In http://reviews.llvm.org/D9044#163999, @mclow.lists wrote:

> > The new tests don't test the instability of the reservoir sampling 
> > algorithm. Should they?
>
>
> Yes - but I won't hold up accepting this because that test is missing.
>
> The phrase in the standard "stable if and only if"...


Fixed anyway, this is easy.


================
Comment at: include/experimental/algorithm:85
@@ +84,3 @@
+  return _VSTD_LFTS::__sample(
+      __first, __last, __out, __n, __g,
+      typename iterator_traits<_PopulationIterator>::iterator_category(),
----------------
mclow.lists wrote:
> I think that should forward the RNG: 
> `_VSTD::forward<_UniformRandomNumberGenerator>(__g)`
> 
> 
> 
done

================
Comment at: 
test/std/experimental/algorithms/alg.random.sample/sample.stable.pass.cpp:44
@@ +43,3 @@
+    for (int j = 1; j < kSampleSize; ++j)
+      assert(oa[j] > oa[j - 1]);
+  }
----------------
mclow.lists wrote:
> Couldn't this just be a call to `std::is_sorted(first, last)` ?
done

http://reviews.llvm.org/D9044

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/



_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to