Re: Unwrapping Java Character to char

> On Jan 3, 2024, at 23:49, Robert Dodier <[email protected]> =
wrote:
>=20
> Hi Blake, thanks for your reply.
>=20
> I actually don't need to convert to ASCII -- what I really want to do
> is to call a Java method which has a char argument. I got the
> following error:
>=20
> CL-USER(19): (jss:new "org.armedbear.lisp.LispCharacter" #\U2502)
> #<THREAD "interpreter" native {41BFB5BC}>: Debugger invoked on
> condition of type JAVA-EXCEPTION
>  Java exception 'java.lang.NoSuchMethodException:
> LispCharacter(java.lang.Character)'.

Do you need to 1) call an arbitrary Java method with a char argument,
or 2) just programmatically call the org.armedbear.lisp.LispCharacter
constructor to get a Lisp character?

For 1) calling arbitrary Java methods, one should be able to use
JSS to do the following:

    (#0"charValue" (#"valueOf" 'java.lang.Character #\U2502))

The #0 returns the raw value of the method invocation without
attempting to convert through the Lisp/Java FFI.

For 2) why can=E2=80=99t you use CODE-CHAR
  =20
    (char-code #\U2502)
    =3D> 9474
    (code-char 9474)=20
    =3D>#\=E2=94=82


> It appears that the Lisp character #\U2502 has been converted to a
> Java Character, but that's not acceptable to the
> org.armedbear.lisp.LispCharacter constructor, which is declared to
> take a char argument; see line 70 of
> src/org/armedbear/lisp/LispCharacter.java in the current version
> (commit bba779e).

Indeed it does change back to a java.lang.Character instance, even
when I use the method to return the raw Java type.  I cant figure
out a way to coerce calling arguments to a primitive char.  Maybe
I haven't found the right invocation yet, but otherwise we will have
to fix the implementation.

>=20
> Now a complicating factor is that the LispCharacter(char) constructor
> is declared private -- I don't know if that's the actual problem, and
> the error message about the argument type is misleading. Are Java
> private methods, variables, and constructors visible from Lisp?

JSS:NEW should be able to invoke private constructors.  That is one
of the features which distinguishes it from JAVA:JNEW.

--=20
"A screaming comes across the sky.  It has happened before but there is =
nothing=20
to compare to it now."