CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: d...@cvs.openbsd.org2021/07/02 21:05:11 Modified files: security/py-potr: Makefile Log message: Drop maintainer
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: abie...@cvs.openbsd.org 2021/07/02 20:37:46 Modified files: infrastructure/bin: modgo-gen-modules-helper lang/go: go.port.mk Log message: Attempt to make the experience of updateing a Go port a bit easier. - Add a version argument (second) to modgo-gen-modules-helper. - Change modgo-gen-modules target to fetch module info for the current MODGO_VERSION of a port. Setting MODGO_VERSION=latest will fetch the latest version the Go ecosystem knows about. Similarly, if MODGO_VERSION is empty, fetch the latest verstion Go knows about. While here: - Add '-trimpath' to MODGO_FLAGS. OK danj@
NEW:lang/bwbasic
hello, Here is a new port, Bywater BASIC (bwBASIC). Original archive has no manpage, but NetBSD-pkgsrc creates it from bwbasic.doc with large patch. Following FreeBSD-ports, our ports has no manpage. ok? DESCR: The Bywater BASIC Interpreter (bwBASIC) implements a large superset of the ANSI Standard for Minimal BASIC (X3.60-1978), a significant subset of the ANSI Standard for Full BASIC (X3.113-1987), and many classic BASIC dialects in C. bwBASIC seeks to be as portable as possible. -- SASANO Takayoshi (JG1UAA) bwbasic.tgz Description: Binary data
Re: net/igmpproxy - chroot and drop privileges
On 2021/07/02 21:56, Bjorn Ketelaars wrote:
> Enclosed is a diff for net/igmpproxy, which puts igmpproxy in an
> unprivileged chroot after startup. I'm currently discussing a more
> extensive diff with upstream.
>
> We normally do not add features to our ports, but I was wondering if
> this addition makes sense to commit as it increases security a bit.
There is past history of adding privdrop to ports in some cases
and I think it is worthwhile for a root network daemon.
> Run tested on amd64 in combination with an iptv setup.
>
> While here add daemon_flags="${SYSCONFDIR}/igmpproxy.conf" to
> igmpproxy.rc as igmpproxy will complain if no configuration file is
> given.
>
> Thoughts/tests/comments/OK?
I don't think it's likely that pledge will work out without major
refactoring, so the setuid probably is the only way to remove those
privileges.
For the filesystem access side, I'm wondering if unveil might be a bit
simpler than chroot. OTOH, that might be harder to convince upstream to
add, and if there's chance that they might accept it, it would be nice
to be able to use whatever they have directly.
> diff --git infrastructure/db/user.list infrastructure/db/user.list
> index bfb3d70510e..f2eb6a60b8d 100644
> --- infrastructure/db/user.list
> +++ infrastructure/db/user.list
> @@ -376,3 +376,4 @@ id user group port
> 865 _vger_vger net/vger
> 866 _navidrome _navidrome audio/navidrome
> 867 _notify_push www/nextcloud_notify_push
> +868 _igmpproxy _igmpproxy net/igmpproxy
> diff --git net/igmpproxy/Makefile net/igmpproxy/Makefile
> index f87a2b4fc45..6d971d68749 100644
> --- net/igmpproxy/Makefile
> +++ net/igmpproxy/Makefile
> @@ -3,7 +3,7 @@
> COMMENT =multicast router utilizing IGMP forwarding
>
> VERSION =0.3
> -REVISION = 0
> +REVISION = 1
> DISTNAME = igmpproxy-${VERSION}
> CATEGORIES = net
> MASTER_SITES =
> https://github.com/pali/igmpproxy/releases/download/${VERSION}/
> diff --git net/igmpproxy/patches/patch-src_igmpproxy_c
> net/igmpproxy/patches/patch-src_igmpproxy_c
> index 021692a14b3..a4aee9e92f0 100644
> --- net/igmpproxy/patches/patch-src_igmpproxy_c
> +++ net/igmpproxy/patches/patch-src_igmpproxy_c
> @@ -3,7 +3,7 @@ $OpenBSD: patch-src_igmpproxy_c,v 1.1 2021/01/12 17:59:49
> sthen Exp $
> Index: src/igmpproxy.c
> --- src/igmpproxy.c.orig
> +++ src/igmpproxy.c
> -@@ -37,13 +37,10 @@
> +@@ -37,13 +37,11 @@
> * February 2005 - Johnny Egeland
> */
>
> @@ -14,12 +14,23 @@ Index: src/igmpproxy.c
> -
> #include "igmpproxy.h"
>
> ++#include
> +#include
> +
> static const char Usage[] =
> "Usage: igmpproxy [-h] [-n] [-d] [-v [-v]] \n"
> "\n"
> -@@ -123,6 +120,25 @@ int main( int ArgCn, char *ArgVc[] ) {
> +@@ -68,6 +66,9 @@ static int sighandled = 0;
> + #define GOT_SIGUSR1 0x04
> + #define GOT_SIGUSR2 0x08
> +
> ++#define CHROOT_DIR "/var/empty"
> ++#define NOPRIV_USER "_igmpproxy"
> ++
> + // Holds the indeces of the upstream IF...
