Hi Claudio,
I had a look at the code, and it looks that in dfs_search function, where the comment says that "BACTRACKING is happening HERE" there is something wrong (it looks neighbours are not all considered, because as you said next_node returns only one neighbour).
I think that the definition of 
dfs_search (x:xs) l_closed
is wrong. I'd start to write it from the beginning, instead of correcting this code.

Cheers,
Ut

Mail priva di virus. www.avg.com

Il giorno mer 21 ott 2020 alle ore 00:37 Claudio Cesar de Sa <ccs1664@gmail.com> ha scritto:
Hi everyone

Initially, I hope that this list is active yet.  For some days,
I had been trying to do a Depth First Search on a graph from scratch  following something very didactical (clear, elegant - legible, any worry about efficiency).

This DFS keeps a boolean list to mark the nodes already visited and another with current or valid nodes ( the stack of DFS classic).

The code is naive and well documented and is here:

https://github.com/claudiosa/CCS/blob/master/haskell/dfs_graph.hs
(excessively commented)

but unfortunately, it is not working.  My guess is  that the problem starts on new_node function.  I am not sure if in DFS, when a current node gets a next neighbour to stack, it gets all the neighbours. 
In this code, it is getting one node per time.

The functions next_node and one_node seem very non-sense. Could someone help me?

Thanks in advance



claudio

**********************************************************************
***********************************************************************
_______________________________________________
Beginners mailing list
Beginners@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners