is a test function for `cachematrix.R`

#

# Test init:

# 1. Create the cached matrix

# 2. Calculate its inverse (inv1) --> this should be slow

#

# Test cases:

# 1. Is the result really an inverse metrix? It should be true that `inv(A)*A = I`.

# 2. Recalculate the inverse (inv2) --> matrix has not changed, it should be

# the same as inv1 and the computations should be very fast,

# i.e., this is the only case when "getting cached data" appears

# 3. Change the matrix A and recalculate the inverse --> matrix has changed,

# however the data are the same, therefore the computation

# should be slow, but the result should be the same as inv1

# 4. Change the matrix again, but this time generate new data with a different

# seed --> matrix has changed, the computation should be slow and the value

# should be different from inv1

#

testCacheMatrix <- function() {

source("cachematrix.R");

set.seed(1);

A <- makeCacheMatrix(matrix(rnorm(100), 10, 10));

inv1 <- cacheSolve(A);

if(sum((inv1 %*% A$get()) - diag(10)) > 2e-9) { # 2e-9 == zeroTol

print("ERROR! inv1 != inverse of A");

} else {

print("OK.");

}

inv2 <- cacheSolve(A);

if(!identical(inv1, inv2)) {

print("ERROR! inv1 != inv2");

} else {

print("OK.");

}

A$set(A$get());

inv3 <- cacheSolve(A);

if(!identical(inv1, inv3)) {

print("ERROR! inv1 != inv3");

} else {

print("OK.");

}

set.seed(2);

A$set(matrix(rnorm(1000), 10, 10));

inv4 <- cacheSolve(A);

if(identical(inv1, inv4)) {

print("ERROR! inv1 == inv3");

} else {

print("OK.");

}

}