>From a security point of view, my understanding is we're "hanging out to dry" anyway: a package can run a build script that could in theory do virtually anything the user could.
If we accept that the of scripts you can choose to call doesn't seem that much more dangerous. Was security the reason for it not being a good idea? On Monday, 9 June 2014 15:14:47 UTC+1, Kevin Squire wrote: > > I've found it useful. I've often installed python packages simply for the > scripts they include, and it's convenient that they are accessible > immediately from the command line. > > In contrast, the R packages I've used rarely provide self-contained > scripts, and I've found it frustrating to download an R package and have to > load up R and execute a series of commands that (in my opinion) would have > more naturally been executed as a script (with command line arguments, > etc.). > > Cheers, > Kevin > > > On Mon, Jun 9, 2014 at 7:04 AM, Stefan Karpinski <[email protected] > <javascript:>> wrote: > >> My question is if this is a good idea or not. I'm not really sure. >> >> >> On Mon, Jun 9, 2014 at 5:57 AM, Samuel Colvin <[email protected] >> <javascript:>> wrote: >> >>> In python pip packages you can declare "bin" scripts which become >>> available in path to execute. It's a simple feature but it allows you to >>> use the system to distribute simple "programs" as well as libraries. >>> >>> I can see there's no obvious way of providing the same functionality in >>> julia, but if I did have a package with a script that people might want to >>> execute, how would I proceed? >>> >>> Is there any plan for an optional directory in packages which would be >>> added to PATH? >>> >>> Perhaps the best approach for now is just to give some direction in the >>> README, eg.: >>> >>> "just run `cp ~/.julia/v0.3/packname/script.jl .` to copy the script to >>> the local directory and run it from there." ??? >>> >> >> >