> + int upStreamIfIdx[MAX_UPS_VIFS];
> +
> +@@ -123,6 +124,25 @@ int main( int ArgCn, char *ArgVc[] ) {
>
> openlog("igmpproxy", LOG_PID, LOG_USER);
>
> @@ -45,25 +56,42 @@ Index: src/igmpproxy.c
> // Write debug notice with file path...
> my_log(LOG_DEBUG, 0, "Searching for config file at '%s'" ,
> configFilePath);
>
> -@@ -142,16 +158,8 @@ int main( int ArgCn, char *ArgVc[] ) {
> +@@ -140,18 +160,25 @@ int main( int ArgCn, char *ArgVc[] ) {
> + break;
> + }
>
> - if ( !NotAsDaemon ) {
> +-if ( !NotAsDaemon ) {
> ++// Drop privileges
> ++{
> ++struct passwd *pw;
>
> -// Only daemon goes past this line...
> -if (fork()) exit(0);
> --
> ++pw = getpwnam(NOPRIV_USER);
> ++if (pw == NULL)
> ++my_log(LOG_ERR, 0, "unknown user %s", NOPRIV_USER);
> +
> -// Detach daemon from terminal
> -if ( close( 0 ) < 0 || close( 1 ) < 0 || close( 2 ) < 0
> -|| open( "/dev/null", 0 ) != 0 || dup2( 0, 1 ) < 0 || dup2(
> 0, 2 ) < 0
> -|| setpgid( 0, 0 ) < 0
> -) {
> ++if (chroot(CHROOT_DIR) != 0 || chdir("/") != 0 ||
> ++ setgroups(1, >pw_gid) != 0 ||
> ++ setresgid(pw->pw_gid, pw->pw_gid, pw->pw_gid) != 0 ||
> ++ setresuid(pw->pw_uid, pw->pw_uid, pw->pw_uid) != 0)
> ++my_log(LOG_ERR, 0, "cannot drop privileges");
> ++}
> ++
> ++if ( !NotAsDaemon ) {
> ++
> +if ( daemon(1, 0 ) < 0 )
> my_log( LOG_ERR, errno, "failed to detach daemon" );
> -}
> }
>
> // Go to the main loop.
> -@@ -207,6 +215,8 @@ int igmpProxyInit(void) {
> +@@ -207,6 +234,8 @@ int igmpProxyInit(void) {
> }
>
> for ( Ix = 0; (Dp = getIfByIx(Ix)); Ix++
net/igmpproxy - chroot and drop privileges
Enclosed is a diff for net/igmpproxy, which puts igmpproxy in an
unprivileged chroot after startup. I'm currently discussing a more
extensive diff with upstream.
We normally do not add features to our ports, but I was wondering if
this addition makes sense to commit as it increases security a bit.
Run tested on amd64 in combination with an iptv setup.
While here add daemon_flags="${SYSCONFDIR}/igmpproxy.conf" to
igmpproxy.rc as igmpproxy will complain if no configuration file is
given.
