On Mon, Jul 27, 2020 at 01:38:22AM -0400, Kurt Mosiejczuk wrote:
> hashlink has been dying for a bit, probably to do with the 1.11 update.
>
> Errors were complaining about no definition of char16_t. Based on what I
> saw used for fixes for clang without c++, I changed it to use uint16_t
> for the uchar typedef. This fixes the build on sparc64 and doesn't break
> it on amd64.
>
> ok?
ok thfr@
tested on amd64 without issues.
IMO should bump REVISION
Your bulk from July 28 suggests you've been doing the build with this
patch already.
>
> (cc maintainer)
>
> --Kurt
>
> Index: patches/patch-src_hl_h
> ===================================================================
> RCS file: /cvs/ports/lang/hashlink/patches/patch-src_hl_h,v
> retrieving revision 1.3
> diff -u -p -r1.3 patch-src_hl_h
> --- patches/patch-src_hl_h 11 Apr 2020 09:59:20 -0000 1.3
> +++ patches/patch-src_hl_h 27 Jul 2020 05:36:25 -0000
> @@ -3,10 +3,12 @@ $OpenBSD: patch-src_hl_h,v 1.3 2020/04/1
> add OpenBSD to ifdef
> don't typedef char{16,32}_t in clang with C++
>
> +Use uint16_t for uchar otherwise. char16_t doesn't exist
> +
> Index: src/hl.h
> --- src/hl.h.orig
> +++ src/hl.h
> -@@ -234,11 +234,13 @@ typedef uint16_t uchar;
> +@@ -234,15 +234,17 @@ typedef uint16_t uchar;
> # define USTR(str) u##str
> #else
> # include <stdarg.h>
> @@ -20,4 +22,9 @@ Index: src/hl.h
> +#endif
> #else
> # include <uchar.h>
> + #endif
> +-typedef char16_t uchar;
> ++typedef uint16_t uchar;
> + # undef USTR
> + # define USTR(str) u##str
> #endif
