[Agda] Patch for compilation with GHC-7

kahl at cas.mcmaster.ca kahl at cas.mcmaster.ca
Tue Nov 2 17:47:26 CET 2010


On 02 Nov 2010 13:38, Nils Anders Danielsson <nad at Cs.Nott.AC.UK> wrote:
 > 
 > > Since I am still getting segfaults and GHC-run-time-system errors
 > > with GHC-6.12.3 and Agda-2.2.8, I am now trying GHC-7 for comparison.
 > 
 > Are these errors caused by GHC, or should we try to do something about
 > them? If so, can you please open a ticket on the bug tracker?

As long as I get segfaults and RTS errors, I consider the problem
as being on the side of GHC.

(They already fixed one core-lint error that arises with profiling
 and GHC-6.12, and one RTS sanity problem.)


 > 
 > > (Now if I say ``make install''

I should have said ``./Setup install'', and may have gotten that wrong in a
hurry. I will check this when I get home. I hope that  ``./Setup install''
does not call ``cabal update''...

But I definitely noticed that I had to pass ``-no-test''
to ``darcs record'', since the test wanted to call ``cabal install''.

 > 
 >    make CABAL_OPTIONS="--with-compiler=<path to ghc>" install
 > 
 >    Re: Advice on Multiple GHC installations
 >    http://thread.gmane.org/gmane.comp.lang.haskell.glasgow.user/18399/focus=18403

The fact that you have to point to some mailing list posting
pretty much highlights the main problem I have seen with cabal(-install)
every time I checked so far: lack of complete and precise documentation.
So I am sort of taking the stance: ``If they don't want to tell me exactly
what their tool is going to do if I invoke it, I won't invoke their tool.''

But thanks for the hint!


Wolfram

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 2010-11-01_GHC7_signatures.patch.gz
Type: application/octet-stream
Size: 4946 bytes
Desc: not available
Url : http://lists.chalmers.se/pipermail/agda/attachments/20101102/84a90a1b/2010-11-01_GHC7_signatures.patch.obj
-------------- next part --------------





More information about the Agda mailing list