Jul 28, 2016
Jul 13, 2016
micha commented on pull request hoplon/hoplon#139
@micha

i like the idea of using bidi to construct the route parsing functions then using that in a formula cell with the hoplon core route-cell cell. the …

Jul 10, 2016
@micha
  • @micha f08d23f
    Add ICustomAttribute protocol and definterval macro
Jul 10, 2016
Jul 7, 2016
micha commented on issue mcohen01/amazonica#223
@micha

I have a filthy workaround that has been working reliably for me: (defn invoke-lambda [fn-name data & [bbuf?]] (let [coerce (if bbuf? json-bbuf jso…

Jul 6, 2016
Jul 6, 2016
micha commented on issue boot-clj/boot#185
@micha

Perhaps trying option 1 from here: http://stackoverflow.com/a/15409446/3096353 (what you have above is option 2). Perhaps this would be good to add…

micha created tag 0.2.2 at hoplon/boot-hoplon
Jul 5, 2016
Jul 5, 2016
@micha
  • @micha 1a11446
    0.2.2 -- separate ns+ into own task, fix ns+ dep issues
micha created tag 0.2.1 at hoplon/boot-hoplon
Jul 4, 2016
Jul 4, 2016
@micha
  • @micha bb67111
    0.2.1 -- fix issue with ns+ macro resolution, add warnings
Jul 4, 2016
@micha
micha created branch master at micha/boot-aot-shim
Jul 4, 2016
micha created repository micha/boot-aot-shim
Jul 4, 2016
micha created tag 0.2.0 at hoplon/boot-hoplon
Jul 4, 2016
Jul 4, 2016
@micha
  • @micha d44d3e8
    0.2.0 - Add support for .cljs files in addition to .hl
Jun 30, 2016
micha commented on pull request boot-clj/boot#474
@micha

Seems like that's the symptom, not the cause. Those keys will remain in the weak hash map until the pods are collected by GC, not the other way aro…

Jun 30, 2016
micha commented on pull request boot-clj/boot#474
@micha

The pod/pods is a [WeakHashMap][1], shouldn't interfere with GC I don't think. 1: https://docs.oracle.com/javase/7/docs/api/java/util/WeakHashMap.html

Jun 30, 2016
micha commented on pull request boot-clj/boot#474
@micha

The pod pool creates new pods, it doesn't reuse them...

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Or (alter-var-root #'boot.pod/destroy-pod #(comp % my-destroy-function))...

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

You can use set! on the binding, but it just seems like a last resort. If we can solve this by reflection or debugger APIs and without any hooks th…

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Yes, that's why maybe we need both. I mean if I make a task that craetes a pod and my task is loading core.async in the pod, then my task should al…

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Regarding the *hooks* vs. pod/hooks, it might be a "bothism"; we might want both. The *hooks* thing is good when you need to inject finalization pr…

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Possibly... although it seems like there is a more elegant solution somewhere, maybe? Another option to consider is reaping defunct pods out-of-ban…

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Or an atom in boot.pod that you can conj thunks onto. Then destroy-pod could do something like (with-eval-in p (doseq [x @boot.pod/destroy-hooks] (x)…

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

makes sense to me

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Ah, maybe something like a shutdown hook that can be installed in the pod to perform cleanup...destroy-pod would call those before finishing?

Jun 29, 2016
micha commented on pull request boot-clj/boot#474
@micha

Perhaps JMX can be used to find leaks automatically? The pod-pool function already accepts a :destroy option that can be used to finalize things in…

Jun 22, 2016
Jun 22, 2016
@micha
  • @micha f3ad9d7
    Datadog protocol refactor