Thanks! That's really nice. Where it says "If I had been using IJulia, the call to areload() would have been unnecessary", that's really what I'm looking for. I wonder what makes it hard or undesirable to do that at the command line REPL. (If it did, this would be exactly what I was looking for.) In IJulia notebook, it does run the reloaded code upon saving, but doesn't show any output until I've interacted with notebook again in some way. (I'm using Gadfly, and also from the reloaded code don't show up for some reason.)
It seems like Autoreload is mainly intended to support a workflow of developing e.g. a package in a text editor, but still doing anything "interactive" in IJulia, which is why it doesn't quite work for my desire to do *everything* in the text editor. (I wonder if I'm wrong to desire that?) Still, I'm sure Autoreload will be very helpful to me, and I appreciate Jonathan for making it and you for pointing it out to me! David On Wed, Mar 12, 2014 at 8:52 AM, Patrick O'Leary <[email protected]> wrote: > You might take a look at https://github.com/malmaud/Autoreload.jl and see if > it meets your needs. > > > On Wednesday, March 12, 2014 1:20:42 AM UTC-5, David Chudzicki wrote: >> >> Hi-- >> >> The workflow I'd like is to edit a file and see the effect of what I've >> done upon saving. Best would be something like conttest julia mycode.jl, >> which runs the script through Julia whenever it changes, but I'm finding >> this too slow (since loading some of the packages my script uses takes a few >> seconds). >> >> So as an alternative, I thought maybe I should try to include my script in >> a single, continuing Julia session whenever it changes. Can anyone help me >> out with the best way to do that? >> >> I was thinking something like this: >> >> function watch_and_include(filename) >> watch_file((f,e,s) -> { >> print("\n--------\n") >> include("mycode.jl") >> watch_and_include(filename) >> }, >> filename) >> end >> >> watch_and_include("mycode.jl") >> >> ... but that seems kick off the inner watch_and_include too many times! >> >> If anyone has any overall workflow suggestions, I'd appreciate that too. >> >> Thanks, >> David >> > -- David J. Chudzicki blog.davidchudzicki.com [email protected] (518) 366-7303 Data Scientist Kaggle (we're hiring!)
