[Issue 11783] Make std.datetime unittesting faster

[d-bugmail at puremagic.com]

https://d.puremagic.com/issues/show_bug.cgi?id=11783


Walter Bright <bugzilla at digitalmars.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bugzilla at digitalmars.com


--- Comment #11 from Walter Bright <bugzilla at digitalmars.com> 2013-12-22 11:00:48 PST ---
I want to chime in and agree that 38 seconds to unittest datetime is too slow.
I want to add that unittesting datetime consumes more than 500Mb of memory (how
much I don't know, but it exceeds all memory and fails on one of my older
boxes).

It just seems patently absurd to me that date/time consumes more than 500Mb of
memory.

I agree that the datetime tests should be thorough. But I see no reason why
they can't be both thorough and run in a reasonable amount of time in a
reasonable amount of space.

Let's assume that exceptions are the bottleneck. I posit that there's nothing
about date/time calculations that require exceptions. I suggest that datetime
separate out date/time calculations into separate nothrow functions, and have
the throwing happen in a layer over the top of them. Then have the unittests
test the non-throwing calculation engine, and you'll only need a handful to
check to see that the throwing works.

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------