[Phoenix] [MSVC] 8 9 and 10 COMDAT error
I am attempting to sort out the bugs and test failures in Boost Phoenix. I have one test (bind_rv_sp_test) which fails at the link stage with the following error on MSVC 8 9 and 10 but not later bind_rv_sp_test.obj : fatal error LNK1179: invalid or corrupt file: duplicate COMDAT '?value@?$result_@P8Y@@AE?AV?$shared_ptr@UX@@@boost@@XZ@?$is_mem_fun_pointer_select@$0A@@detail@boost@@2_NB' I have been researching this error and attempting to find a workaround for Phoenix. Has anyone else working on Boost encountered this error and have any suggestions about it. It does appear to have been around for some years and there are a lot of suggestions but no clear resolution I have found. How important are these compilers, and would it be O.K. to mark this test as expected to fail? I don't have these compilers so for any new tests I am relying on the develop tests to check out experiments. Thank you John
On 17/01/2015 13:04, Fletcher, John P wrote:
bind_rv_sp_test.obj : fatal error LNK1179: invalid or corrupt file: duplicate COMDAT '?value@?$result_@P8Y@@AE?AV?$shared_ptr@UX@@@boost@@XZ@?$is_mem_fun_pointer_select@$0A@@detail@boost@@2_NB'
I have been researching this error and attempting to find a workaround for Phoenix.
Has anyone else working on Boost encountered this error and have any suggestions about it. It does appear to have been around for some years and there are a lot of suggestions but no clear resolution I have found.
A possible workaround might be to try compiling without /Gy and /ZI.
________________________________________ From: Boost [boost-bounces@lists.boost.org] on behalf of Mathias Gaunard [mathias.gaunard@ens-lyon.org] Sent: 17 January 2015 16:24 To: boost@lists.boost.org Subject: Re: [boost] [Phoenix] [MSVC] 8 9 and 10 COMDAT error
Has anyone else working on Boost encountered this error and have any suggestions about it. It does appear to > >have been around for some years and there are a lot of suggestions but no clear resolution I have found.
A possible workaround might be to try compiling without /Gy and /ZI.
Mathias Thank you for this. Unfortunately I don't control how the tests are compiled, so I cannot implement it. I am trying to figure out a workaround which will run. John
participants (2)
-
Fletcher, John P
-
Mathias Gaunard