[petsc-dev] Build Fails And Warnings In KSP

[Satish Balay]

On Mon, 21 Dec 2020, Barry Smith wrote:

> 
> 
> > On Dec 21, 2020, at 3:52 PM, Satish Balay <balay at mcs.anl.gov> wrote:
> > 
> > On Mon, 21 Dec 2020, Barry Smith wrote:
> > 
> >> 
> >>  There is a custom fortran stub but the C version does not have a /*@C 
> >> 
> >>  I have also reported directly in the MR other broken things about the merge (MergeTrain is the way to go).
> > 
> > I've been avoiding MergeTrain as it likely doubles the CI load..
> 
>    One per MR which is likely much less than doubles.

Sure - its not double - but is 2-3h for each.. [and retries on any failures would have to redo the whole thin - not just the failed job]

But if I start merge train  on 2 or 3 MRs - then anyone else who might need the CI at that time would have their jobs waiting for 4-9h..

Satish

> 
> 
> > 
> > I have a (potential) fix in branch balay/fix-MR3489 at https://gitlab.com/petsc/petsc/-/merge_requests/3496
> > 
> > Satish
> > 
> >> 
> >>  Barry
> >> 
> >> 
> >>> On Dec 21, 2020, at 3:40 PM, Satish Balay via petsc-dev <petsc-dev at mcs.anl.gov> wrote:
> >>> 
> >>> Hm - this is from https://gitlab.com/petsc/petsc/-/merge_requests/3489
> >>> 
> >>> Maybe I made a mistake in verifying the pipeline tests passing [before merging]
> >>> 
> >>> Matt, Can you check on this? [or I can revert this merge]
> >>> 
> >>> Satish
> >>> 
> >>> On Mon, 21 Dec 2020, Jacob Faibussowitsch wrote:
> >>> 
> >>>> Hello All,
> >>>> 
> >>>> Seems a recent change introduced the following errors (I have tried make clean, make allfortranstubs, and then rebuild):
> >>>> 
> >>>> 
> >>>>    CLINKER master-arch-darwin-c-debug/lib/libpetsc.3.014.2.dylib
> >>>> duplicate symbol '_kspconvergedrateview_' in:
> >>>>   master-arch-darwin-c-debug/obj/ksp/ksp/interface/ftn-auto/itfuncf.o
> >>>>   master-arch-darwin-c-debug/obj/ksp/ksp/interface/ftn-custom/zitfuncf.o
> >>>> ld: 1 duplicate symbol for architecture x86_64
> >>>> clang: error: linker command failed with exit code 1 (use -v to see invocation)
> >>>> gmake[3]: *** [gmakefile:112: master-arch-darwin-c-debug/lib/libpetsc.3.014.2.dylib] Error 1
> >>>> gmake[2]: *** [/Users/jacobfaibussowitsch/NoSync/petsc/lib/petsc/conf/rules:49: libs] Error 2
> >>>> **************************ERROR*************************************
> >>>> Error during compile, check master-arch-darwin-c-debug/lib/petsc/conf/make.log
> >>>> Send it and master-arch-darwin-c-debug/lib/petsc/conf/configure.log to petsc-maint at mcs.anl.gov<mailto:petsc-maint at mcs.anl.gov>
> >>>> ********************************************************************
> >>>> gmake[1]: *** [makefile:40: all] Error 1
> >>>> gmake: *** [GNUmakefile:9: all] Error 2
> >>>> 
> >>>> 
> >>>> And warnings (although they appear harmless):
> >>>> 
> >>>> 
> >>>> src/ksp/ksp/interface/itfunc.c:2283:39: warning: passing 'PetscReal **' (aka 'double **') to parameter of type 'const PetscReal **' (aka 'const double **') discards qualifiers in nested pointer types [-Wincompatible-pointer-types-discards-qualifiers]
> >>>>   ierr = KSPGetResidualHistory(ksp, &hist, &n);CHKERRQ(ierr);
> >>>> 
> >>>> src/ksp/ksp/interface/itfunc.c:2167:64: note: passing argument to parameter 'a' here
> >>>> PetscErrorCode KSPGetResidualHistory(KSP ksp, const PetscReal *a[],PetscInt *na)
> >>>>                                                                                                             ^
> >>>> src/ksp/ksp/interface/itfunc.c:2306:36: warning: passing 'PetscReal **' (aka 'double **') to parameter of type 'const PetscReal **' (aka 'const double **') discards qualifiers in nested pointer types [-Wincompatible-pointer-types-discards-qualifiers]
> >>>>   ierr = KSPGetErrorHistory(ksp, &hist, &n);CHKERRQ(ierr);
> >>>>                                                     ^~~~~
> >>>> /Users/jacobfaibussowitsch/NoSync/petsc/src/ksp/ksp/interface/itfunc.c:2243:61: note: passing argument to parameter 'a' here
> >>>> PetscErrorCode KSPGetErrorHistory(KSP ksp, const PetscReal *a[], PetscInt *na)
> >>>>                                                                                                       ^
> >>>> 2 warnings generated.
> >>>> 
> >>>> Best regards,
> >>>> 
> >>>> Jacob Faibussowitsch
> >>>> (Jacob Fai - booss - oh - vitch)
> >>>> Cell: (312) 694-3391
> >>>> 
> >>>> 
> >>> 
> >> 
> > 
>