Custom Query (1030 matches)

I wrote some code to exercise CCL's math functions and found that there were cases where intermediate computations were causing floating point overflows.

Functions affected are exp, tan, atan, tanh and atanh:

exp: fails for very large negative real rationals.

tanh: fails for complex arguments with very large real parts.

atanh: fails for numbers of the form 1 + iy for very small y.

tan and atan: analogous failures as for tanh and atanh.

I'm attaching (i) the code to exercise the functions (you just load it into Lisp) and (ii) a file of patches that fix the issues. (Apart from the exp issue, these are present in the fix to 869 - sorry.)

The ability to change the EXTERNAL-FORMAT of an open stream is sometimes useful and CCL defines (SETF STREAM-EXTERNAL-FORMAT) to do this.

ANSI CL doesn't allow an implementation to define SETF functions or expanders on CL function names that aren't explicitly allowed to work with SETF, so this function should be renamed to something like CCL:SET-STREAM-EXTERNAL-FORMAT.

The reader signals an error when the first sequence in the initial contents list isn't a list, but a vector, bitvector or string when parsing a 2-dimensional array literal:

? #2a("ab" #*01)

Error: Reader error on #<CCL::RECORDING-CHARACTER-INPUT-STREAM #x2100CE960D>,

near position 14:

Initial contents for #A is inconsistent with dimensions: #2A("ab" #<SIMPLE-BIT-VECTOR 2>)

While executing: CCL::SIGNAL-READER-ERROR, in process listener(1).

If the first sequence is a list, everything works fine:

? #2a((1 2) "ab" #*10) #2A((1 2) (#\a #\b) (1 0))

Creating the array with make-array also works fine:

? (make-array '(2 2) :initial-contents '("ab" #*10)) #2A((#\a #\b) (1 0))