chain()

chain(<attribute list>, <base>[, <arg0>[, ... , <argN>]])

chain() threads a value through a sequence of user-defined attributes — the functional “pipeline”, or thread-first “arrow” (as in Clojure’s ->), pattern. <attribute list> is a space-separated list of [<object>/]<attribute> names.

The <base> value is passed as %0 to the first attribute. The result of each attribute is then passed as %0 to the next, threading one accumulating value along the whole list. Any extra <arg0>, <arg1>, ... are passed unchanged as %1, %2, … to every attribute, so each step can reach the same side inputs. The result of the last attribute is returned. If <attribute list> is empty, <base> is returned unchanged.

chain steps: stopping early, evaluation, and how chain() differs from fold()
chain example: threading a string through two attributes

Steps and fold()

A step may call ibreak() to short-circuit the pipeline: the remaining attributes are skipped, and the value produced by that step is returned. chain() counts as an iteration level, so itext(0) and inum(0) inside a step give the running value and the step number.

Each attribute is evaluated as by ufun(). Object names in the list may not contain spaces (use “me” or a dbref), since spaces separate the attributes.

chain() is close kin to fold(): fold() walks a list of data, passing each element in turn as %0; chain() walks a list of attributes, passing one threaded result as %0 while carrying the same optional arguments to each step.

Example

Example — thread a string through two steps (%0 is the running value, %1 the shared side-arg):

> &WRAP me=%1%0%1
> &SHOUT me=ucstr(%0)!
> say chain(WRAP SHOUT, hello, *)
You say, "*HELLO*!"

WRAP wraps the base “hello” in the side-arg ”*” to make “hello”; that result becomes %0 for SHOUT, which upper-cases it and appends ”!”.

::: seealso