3 ms·
> But with the schema parser written in C++, things become much simpler. [see article for more context] I think this is debatable part. First, calling Haskell
by drdaeman 13y ago
> But with the schema parser written in C++, things become much simpler. [see article for more context]
I think this is debatable part.
First, calling Haskell from Python is quite easy. I've wanted that once and things went surprisingly well, see yourself: https://github.com/drdaeman/haskell-library-ffi-example https://github.com/drdaeman/haskell-library-ffi-example
Second, not everyone has C++ compiler installed and readily available (although, it's true that it's far more likely that there's already a C++ compiler than Haskell one). But in the end, I don't think it really matters whenever one does `$package-install-command g++` or `$package-install-command ghc`, as, I believe, neither comes on out of box installs on most popular non source-based GNU distros, and certainly not on Windows. I don't know how things are in *BSD land.
- kentonv 13y agoDo you really want two garbage collectors running in one process? Isn't one already one too many? :) Seriously, though, yeah, it's definitely debatable. If I had been loving writing my code in Haskell, the challenges of FFI probably wouldn't have been enough to sway me.
- Argorak 13y agoHm. If not, you have the problem of making sure that the calling runtime handles memory correctly for the non-GCed space, so it is definitely debatable. E.g. RMagick was famous for suffering from problems of that kind.