Thoughts/tests/comments/OK?
diff --git infrastructure/db/user.list infrastructure/db/user.list
index bfb3d70510e..f2eb6a60b8d 100644
--- infrastructure/db/user.list
+++ infrastructure/db/user.list
@@ -376,3 +376,4 @@ id usergroup port
865 _vger _vger net/vger
866 _navidrome _navidrome audio/navidrome
867 _notify_push www/nextcloud_notify_push
+868 _igmpproxy _igmpproxy net/igmpproxy
diff --git net/igmpproxy/Makefile net/igmpproxy/Makefile
index f87a2b4fc45..6d971d68749 100644
--- net/igmpproxy/Makefile
+++ net/igmpproxy/Makefile
@@ -3,7 +3,7 @@
COMMENT = multicast router utilizing IGMP forwarding
VERSION = 0.3
-REVISION = 0
+REVISION = 1
DISTNAME = igmpproxy-${VERSION}
CATEGORIES = net
MASTER_SITES = https://github.com/pali/igmpproxy/releases/download/${VERSION}/
diff --git net/igmpproxy/patches/patch-src_igmpproxy_c
net/igmpproxy/patches/patch-src_igmpproxy_c
index 021692a14b3..a4aee9e92f0 100644
--- net/igmpproxy/patches/patch-src_igmpproxy_c
+++ net/igmpproxy/patches/patch-src_igmpproxy_c
@@ -3,7 +3,7 @@ $OpenBSD: patch-src_igmpproxy_c,v 1.1 2021/01/12 17:59:49 sthen
Exp $
Index: src/igmpproxy.c
--- src/igmpproxy.c.orig
+++ src/igmpproxy.c
-@@ -37,13 +37,10 @@
+@@ -37,13 +37,11 @@
* February 2005 - Johnny Egeland
*/
@@ -14,12 +14,23 @@ Index: src/igmpproxy.c
-
#include "igmpproxy.h"
++#include
+#include
+
static const char Usage[] =
"Usage: igmpproxy [-h] [-n] [-d] [-v [-v]] \n"
"\n"
-@@ -123,6 +120,25 @@ int main( int ArgCn, char *ArgVc[] ) {
+@@ -68,6 +66,9 @@ static int sighandled = 0;
+ #define GOT_SIGUSR1 0x04
+ #define GOT_SIGUSR2 0x08
+
++#define CHROOT_DIR "/var/empty"
++#define NOPRIV_USER "_igmpproxy"
++
+ // Holds the indeces of the upstream IF...
+ int upStreamIfIdx[MAX_UPS_VIFS];
+
+@@ -123,6 +124,25 @@ int main( int ArgCn, char *ArgVc[] ) {
openlog("igmpproxy", LOG_PID, LOG_USER);
@@ -45,25 +56,42 @@ Index: src/igmpproxy.c
// Write debug notice with file path...
my_log(LOG_DEBUG, 0, "Searching for config file at '%s'" ,
configFilePath);
-@@ -142,16 +158,8 @@ int main( int ArgCn, char *ArgVc[] ) {
+@@ -140,18 +160,25 @@ int main( int ArgCn, char *ArgVc[] ) {
+ break;
+ }
- if ( !NotAsDaemon ) {
+-if ( !NotAsDaemon ) {
++// Drop privileges
++{
++struct passwd *pw;
-// Only daemon goes past this line...
-if (fork()) exit(0);
--
++pw = getpwnam(NOPRIV_USER);
++if (pw == NULL)
++my_log(LOG_ERR, 0, "unknown user %s", NOPRIV_USER);
+
-// Detach daemon from terminal
-if ( close( 0 ) < 0 || close( 1 ) < 0 || close( 2 ) < 0
-|| open( "/dev/null", 0 ) != 0 || dup2( 0, 1 ) < 0 || dup2(
0, 2 ) < 0
-|| setpgid( 0, 0 ) < 0
-) {
++if (chroot(CHROOT_DIR) != 0 || chdir("/") != 0 ||
++ setgroups(1, >pw_gid) != 0 ||
++ setresgid(pw->pw_gid, pw->pw_gid, pw->pw_gid) != 0 ||
++ setresuid(pw->pw_uid, pw->pw_uid, pw->pw_uid) != 0)
++my_log(LOG_ERR, 0, "cannot drop privileges");
++}
++
++if ( !NotAsDaemon ) {
++
+if ( daemon(1, 0 ) < 0 )
my_log( LOG_ERR, errno, "failed to detach daemon" );
-}
}
// Go to the main loop.
