Paul,
Am 11.02.26 um 11:36 PM schrieb Paul Richard Thomas:
Hi Harald,
I forgot to do the git add. Here it is.
this looks to me like the same as the previous one
minus the intrinsic.texi part and minus the testcase.
Can you please check?
Harald
Paul
On Wed, 11 Feb 2026 at 20:29, Harald Anlauf <[email protected]> wrote:
Hi Paul,
I do not see the promised updated patch here.
Is it only me, or did you forget to attach it?
Best,
Harald
Am 11.02.26 um 10:00 AM schrieb Paul Richard Thomas:
Hello Sandra and Harald,
Many thanks for taking a look at the patch for PR99250. I admit that I
was in something of a rush to get it out of the door so that I could
return to the last few PDT problems.
The use of the KIND argument now works correctly and the testcase has
been completely revamped to reflect this. The testcase runs correctly
using the shared memory gfortran, for which Jerry has just posted a
series of patches. Note that GFORTRAN_NUM_IMAGES=multiple_of_3 is
required.
When converted to .html, intrinsic.texi looks OK and is now ordered correctly.
Regtests OK on FC43/x86_64 mainline. OK to push to mainline?
Regards
Paul
On Mon, 9 Feb 2026 at 23:35, Sandra Loosemore <[email protected]> wrote:
On 2/9/26 09:54, Paul Richard Thomas wrote:
Hi All,
This is a boilerplate implementation of the COSHAPE intrinsic. The
testcase is placed in the main gfortran.dg directory so that it can
make us of the option -fcoarry=lib to test the number of references to
_gfortran_caf_num_images (One to set no_images and to for the upper
bound of the last codimension).
Regtests on FC43/x86_64. OK for mainline?
Paul
PS Could somebody please check the intrinsic.texi entries?
Ack....
diff --git a/gcc/fortran/intrinsic.texi b/gcc/fortran/intrinsic.texi
index cf81791b8b3..accd7825e38 100644
--- a/gcc/fortran/intrinsic.texi
+++ b/gcc/fortran/intrinsic.texi
@@ -127,6 +127,7 @@ Some basic guidelines for editing this document:
* @code{COS}: COS, Cosine function
* @code{COSD}: COSD, Cosine function, degrees
* @code{COSH}: COSH, Hyperbolic cosine function
+* @code{COSHAPE}: COSHAPE, Determine the coshape of a array
* @code{COSPI}: COSPI, Circular cosine function
* @code{COTAN}: COTAN, Cotangent function
* @code{COTAND}: COTAND, Cotangent function, degrees
@@ -4635,6 +4636,34 @@ Inverse function: @*
+@node COSHAPE
+@section @code{COSHAPE} --- Coshape of a coarray
+@fnindex COSHAPE
+@cindex coarray, coshape
+
+@table @asis
+@item @emph{Synopsis}:
+@code{RESULT = COSHAPE(COARRAY [, KIND])}
+
+@item @emph{Description}:
+Returns the shape of the cobounds of a coarray.
+
+@item @emph{Standard}:
+Fortran 2018
+
The subheadings have been reordered in the Fortran intrinsics
documentation. "Standard" now comes after "See also", or at the very
end of the node if there's no "See also".
+@item @emph{Class}:
+Inquiry function
+
+@item @emph{Arguments}:
+@multitable @columnfractions .15 .70
+@item @var{COARRAY} @tab Shall be an coarray, of any type.
s/an coarray/a coarray/
+@item @var{KIND} @tab (Optional) A scalar @code{INTEGER} constant
+expression indicating the kind parameter of the result.
I think there should be
@item @emph{Return value}:
...blah blah...
after this. You could probably adapt the language from the SHAPE intrinsic.
And, probably also add
@item @emph{See also}:
pointing at SHAPE.
+@end multitable
+@end table
+
+
+
@node COSPI
@section @code{COSPI} --- Circular cosine function
@fnindex COSPI
-Sandra
