I have a `get-summary` function that builds stats and returns them as a web
service. Under the hood, it calls quite a few map, group-by, filter, etc.
functions.
I’m experimenting with transducers, and `sequence xform` does the trick
most of the time. But I want to understand `eduction` use
>
> => CompilerException java.lang.Exception: namespace 'foo.bar' not found
> after loading '/foo/bar', compiling:(*cider-repl foo*:65:7)
>
An error like this usually means that the ns declaration does not match the
expected namespace name, based on the file path.
Also be aware that
Their respective docstrings give a hint here:
Sequence – “returns a lazy sequence”.
Eduction – “returns a reducible/iterable …”.
(type (sequence identity [1 2 3])) => clojure.lang.LazySeq
(type (eduction identity [1 2 3])) => clojure.core.Eduction
If you look at that cheshire.generate/generate
If you do have namespace names that do not correspond with the file name
they are placed in, in a Clojure/Java files, Eastwood can find them for you
quickly. Eastwood doesn't analyze ClojureScript files, though.
Andy
https://github.com/jonase/eastwood
On Fri, Dec 22, 2017 at 9:41 AM, Stuart
Even though evaluation and realization of sequences is lazy, they cache
their result after evaluation. If you consume an object from the `sequence`
function more than once, the work is only done once. Sequences are iterable
and reducible, but not as efficiently because of the caching and
Ah, ok. I guess I was thrown off because evaluating sequence and eduction
results both produce output at the REPL, but only sequence can be output as
JSON at the REPL.
On Friday, December 22, 2017 at 1:46:59 PM UTC-5, Sean Corfield wrote:
>
> Their respective docstrings give a hint here:
>
>
On Fri, 22 Dec 2017, at 18:41, Stuart Sierra wrote:
> Also be aware that AOT-compilation[1] causes all sorts of havoc with
> namespace reloading, because AOT-compiled .class files get read by a
> different ClassLoader.>
Is there any place explaining how exactly Clojure compiles and loads the
