[GHC] #9370: OPTIONS_GHC (or its absence) leaks between modules compiled in parallel

[GHC]

#9370: OPTIONS_GHC (or its absence) leaks between modules compiled in parallel
-------------------------------------+-------------------------------------
              Reporter:  carter      |            Owner:
                  Type:  bug         |           Status:  new
              Priority:  normal      |        Milestone:  7.10.1
             Component:  Compiler    |          Version:  7.8.3
            Resolution:              |         Keywords:  parmake
      Operating System:              |     Architecture:  Unknown/Multiple
  Unknown/Multiple                   |       Difficulty:  Unknown
       Type of failure:  Compile-    |       Blocked By:
  time performance bug               |  Related Tickets:  #910 #9221
             Test Case:              |
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Brilliant work!

 I have no idea about how the leakage is taking place... that will need
 someone who knows how the `-j` stuff works in GHC.

 But I ''am'' puzzled by the `OPTIONS_GHC` flags in
 `Text.XmlHtml.HTML.Meta`.  That pretty much switches off all optimisation
 in GHC, and you should need a jolly good reason do to that. Indeed, the
 necessity of doing so signals either a bug in GHC, or else some bad
 rewrite rules that go into a loop.  It would be good to know which.

 Is the `OPTIONS_GHC` stuff accompanied with a prominent comment to explain
 such a bizarre options line?

 If not, would someone care to investigate?  (All this is separate from the
 flag leakage question, of course.)

 Simon

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9370#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler