This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] fix PR ada/42554 by dropping -c from ranlib on darwin10 and later

On Sat, Feb 06, 2010 at 04:09:15PM -0800, Mike Stump wrote:
> On Feb 3, 2010, at 8:20 PM, Jack Howarth wrote:
>> Okay for gcc trunk?
> Ok.

   Actually, I am having second thoughts about this patch.
While it doesn't produce any failures in the testsuites,
I am worried about untested cases where the -static-libgfortran
flag is used. Looking through the testresults posting from
the time frame where Geoff introduced this patch, I didn't see
any g77 test cases that it eliminated so I suspect the failures
he was addressing were outside of the normal testsuite.
   The optimal solution of course would be for Apple to revisit
the failure of the darwin10 linker not to properly ignore
duplicate common symbols from static libraries. My understanding
is that the fix present in Xcode 3.1 was regressed out of 3.2
due to a conflict with another project. Perhaps an alternative
fix could be found for Xcode 3.2.2? 
   The other alternatives are to 1) attempt to prevent the usage
of '-c' with ranlib in just the ada and libada subdirectories or
2) have the ada developers properly fix their Makefiles so that
the entire object file list of libgnat isn't constantly relinked
when building ada utilities.
    Could you ping radar 6320843 to see if it could be addressed
in the near term?

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]