David Craven <da...@craven.ch> writes:
> * gnu/packages/gps.scm (gpsbabel)[arguments]: Require gnu++11. Disable tests
> on all platforms due to rounding error.
> ---
> gnu/packages/gps.scm | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/gnu/packages/gps.scm b/gnu/packages/gps.scm
> index ccd9636..8124990 100644
> --- a/gnu/packages/gps.scm
> +++ b/gnu/packages/gps.scm
> @@ -48,7 +48,9 @@
> "0xf7wmy2m29g2lm8lqc74yf8rf7sxfl3cfwbk7dpf0yf42pb0b6w"))))
> (build-system gnu-build-system)
> (arguments
> - `(#:configure-flags '("--with-zlib=system")
> + `(#:configure-flags
> + '("--with-zlib=system"
> + "CXXFLAGS=-std=gnu++11")
The ' should be under the #, so the two lines above should be moved one
column to the left.
> #:phases
> (modify-phases %standard-phases
> (add-before 'configure 'pre-configure
> @@ -59,7 +61,8 @@
> ;; On i686, 'raymarine.test' fails because of a rounding error:
> ;; <http://hydra.gnu.org/build/133040>. As a workaround, disable
> tests
> ;; on these platforms.
> - #:tests? ,(not (string-prefix? "i686" (%current-system)))))
> + ;; On x86_64 with -std=gnu++11 tests also fail due to rounding error.
> + #:tests? #f))
Please add a FIXME to this comment. We should not be in the habit of
simply disabling test suites that fail and forgetting about them.
I don't see why there should be a rounding error on x86_64. Unlike
i686, double-rounding doesn't happen on that platform, and it's what
developers are mostly testing on.
Anyway, okay for now with the FIXME comment.
Mark
