[string-templates] Simplify custom processor creation

Tagir Valeev amaembo at gmail.com
Mon Nov 27 17:00:13 UTC 2023


Hello!

As we expect that people will create custom template processors, we can
simplify their lives.

First, it could be common to add a finisher transformation to an existing
processor. I think that for many purposes it would be enough to use STR
processor as a starter, and then create a custom domain object from the
resulting string. This could be simplified if we add a method 'andThen' to
the Processor interface:

default <RR> Processor<RR, E> andThen(Function<R, RR> finisher) {
    Objects.requireNonNull(finisher);
    return st -> finisher.apply(process(st));
}

Second, I think that many processors may want to keep string parts intact
but handle embedded expressions in a special way, effectively replacing the
default 'String.valueOf' behavior of the standard STR processor. We can
provide a way doing this, encapsulating all the ceremony. Something like
this:

static StringTemplate.Processor<String, RuntimeException>
withToString(Function<Object, String> toStringFunction) {
    Objects.requireNonNull(toStringFunction);
    return st -> {
        StringBuilder sb = new StringBuilder();
        Iterator<String> fragIter = st.fragments().iterator();

        for (Object value : st.values()) {
            sb.append(fragIter.next());
            sb.append(toStringFunction.apply(value));
        }

        sb.append(fragIter.next());
        return sb.toString();
    };
}

withToString(String::valueOf) should be equivalent to the STR template (in
functionality, not in performance)

Now, one can easily build interesting things like:

StringTemplate.Processor<Pattern, RuntimeException> REGEXP =
   withToString(obj -> Pattern.quote(obj.toString())).andThen(Pattern::compile);


What do you think?

With best regards,
Tagir Valeev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/amber-spec-observers/attachments/20231127/9c3dd511/attachment.htm>


More information about the amber-spec-observers mailing list