#2933: LDFLAGS ignored by build system
---------------------------------+------------------------------------------
    Reporter:  duncan            |        Owner:             
        Type:  bug               |       Status:  new        
    Priority:  normal            |    Milestone:  6.10 branch
   Component:  Build System      |      Version:  6.10.1     
    Severity:  normal            |   Resolution:             
    Keywords:                    |   Difficulty:  Unknown    
    Testcase:                    |           Os:  Solaris    
Architecture:  Unknown/Multiple  |  
---------------------------------+------------------------------------------
Comment (by duncan):

 Replying to [comment:1 simonmar]:
 > I'm surprised if `--with-gmp-libraries` doesn't work - as far as I can
 see it should, because we add those directories to the `library-dirs`
 field of the RTS `package.conf`.  Incedentally GMP should really have its
 own package, which would then be a good place to put these options, along
 with any -R options you also want to add.

 Yes. It adds it to `include-dirs` and `library-dirs`. Those translate into
 -I and -L, but for shared libs we also need -R if the dir is not on the
 standard runtime linker path.

 Similarly, `-R` is not used when linking stage2 so stage2 still does not
 work without `LD_LIBRARY_PATH`.

 As for `LDFLAGS`, I'd advocate improving `--with-gmp-libraries` first. It
 is probably enough in most cases.

 So the suggestion is that `--with-gmp-libraries` should add `ld-options:
 -R${GMP_LIB_DIR}` as well. And it should be used for linking stage2.

 I know some people think -R/-rpath is evil, though I'm never sure quite
 why. If a library is not on the standard lib path then surely we need both
 -L and -R?

 So the `LDFLAGS` is just a backup plan in case people come and tell us
 that -R is evil and should not be used by default.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2933#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Reply via email to