-@@ -207,6 +215,8 @@ int igmpProxyInit(void) {
+@@ -207,6 +234,8 @@ int igmpProxyInit(void) {
}
for ( Ix = 0; (Dp = getIfByIx(Ix)); Ix++ ) {
diff --git net/igmpproxy/pkg/PLIST net/igmpproxy/pkg/PLIST
index 1cbe5a08909..32ebfcdd078 100644
--- net/igmpproxy/pkg/PLIST
+++ net/igmpproxy/pkg/PLIST
@@ -1,4 +1,6 @@
@comment $OpenBSD: PLIST,v 1.5 2021/01/12 17:59:50 sthen Exp $
+@newgroup _igmpproxy:868
+@newuser _igmpproxy:868:868:daemon:IGMP multicast routing
daemon:/var/empty:/sbin/nologin
@rcscript ${RCDIR}/igmpproxy
@man man/man5/igmpproxy.conf.5
@man man/man8/igmpproxy.8
diff --git net/igmpproxy/pkg/igmpproxy.rc net/igmpproxy/pkg/igmpproxy.rc
index f4366b88351..3718a74d8a2 100644
--- net/igmpproxy/pkg/igmpproxy.rc
+++ net/igmpproxy/pkg/igmpproxy.rc
@@ -3,6 +3,7 @@
# $OpenBSD: igmpproxy.rc,v 1.2 2018/01/11 19:27:05 rpe Exp $
daemon="${TRUEPREFIX}/sbin/igmpproxy"
+daemon_flags="${SYSCONFDIR}/igmpproxy.conf"
. /etc/rc.d/rc.subr
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 13:31:34 Modified files: security : Makefile Log message: +pass-import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 13:31:13 Log message: import ports/security/pass-import, ok kmos pass-import is an extension for "password-store", allowing you to import password databases from various password managers to a pass(1) repository. By default, entries are imported at the root of the password store and only the main data (password, login, email, URL, group) are kept. This behaviour can be changed using the provided options. pass-import handles duplicates and is compatible with browserpass. It imports OTP secrets in a way that is compatible with pass-otp. pass-import also provides a pimport script that allows importing passwords to a small number other password managers. For instance, you can import passwords to a Keepass database or a generic CSV file. Status: Vendor Tag: sthen Release Tags: sthen_20210702 N ports/security/pass-import/Makefile N ports/security/pass-import/distinfo N ports/security/pass-import/pkg/DESCR N ports/security/pass-import/pkg/PLIST N ports/security/pass-import/patches/patch-setup_py No conflicts created by this import
Re: textproc/swish-e: fix upcoming build failure
On 2021/07/02 19:01, Theo Buehler wrote: > On Fri, Jul 02, 2021 at 04:47:27PM +0100, Stuart Henderson wrote: > > On 2021/07/02 16:46, Theo Buehler wrote: > > > zlib 1.2.9 added an uncompress2 function that clashes with swish's > > > internal compress2 function. The build is fixed by the following > > > mechanical diff that renames the internal function to swish_uncompress2 > > > as was done in alpine: > > > > > > https://git.alpinelinux.org/aports/tree/main/swish-e/zlib-shadowing.patch?h=3.14-stable > > > > > > I couldn't come up with a preprocessor trick that would allow avoiding > > > the bulk rename, but hopefully there's something cleverer that would > > > avoid a big diff. > > > > > > HOMEPAGE is broken, so I removed it. > > > > here's the trick :) the actual symbol in libz is z_uncompress2, we can > > just get rid of the macro then there's no more conflict. > > Unfortunately that doesn't work. > > It's not a linking problem (which your trick could have solved > potentially) but a problem of conflicting prototypes in "compress.h" > and : > > In file included from compress.c:54: > /usr/include/zlib.h:1283:21: error: conflicting types for 'uncompress2' > ZEXTERN int ZEXPORT uncompress2 OF((Bytef *dest, uLongf *destLen, > ^ > ./compress.h:39:5: note: previous declaration is here sigh, for some reason I though 'make install' would install headers.. > Unless you have another clever idea, I suggest to go with my original > diff, pointing to your hosted tarball and with 'make update-plist'. OK.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bcal...@cvs.openbsd.org 2021/07/02 11:47:25 Modified files: sysutils/diffoscope: Makefile distinfo Log message: Update to diffoscope-177
