[Patch,Fortran] PR36158 - Add transformational version of BESSEL_JN intrinsic

Daniel Kraft d@domob.eu
Tue Aug 17 13:24:00 GMT 2010


Tobias Burnus wrote:
>  Fortran 2008 has
>   BESSEL_JN(N, X)
> as elemental function (which is implemented) and
>   BESSEL_JN(N1, N2, X)
> as transformational function - which this patch adds.
> 
> Note: This patch only does the compile-time implementation - a run-time 
> implementation is still pending.
> 
> Build and regtested on x86-64-linux.
> OK for the trunk? (Pending the issue below.)

Ok.  But:

+      if (n < 0 && gfc_notify_std (GFC_STD_GNU, "Extension: Negativ 
argument "
+				    "N at %L", &order->where) == FAILURE)

It's "negative" in English, I think.  The same is later, also.

> Note: I get the following failures, which seem to be unrelated:
> a) gfortran.dg/widechar_intrinsics_5.f90: Fails (coredump) at -O1 but 
> not at -O0
> b) gfortran.dg/trim_optimize_1.f90: Fails in the dump for having more 
> than 2 memmove
> On gcc-testresults, I do not see those failures, but the builds there 
> are also a tad older. I am currently bootstrapping to make sure. (b) 
> seems to be a FE issue (dependency.c, frontend-passes.c) - but I do not 
> have any modifications there; (a) looks like a FE->ME issue, but I do 
> not see how this patch could affect those.

I see (b) but not (a) on my system with the build of my last ASSOCIATE 
patch, but there has been some partial svn update so it may be wrong.

Yours,
Daniel

-- 
http://www.pro-vegan.info/
--
Done:  Arc-Bar-Cav-Ran-Rog-Sam-Tou-Val-Wiz
To go: Hea-Kni-Mon-Pri



More information about the Gcc-patches mailing list