If anyone is interested, Typeclassopedia pointed me to Composing Monads by
Jones and Duponcheel (1993), which contains exactly my implementation along
with some other nice patterns for composing Monads via Traversable.sequence
(called swap in the paper) and related operators. It would be
Hi everyone,
I wanted to announce that FP Complete is now running a Jenkins job to build
Stackage with GHC 7.8. You can see the current results in the relevant
Github issue[1]. Essentially, we're still trying to get version bounds
updated so that a build can commence.
I'd like to ask two things
Niklas Hambüchen mail at nh2.me writes:
In sets, the order does not matter, while for nub it does.
Let's be careful here!. Niklas, when you say order, do you mean:
* the _ordering_ from the Ord instance? Or
* the relative sequence of elements in the list?
... the fact that Set is used
On 13/10/13 21:42, AntC wrote:
Niklas Hambüchen mail at nh2.me writes:
In sets, the order does not matter, while for nub it does.
Let's be careful here!. Niklas, when you say order, do you mean:
* the _ordering_ from the Ord instance? Or
* the relative sequence of elements in the list?
Niklas Hambüchen mail at nh2.me writes:
On 13/10/13 21:42, AntC wrote:
...
If you use the Set library, that fact may be very visible!
Because Set re-sequences the whole list, as per its Ord instance.
But List.nub preserves the list sequence
(except for omitting duplicates).
I
On 14/10/13 03:20, AntC wrote:
Thanks Niklas, I hadn't spotted those benchmarks back in July.
No worries :)
I'm surprised at that result for singletons
(and for very small numbers of elements which are in fact each different).
I think one of the main reasons for the performance difference
Niklas Hambüchen mail at nh2.me writes:
On 14/10/13 03:20, AntC wrote:
...
Then here's a further possible optimisation, instead of making
separate calls to `member` and `insert`:
This I understand again. Where do you get insert' from? containers
doesn't seem to have it. Do you