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: bcal...@cvs.openbsd.org 2021/07/02 11:45:36 Modified files: sysutils/ugrep : Makefile distinfo sysutils/ugrep/patches: patch-configure Log message: Update to ugrep-3.3.4 Changelog: https://github.com/Genivia/ugrep/releases/tag/v3.3.4
Re: textproc/swish-e: fix upcoming build failure
On Fri, Jul 02, 2021 at 04:47:27PM +0100, Stuart Henderson wrote: > On 2021/07/02 16:46, Theo Buehler wrote: > > zlib 1.2.9 added an uncompress2 function that clashes with swish's > > internal compress2 function. The build is fixed by the following > > mechanical diff that renames the internal function to swish_uncompress2 > > as was done in alpine: > > > > https://git.alpinelinux.org/aports/tree/main/swish-e/zlib-shadowing.patch?h=3.14-stable > > > > I couldn't come up with a preprocessor trick that would allow avoiding > > the bulk rename, but hopefully there's something cleverer that would > > avoid a big diff. > > > > HOMEPAGE is broken, so I removed it. > > here's the trick :) the actual symbol in libz is z_uncompress2, we can > just get rid of the macro then there's no more conflict. Unfortunately that doesn't work. It's not a linking problem (which your trick could have solved potentially) but a problem of conflicting prototypes in "compress.h" and : In file included from compress.c:54: /usr/include/zlib.h:1283:21: error: conflicting types for 'uncompress2' ZEXTERN int ZEXPORT uncompress2 OF((Bytef *dest, uLongf *destLen, ^ ./compress.h:39:5: note: previous declaration is here int uncompress2(unsigned char **buffer); (The second problem is that I wrote compress2 by accident. You'd need "#undef uncompress2" in your compress.c patch, and the patch for docprop_write.c is not necessary) Unless you have another clever idea, I suggest to go with my original diff, pointing to your hosted tarball and with 'make update-plist'.
Re: UPDATE: net/nextcloudclient-3.2.3
Ping. Em sex., 25 de jun. de 2021 às 17:40, Adriano Barbosa escreveu: > > Hi. > Update for net/nextcloudclient v3.2.3. > Changelog: https://github.com/nextcloud/desktop/releases/tag/v3.2.3 > > graphics/inkscape is now required for building as per > https://github.com/nextcloud/desktop/pull/3466 > > Thank you for testing. > -- > Adriano Barbosa -- Adriano
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: fcam...@cvs.openbsd.org 2021/07/02 10:08:07 Modified files: devel/cvsgraph : Makefile distinfo Log message: Update cvsgraph to 1.7.2.
Re: OpenBSD ports: graywolf broken
Weekly ping!
Il 28 giugno 2021 14:02:31 CEST, Alessandro De Laurenzis
ha scritto:
>Hello Jasper,
>
>There are indeed some header files named "dialog.h" spread into the
>source code, causing a clash when /usr/local/include/dialog.h is
>present.
>
>I'm not an expert of the CMake ecosystem, but from a quick search on
>the
>net, the usual solution to this kind of issues is to explicitly add the
>
>current source and build directories to the include path, by mean of
>the
>CMAKE_INCLUDE_CURRENT_DIR variable.
>
>Please find enclosed a diff with bumped revision and re-based patches
>(while there, I also fixed some indentations in the port's Makefile).
>
>Of course, alternative approach suggestions are welcome.
