On 2026/09/29 17:28, Andrew Hewus Fresh wrote:
> There is a new compile bug this adds, but it builds fine for me on my
> sparc64, and I didn't turn on my only gcc architecture (alpha) to try
> it, since checking out the ports tree and building it would likely take
> well past the release.
> 
> https://github.com/perl5-dbi/dbi/issues/198
> 
> That said, reverse dependency tests have the same results as pre-upgrade.
> 
> It does fix a couple of CVE: https://metacpan.org/dist/DBI/changes
> 
> I did try cherry-picking the commit from the above issue as a patch and
> DBI tests pass the same as without.
> 
> https://github.com/perl5-dbi/dbi/commit/ff4f1baf0389dfa5e7282e1d8d7423c1b5e06b76.patch
> 
> Comments, OK, wait until after unlock?

Since you already found it, I would cherrypick the ff4f1baf patch too.
It just moves the variable declarations up. We don't build packages
for any base-gcc archs any more but there's no reason to break them
when it's easily avoidable.

With that, OK.


> Index: Makefile
> ===================================================================
> RCS file: /cvs/ports/databases/p5-DBI/Makefile,v
> diff -u -p -r1.77 Makefile
> --- Makefile	20 Sep 2026 03:14:14 -0000	1.77
> +++ Makefile	28 Sep 2026 18:44:52 -0000
> @@ -2,7 +2,7 @@ COMMENT =	the standard database interfac
>  
>  MODULES =	cpan
>  CPAN_AUTHOR =	HMBRAND
> -DISTNAME =	DBI-1.653
> +DISTNAME =	DBI-1.654
>  EXTRACT_SUFX =	.tgz
>  FIX_EXTRACT_PERMISSIONS =	Yes
>  CATEGORIES =	databases
> Index: distinfo
> ===================================================================
> RCS file: /cvs/ports/databases/p5-DBI/distinfo,v
> diff -u -p -r1.35 distinfo
> --- distinfo	20 Sep 2026 03:14:14 -0000	1.35
> +++ distinfo	28 Sep 2026 18:44:52 -0000
> @@ -1,2 +1,2 @@
> -SHA256 (DBI-1.653.tgz) = qYwh/Tfu2PhBFyh10XXZcv6H8GPX0NKjt3ZZCLsl61g=
> -SIZE (DBI-1.653.tgz) = 742238
> +SHA256 (DBI-1.654.tgz) = 2oU/84KugADpZ3Z61O2QxdxzAEhY9t7Ej8b0NAwwaSI=
> +SIZE (DBI-1.654.tgz) = 742647
>