Fix documentation of ar[f/b]_urandom

doc/source/arb.rst

@@ -370,15 +370,11 @@ Random number generation
     that representing the endpoints as exact rational numbers would
     cause overflows.
 
-.. function:: void arb_randtest_uniform(arb_t x, flint_rand_t state, slong prec)
+.. function:: void arb_urandom(arb_t x, flint_rand_t state, slong prec, arf_rnd_t rnd)
 
-    Sets the midpoint and radius of *x* to uniformly distributed random numbers
-    in `[0, 1)`.
-
-.. function:: void arb_randtest_uniform_exact(arb_t x, flint_rand_t state, slong prec)
-
-    Sets the midpoint of *x* to a uniformly distributed random number in
-    `[0, 1)` whose radius is `0`.
+    Sets *x* to a uniformly distributed random number in the interval
+    `[0, 1]`. The method uses rounding from integers to floats based on the
+    rounding mode *rnd*.
 
 Radius and interval operations
 -------------------------------------------------------------------------------

doc/source/arf.rst

@@ -482,9 +482,11 @@ Random number generation
     Identical to :func:`arf_randtest`, except that the output occasionally
     is set to an infinity or NaN.
 
-.. function:: void arf_randtest_uniform(arf_t res, flint_rand_t state, slong bits)
+.. function:: void arf_urandom(arf_t res, flint_rand_t state, slong bits, arf_rnd_t rnd)
 
-    Sets *res* to a uniformly distributed random number in `[0, 1)`.
+    Sets *res* to a uniformly distributed random number in the interval
+    `[0, 1]`. The method uses rounding from integers to floats based on the
+    rounding mode *rnd*.
 
 Input and output
 -------------------------------------------------------------------------------

doc/source/mag.rst

@@ -257,10 +257,6 @@ Random generation
 
     Like :func:`mag_randtest`, but also sometimes sets *res* to infinity.
 
-.. function:: void mag_randtest_uniform(mag_t res, flint_rand_t state)
-
-    Sets *res* to a uniformly distributed random number in `[0, 1)`.
-
 Arithmetic
 -------------------------------------------------------------------------------