Custom Query (1630 matches)

The implementation of C_i_greaterp does not throw an error when the first argument is not a number.

See example at ​http://paste.lisp.org/display/99206

I believe this is because it does not check the argument type of x, only y. Instead if x is not a fixnum or flonum, C_i_greaterp just returns #f.

In other words it is missing: else barf(C_BAD_ARGUMENT_TYPE_ERROR, ">", x);

I also believe C_i_nequalp, C_i_greaterp, C_i_lessp, C_i_greater_or_equalp, C_i_less_or_equalp are all affected.

Now that the bootstrap target has been removed (?), instructions for bootstrapping need to be fixed to explain how to bootstrap. Apparently this is as simple as invoking make in a release tarball. But the docs don't explain this nor do they say what the new bootstrap compiler will be called. Also, I would like to know if it is possible to only perform the bootstrap step (as opposed to full build).

On IRC, we were discussing the odd status of irregex, which requires that you (use regex) and (import irregex) to use it. As regex is now just a wrapper around irregex, and irregex has no dependencies on regex, it seemed to us it might be good to grant irregex full unit status.

The patch attached does so without breaking any existing code. To use irregex, you can continue to

(use regex) (import irregex)

or now simply

(use irregex)

for only the irregex API, or even

(use regex irregex)

for both, which is (IMO) less confusing.

Feel free to chime in, anyone.