search

jeroen / openssl

OpenSSL bindings for R
Other
63 stars 21 forks source link

I was going to do this #1

Closed Ironholds closed 9 years ago

Ironholds commented 9 years ago

...that's my only issue. Seriously, I found this package while searching around for existing bindings to make sure I wasn't duplicating effort.

What would you think of extending it to allow for character string hashing under the various algorithms openSSL supports? At the moment the only thing for this is digest, which isn't vectorised. As a general asset for researchers, "throw the UUIDS in the dataset you want to anonymise into this function and they'll come out very difficult to trace back" would be...useful.

Let me know if you think it's a good idea, and I'll build the necessary code myself - like I said, it's something I've been thinking about for a while.

Ironholds commented 9 years ago

Actually, I spoke too soon - implemented in C! I'm only familiar with the Rcpp binding into the library. Bah.

jeroen commented 9 years ago

Sounds interesting. The C interface to R is not that difficult for simple arguments (strings, numbers). Just look at the example and hadleys intro.

jeroen commented 9 years ago

I like to avoid the Rcpp dependency for packages to keep open the option for code to be adopted by base R.

Ironholds commented 9 years ago

Makes sense! At the moment I'm finding a far bigger blocker is the quality of the openSSL docs.

Ironholds commented 9 years ago

Hrm; I've implemented a R/C++ version here. I'll punt on moving it into C for the time being.

jeroen commented 9 years ago

Mind if I adopt some of your code?

Ironholds commented 9 years ago

For the same purpose? I mean legally, my code is MIT licensed and so you're free to do whatever you want with it. But, I just spent two solid days writing this thing and an associated set of vignettes and the start of a paper, so as a matter of protocol I'd rather you not.

jeroen commented 9 years ago

Oh okay. You should probably pick a license that matches those feelings. I cant promise that I will never use the openssl hashing functions in my package, but I won't look at your code if I do :) Too bad you don't want to collaborate, good luck with the paper :)

jeroen commented 9 years ago

Please let me know if you change your mind though. I think it would be much more efficient (and fun) if we join forces and put our openssl bindings in a single package. I can easily add you as a package author and repository admin. Then you don't have to worry about maintenance and getting things to work on windows, etc (which is a pain) because I've already gone through that :-) I could help out with the paper too if you want. I think it would be a win-win.

I don't think using Rcpp to interface a C library is a very strong argument to duplicate efforts. I'm willing to translate your example to C if you want (which is quite trivial).

Ironholds commented 9 years ago

That sounds perfectly acceptable; my worry was more around me putting in a ton of work and it then vanishing like smoke as you adopted it solo :). I'd suggest translating one of the implementations to C, and I'll then translate the rest - and happy to work on the paper together!