[GHC] #9315: Weird change in allocation numbers of T9203

[GHC]

#9315: Weird change in allocation numbers of T9203
-------------------------------------+-------------------------------------
              Reporter:  nomeata     |             Owner:  simonmar
                  Type:  bug         |            Status:  new
              Priority:  normal      |         Milestone:
             Component:  Runtime     |           Version:  7.9
  System                             |          Keywords:
            Resolution:              |  Operating System:  Unknown/Multiple
Differential Revisions:              |   Type of failure:  None/Unknown
          Architecture:              |         Test Case:
  Unknown/Multiple                   |          Blocking:
            Difficulty:  Unknown     |
            Blocked By:              |
       Related Tickets:              |
-------------------------------------+-------------------------------------

Comment (by simonmar):

 Replying to [comment:21 nomeata]:
 > > validate takes a minute longer
 >
 > Out of more than half an hour. It’s not like validate is fast enough so
 that people can actually sit and watch it. And no matter whether its 36 or
 37 minutes, I guess most will stop running validate themselves and rather
 leave that to Phabricator or other CI stuff, precisely because validate
 takes so long and its convenient to let someone else do it for you.

 But then where do you stop?  By your arguments we should be making
 validate run the full testsuite, which takes 3 hours or so (IIRC).  When
 we first made validate it took 15 mins, now it takes more than double
 that, and we arrived here not by huge leaps, but by adding one minute at a
 time.  So every time I see this happening, I like to question
 (vigorously!) whether that extra minute is paying for itself.  In this
 case I'm not at all convinced.

 The point of validate is to catch bugs that will block other developers if
 they get committed: build breakage or serious bugs, whereas everything
 else we catch with the nightly builds,  If there's a feeling that we
 should change this, then let's take that discussion to the mailing list.
 It's not at all clear-cut (which is why we're having this discussion!).

 What I'd like to see instead is one of the following:

  * Skip the test unless the libraries were built with -O2
  * Skip the test *if* the libraries were built with -O2
  * Have different results according to the optimisation levels

 All of which fix the perceived problem without introducing the regression
 in validate time.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9315#comment:23>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler