euclidian distance line 142 in abc.R

[EwenBurban]

Hi,
I found a bug in the calculation of the euclidian distance in the script abc.R.
I detailed the error ( it’s very subtle but the impact is huge, at least for me) in the following R script.
I hope I’ve detailed the bug correctly.

target=iris[1,1:4]
sumstat=iris[-1,1:4]
### wrong way (used in abc package in abc.R script line 142)

wrong_diff = sumstat[,1] - target[1]
### correct way 
correct_diff = sumstat[,1] - target[,1]

print(paste('the wrong way produce only',length(wrong_diff),' value'))
print(paste('the correct way produce',length(correct_diff),' values'))

## the wrong way is similar to write
test= sumstat[,1] - target[1,1]
print(test==wrong_diff)
## so actually, the gfit function calculate the godness of fit as the euclidian distance between the 
## first element of the reference table and the target rather than the mean distance between the reference 
## table and the target