Investigate alternatives to global environment assignment in fitdistdoublecens

[seabbs]

As the title and as flagged in #61. There are two options I see:

• Find a work around
• Submit a issue/PR to fitdistrplus to change the way that it (optionally) finds custom distribution functions.

I think the current implementation is not CRAN compliant and so I think this needs to be resolved before CRAN submission.

[seabbs]

Perhaps just using withr would be enough.

[seabbs]

I tried with withr::with_environment and that didn't seem to work.

[seabbs]

Submit a issue/PR to fitdistrplus to change the way that it (optionally) finds custom distribution functions.

I'm currently favouring this I think

[seabbs]

@Bisaloo any chance you have any good ideas here? I think this is blocking a CRAN release but its also really useful functionality for people who can't/don't want to use CRAN.

[seabbs]

This is the assignment I can't find a way around

[seabbs]

Issue over on fitdistrplus: lbbe-software/fitdistrplus#33

[seabbs]

The problematic line is here: https://github.com/lbbe-software/fitdistrplus/blob/89f00a69b35bc687f6003355354b8ec712b31c74/R/fitdist.R#L34