Better error message when enrollment rate vectors are longer than `T - minfup`

[nanxstats]

When the length of the enrollment rates vector gamma/R is accidentally set to be longer than T - minfup, for example:

library(gsDesign)

gsSurv(
  k = 2,
  test.type = 4,
  alpha = 0.025,
  beta = 0.1,
  astar = 0,
  timing = 0.75,
  sfu = sfLDOF,
  sfupar = c(0),
  sfl = sfHSD,
  sflpar = c(0),
  lambdaC = log(2)/20,
  hr = 0.65,
  hr0 = 1,
  eta = -log(1-0.02)/18,
  gamma = c(1, 6, 10, 20, 30),
  R = rep(1, 5),
  S = NULL,
  T = 12,
  minfup = 8,
  ratio = 1
)

This yields the cryptic error message:

Error in `rownames<-`(`*tmp*`, value = nameR) : 
  attempt to set 'rownames' on an object with no dimensions

Perhaps the reason is obvious (discrete time unit?), we should probably handle this exception in the logic explicitly.

[nanxstats]

This error becomes more prevalent since the breaking change on default values for T and minfup in gsDesign 3.6.0. I got another inquiry today about this error because some old code can't run - obviously, T and minfup were not explicitly specified before in the gsSurv() call, so they inherited the new defaults which cannot satisfy the constraint.

[nanxstats]

Should do the assertion in both gsSurv() and nSurv()