Feedback on the implementation of StringJoiner

[Roel Spilker]

It is already pushed to the lambda repo:
http://hg.openjdk.java.net/lambda/lambda/jdk/annotate/6fb8cefc938c/src/share/classes/java/util/StringJoiner.java

I'm not sure that that's correct. If the prefix and the first added element
are both empty and the infix is not, I would expect the resulting String to
start with the infix, but in this implementation it starts with the first
added non-empty CharSequence. So I think we still need the somethingAdded
field, or need to initialize the value field with null and construct the
value when the first value is added.

Also, the lambda implementation still does the string concatenation in the
constructor.

Roel



On Thu, Dec 13, 2012 at 8:18 PM, Henry Jen <henry.jen at oracle.com> wrote:

> On 12/13/2012 01:37 AM, Remi Forax wrote:
> > Hi Henry,
> > if we don't change the implementation of StringJoiner now (see my mail
> > about my inability to test the stream pipeline),
> > there are at least two things that can be changed in your current
> > implementation.
> >
>
> We don't change as I explained to you, save an intermediate allocate of
> char[] and two copies for a list and copy of references is not necessary
> better.
>
> > All the static final String can be removed because the string are
> > interned so there is no need to store then in a static fields.
> > Even if you think the code is more clear with that and you may be right,
> > it's not a standard practice, the only code that does that is
> > String.java (as far as I know).
> >
> > You don't need the field 'somethingAdded' because you can test the
> > length of the StringBuilder (value).
> >
>
> Agree.
>
> Cheers,
> Henry
>
>
>