On Tue, August 9, 2016 1:03 am, Gerald Hanuer wrote:
> Hello ports@,
>
> I have started the update of links+2.12 > links+2.13.
>
> The application of patches/patch-https_c succeeds with
> fuzz.
Run 'make update-patches' to fix the line numbers.
> |$OpenBSD: patch-https_c,v 1.3 2015/10/06 22:43:50 sthen Exp $
> |--- https.c.orig Thu Sep 17 18:28:56 2015
> |+++ https.c Tue Oct 6 10:49:16 2015
> --------------------------
> Patching file https.c using Plan A...
> Hunk #1 succeeded at 89 with fuzz 2 (offset 2 lines).
> done
>
> But the build fails with,
>
> https.c: In function 'getSSL':
> https.c:93: error: 'f_randfile' undeclared (first use in this function)
> https.c:93: error: (Each undeclared identifier is reported only once
> https.c:93: error: for each function it appears in.)
> *** Error 1 in /usr/obj/ports/links+-2.13/links-2.13 (Makefile:199
> 'https.o')
> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2674
> '/usr/obj/ports/links+-2.13/.build_done')
> *** Error 1 in /usr/ports/www/links+
> (/usr/ports/infrastructure/mk/bsd.port.mk:2396 'all')
>
> The source file /usr/obj/ports/links+-2.13/links-2.13/https.c
> has been modified since links-2.12, not sure if
> the patch is still relevant.
> This looks to be security sensitive and needs closer scrutiny.
>
> The build succeeds and the port appears to function as expected
> without the patches/patch-https_c, but.....
>
> Hope this helps.
>
> Regards,
> Gerald Hanuer
>
>
> Index: Makefile
> ===================================================================
> RCS file: /cvs/ports/www/links+/Makefile,v
> retrieving revision 1.58
> diff -u -p -r1.58 Makefile
> --- Makefile 17 Jan 2016 17:29:10 -0000 1.58
> +++ Makefile 9 Aug 2016 03:09:52 -0000
> @@ -1,7 +1,7 @@
> # $OpenBSD: Makefile,v 1.58 2016/01/17 17:29:10 sthen Exp $
>
> COMMENT= graphics and text browser
> -VER= 2.12
> +VER= 2.13
> REVISION= 0
REVISION should be removed.
> DISTNAME= links-${VER}
> PKGNAME= links+-${VER}
The fix is easy enough. They moved the variable definition into the #if
block so when you hit the patch's #elif, the variable won't exist. Either
move the definition back up with the others or copy/paste into the #elif
block and update the patch.
I can send a diff tomorrow if you don't get to it. Just back from
vacation and looked at this quickly.
Tim.