>
>Copying ports@
>
>On 27/06/2021 13:38, Jasper Lievisse Adriaanse wrote:
>> Hello,
>>
>> I noticed that the cad/graywolf port doesn’t build when misc/dialog
>is installed:
>>
>> [2/159] : && /usr/ports/pobj/graywolf-0.1.6/bin/cc -O2 -pipe
>-Wno-return-type -DSIZEOF_VOID_P=64 -Wno-implicit-function-declaration
>-DNDEBUG src/
>> date/CMakeFiles/getdate.dir/getdate.c.o -o src/date/getdate
>-L/usr/local/lib -Wl,-rpath-link,/usr/X11R6/lib:/usr/local/lib && :
>> [3/159] /usr/ports/pobj/graywolf-0.1.6/bin/cc -I/usr/local/include
>-I/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/include -Iinclude
>-I/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/src/twmc/. -O2 -pipe
>-Wno-return-type -DSIZEOF_VOID_P=64 -Wno-implicit-function-declaration
>-DNDEBUG -MD -MT src
>> /twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o -MF
>src/twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o.d -o
>src/twmc/CMakeFiles/TimberWolfMC.dir/graph
>> ics.c.o -c
>/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/src/twmc/graphics.c
>> FAILED: src/twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o
>> /usr/ports/pobj/graywolf-0.1.6/bin/cc -I/usr/local/include
>-I/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/include -Iinclude
>-I/usr/ports/pobj/gray
>> wolf-0.1.6/graywolf-0.1.6/src/twmc/. -O2 -pipe -Wno-return-type
>-DSIZEOF_VOID_P=64 -Wno-implicit-function-declaration -DNDEBUG -MD -MT
>src/twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o -MF
>src/twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o.d -o
>src/twmc/CMakeFiles/TimberWolfMC.dir/graphics.c.o -c
>/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/src/twmc/graphics.c
>In file included from
>/usr/ports/pobj/graywolf-0.1.6/graywolf-0.1.6/src/twmc/graphics.c:137:
>> In file included from /usr/local/include/dialog.h:59:
>> /usr/include/ncurses.h:1216:32: error: expected identifier or '('
>> extern NCURSES_EXPORT_VAR(int) COLORS;
>>^
>>
>> Could you have a look at this please?
>>
>> Cheers,
>> Jasper
>>
>
>--
>Alessandro De Laurenzis
>[mailto:jus...@atlantide.mooo.com]
>Web: http://www.atlantide.mooo.com
>LinkedIn: http://it.linkedin.com/in/delaurenzis
--
Inviato dal mio dispositivo Android con K-9 Mail. Perdonate la brevità.
Re: textproc/swish-e: fix upcoming build failure
On 2021/07/02 16:46, Theo Buehler wrote:
> zlib 1.2.9 added an uncompress2 function that clashes with swish's
> internal compress2 function. The build is fixed by the following
> mechanical diff that renames the internal function to swish_uncompress2
> as was done in alpine:
>
> https://git.alpinelinux.org/aports/tree/main/swish-e/zlib-shadowing.patch?h=3.14-stable
>
> I couldn't come up with a preprocessor trick that would allow avoiding
> the bulk rename, but hopefully there's something cleverer that would
> avoid a big diff.
>
> HOMEPAGE is broken, so I removed it.
here's the trick :) the actual symbol in libz is z_uncompress2, we can
just get rid of the macro then there's no more conflict.
I mirrored the distfile too.
