Amos Jeffries wrote:
> Henrik Nordstrom wrote:
>> tor 2009-10-01 klockan 16:47 +0200 skrev Kinkie:
>>> ld: fatal: symbol `void operator delete[](void*)' is multiply-defined:
>>> (file store.o type=FUNC; file SquidNew.o type=FUNC);
>>> ld: fatal: symbol `void*operator new[](unsigned)' is multiply-defined:
>>> (file store.o type=FUNC; file SquidNew.o type=FUNC);
>>> ld: fatal: symbol `void operator delete(void*)' is multiply-defined:
>>> (file debug.o type=FUNC; file SquidNew.o type=FUNC);
>>> ld: fatal: symbol `void*operator new(unsigned)' is multiply-defined:
>>> (file debug.o type=FUNC; file SquidNew.o type=FUNC);
>>> ld: fatal: file processing errors. No output written to ufsdump
>>
>>> I can't find WHERE those may be multiply-defined. Any ideas?
>>
>> include/SquidNew.h
>>
>> Regards
>> Henrik
>>
>
> Defined in Squid.New.h. With an explicit instantiation in SquidNew.o AND
> everywhere else when inlined.
>
> The util.h piece looks like it should be mostly in the compat library
> compat/os/sgi.h with just a wrapped include left there if needed.
>
> ** SquidNew.cc needs some better wrapping tests to ensure the functions
> don't get defined there there when inlined.
>
> And <new> needs to be tested by configure and safety-wrapped as well.
>
> Does the attached patch change anything?
>
> Amos
>
Bah typo. after applying inside include/SquidNew.h you will have to add
the #if in wrapping of <new>
Amos
-- Please be using Current Stable Squid 2.7.STABLE7 or 3.0.STABLE19 Current Beta Squid 3.1.0.14Received on Fri Oct 02 2009 - 01:46:03 MDT
This archive was generated by hypermail 2.2.0 : Fri Oct 02 2009 - 12:00:04 MDT