Changes between Initial Version and Version 1 of Ticket #1547


Ignore:
Timestamp:
Nov 12, 2007 1:17:06 PM (8 years ago)
Author:
simonpj
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #1547

    • Property Test Case changed from to stm package/tests/conc052
    • Property Milestone changed from 6.8 branch to 6.10 branch
    • Property Summary changed from conc052 core lint errors in profc/profasm ways to Arity can decrease with -prof
  • Ticket #1547 – Description

    initial v1  
    1 The conc052 test (in the stm package) on amd64/Linux is giving core lint errors for the profc/profasm ways. 
     1Something like 
     2{{{ 
     3  f = \p\q.body 
     4  x = scc "foo" f (\y.e) 
     5}}} 
     6shows f having arity 2, and hence x having arity 1.  But when we inline f, we get 
     7{{{ 
     8  x = scc "foo" let p = \x.e in \q.body 
     9}}} 
     10and the cheap-and-cheerful arity discovery function (`exprArity`) detects arity of 0, not 1.  And then !CoreLint complains about the inconsistency of arity and strictness info. 
     11 
     12This is unpleasant but not actually a problem. 
     13 
     14It shows up in the `conc052` test (in the stm package). 
    215 
    316Part of #1546 might be the same problem.