Re: gen-class/AOT considered harmful (in libraries)???

2019-08-31 Thread Spenser Truex
I'm not sure if it is harmful. On Thursday, August 22, 2019 at 9:59:44 AM UTC-7, Jim foo.bar wrote: > > Hi folks, > > This is a genuine question that has tripped me quite a few times, so I'd > appreciate some closure - pun intended ;) > > Suppose you are developing a library *foo*. In the absence

Re: gen-class/AOT considered harmful (in libraries)???

2019-08-31 Thread Matching Socks
On Saturday, August 31, 2019 at 5:39:18 AM UTC-4, Didier wrote: > > with only the -xyx fns in it. And have those instantly delegate to a call > to another namespace. > The gen-class :impl-ns option is useful here. It achieves the delegation without spreading AOT to the implementing namespace.

Re: gen-class/AOT considered harmful (in libraries)???

2019-08-31 Thread Didier
Clojure does not guarantee binary compatibility between versions, but almost always happens to be compatible. Yet, it's best not to distribute your Clojure libs as compiled classes, much better to do so as source. For a gen-class though, it works like Java. Java guarantees backward compatibilit

Re: Strange behavior with clojure.java.io/copy

2019-08-31 Thread Matching Socks
A bit awkward to clutter a standard library with too much speculative forgiveness of other software's potential failure to follow a spec. Where would it end? Anyway, the best improvement here would help with all InputStream consumers, not just io/copy. If you suffer from a faulty InputStrea

Re: How to use Clojure core specs in my own library with regards to the EPL license?

2019-08-31 Thread Didier
Ya, I will just license the modifications to the specs under EPL. I do feel though specs are a weird grey area. For example, EPL 2.0 says: > Modified Works shall not include works that contain only declarations, > interfaces, types, classes, structures, or files of the Program solely in > each