Now, for example, (plot (density '(0))) works.
Calculated bandwidth is now bounded by both 1e-308, below which
`kde' produces nonsense, and 1e-14 * max absolute value. The latter
bound ensures the bandwidth is wide enough to make a smooth-looking
curve even in the presence of floating-point rounding in the domain,
by ensuring that at least 100 floating-point numbers or so in the
domain get nonzero density.
It's a little weird to use the gap between floating-point numbers
for this, but it ensures density estimates aren't jagged because
of rounding (at least until you zoom in, in some instances), and
it's at least a decent method of estimating bandwidth for single-
sample density estimators.
No joy in mudville. See
http://drdr.racket-lang.org/28175/ for details
Also fix (just introduced) infinite loop in class/c contract-name
implementation and tweak naming stuff a little so tests pass
Fixed:
- class-field-accessor and class-field mutator,
- the interaction between init args contracts and subclasses,
- object=? on wrapped objects (ie if 'this' flows out and
is compared to the object thatn 'new' returns)
- contract-name for class/c
Also, the code now uses impersonator properties internally to track a
class that is wrapped (instead of a wrapper struct). This simplifies
a bunch of places in the code and paves the way for classes that have
a class/c contract to be chapeone-of the original class (altho that
doesn't work yet)
This allows the typechecker to tell the optimizer or
other downstream analyses what expressions to ignore
because they contain non-typechecked code.
Use it to fix handling of `send`
This change adjusts the meaning of the 'getting-started document
category, which was formerly omitted from the start page. Now,
it means a document that is listed at the beginning and at the same
level as other documents.
(Didn't just revert to the old approach to the "getting started"
entry, because it doesn't work right when "racket-index" is installed
without "racket-doc".)
Closes PR 14348
Some places in the editor API accept an object as an argument
where the object's identity is relevant. In that case, `object=?`
must be used instead of `eq?` to work right with contracts.
Use the field in a `snip%` that is otherwise used by a `text%` owner
to store line information. Reusing the field avoids the need for
an `eq?`-based hash table of snips.
