Skip to content

wheel jobs failing in LALSimIMRSEOBNRROMUtilities.c on MacOS runners

@adam-mercer @duncanmmacleod @michael.puerrer @john-veitch there's a weird failure of wheels jobs on MacOS runners (both arm64 and x86_64)

In file included from LALSimIMRSEOBNRv1ROMEffectiveSpin.c:56:
./LALSimIMRSEOBNRROMUtilities.c:455:3: error: implicit declaration of function 'gsl_bspline_eval_nonzero' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
  gsl_bspline_eval_nonzero(eta,  Bx4, &isx, &iex, bwx);
  ^
./LALSimIMRSEOBNRROMUtilities.c:455:3: note: did you mean 'gsl_spline_eval_integ'?
/opt/local/include/gsl/gsl_spline.h:92:1: note: 'gsl_spline_eval_integ' declared here
gsl_spline_eval_integ(const gsl_spline * spline,
^
In file included from LALSimIMRSEOBNRv1ROMEffectiveSpin.c:56:
./LALSimIMRSEOBNRROMUtilities.c:508:3: error: implicit declaration of function 'gsl_bspline_eval_nonzero' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
  gsl_bspline_eval_nonzero(eta, Bx4, &isx, &iex, bwx);
  ^
2 errors generated.

At first I thought it was non-deterministic, since some pipeline jobs are being reported in a weird order. But actually it seems to be due to a recent compiler/dependencies update on these MacOS runners:

success, branch of !2291 (merged) (3 June):

failure, merge of the same branch (4 June):

more failures on a different branch at !2290 (merged), including two retries from this morning:

(but earlier runs 3 June on that same branch still worked, too)

Edited by David Keitel