Index: Makefile
===
RCS file: /cvs/ports/textproc/swish-e/Makefile,v
retrieving revision 1.14
diff -u -p -r1.14 Makefile
--- Makefile12 Jul 2019 20:50:15 - 1.14
+++ Makefile2 Jul 2021 15:45:45 -
@@ -3,12 +3,11 @@
COMMENT = web/text search engine
DISTNAME = swish-e-2.4.7
-REVISION = 4
+REVISION = 5
SHARED_LIBS = swish-e 2.1
CATEGORIES = textproc www
-HOMEPAGE = http://swish-e.org/
-MASTER_SITES = ${HOMEPAGE}distribution/
+MASTER_SITES = https://spacehopper.org/mirrors/
# GPLv2
PERMIT_PACKAGE = Yes
Index: patches/patch-src_compress_c
===
RCS file: /cvs/ports/textproc/swish-e/patches/patch-src_compress_c,v
retrieving revision 1.1
diff -u -p -r1.1 patch-src_compress_c
--- patches/patch-src_compress_c2 Dec 2014 11:59:42 - 1.1
+++ patches/patch-src_compress_c2 Jul 2021 15:45:45 -
@@ -1,7 +1,17 @@
$OpenBSD: patch-src_compress_c,v 1.1 2014/12/02 11:59:42 naddy Exp $
src/compress.c.origSun Apr 5 03:58:32 2009
-+++ src/compress.c Tue Dec 2 05:06:54 2014
-@@ -995,7 +995,7 @@ voidremove_worddata_longs(unsigned char *worddata,
+
+Index: src/compress.c
+--- src/compress.c.orig
src/compress.c
+@@ -53,6 +53,7 @@ $Id: compress.c 1945 2007-10-22 14:54:07Z karpet $
+ #ifdef HAVE_ZLIB
+ #include
+ #define Z_BUFSIZE 16384
++#undef compress2
+ #endif
+
+ /* Surfing the web I found this:
+@@ -995,7 +996,7 @@ voidremove_worddata_longs(unsigned char *worddata,
progerr("Internal error in remove_worddata_longs");
/* dst may be smaller than src. So move the data */
Index: patches/patch-src_docprop_write_c
===
RCS file: patches/patch-src_docprop_write_c
diff -N patches/patch-src_docprop_write_c
--- /dev/null 1 Jan 1970 00:00:00 -
+++ patches/patch-src_docprop_write_c 2 Jul 2021 15:45:45 -
@@ -0,0 +1,13 @@
+$OpenBSD$
+
+Index: src/docprop_write.c
+--- src/docprop_write.c.orig
src/docprop_write.c
+@@ -38,6 +38,7 @@
+ #include "db.h"
+ #ifdef HAVE_ZLIB
+ #include
++#undef compress2
+ #endif
+
+
Index: pkg/PLIST
===
RCS file: /cvs/ports/textproc/swish-e/pkg/PLIST,v
retrieving revision 1.3
diff -u -p -r1.3 PLIST
--- pkg/PLIST 16 Mar 2015 18:07:57 - 1.3
+++ pkg/PLIST 2 Jul 2021 15:45:45 -
@@ -3,7 +3,7 @@ bin/swish-config
@bin bin/swish-e
bin/swish-filter-test
include/swish-e.h
-lib/libswish-e.a
+@static-lib lib/libswish-e.a
lib/libswish-e.la
@lib lib/libswish-e.so.${LIBswish-e_VERSION}
lib/pkgconfig/swish-e.pc
Re: new: security/pass-import, py-pykeepass, and py-construct bump
On 2021/07/02 10:42, Kurt Mosiejczuk wrote: > On Thu, Jul 01, 2021 at 09:36:37PM +0100, Stuart Henderson wrote: > > pass-import is an extension for password-store aka pass(1) which allows > > importing from a variety of other password managers. I only wanted it > > for one of the csv formats but since I was there anyway I figured it > > would be useful to port pykeepass to allow that to work too, which in > > turn needs a specific version of py-construct. > > > The warning in py-construct's Makefile seems obsolete. plaso's Makefile > > no longer references py-construct (it did previously), and it doesn't > > show up in the code. It is a listed dep in some of plaso's dependencies > > but again doesn't seem to actually show up in the code. So I don't > > think udpating it is a problem (though pykeepass does want this > > specific version at the moment). > > > py-construct diff inline, new ports attached. Any comments/OKs? > > ok kmos on py-construct > > I'd like a comment about why NO_TEST is set for py-pykeepass. Otherwise > ok kmos for import Thanks. > Is there a reason to not just use MODPY_EGG_VERSION directly on pass-import? > > Other than that ok kmos on importing pass-import Partly to save typing, partly to avoid a super-long MASTER_SITES line :)
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 09:29:47 Modified files: security : Makefile Log message: +py-pykeepass
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 09:29:09 Log message: import ports/security/py-pykeepass, ok kmos pykeepass is a Python library to interact with keepass databases (supports KDBX3 and KDBX4). Status: Vendor Tag: sthen Release Tags: sthen_20210702 N ports/security/py-pykeepass/Makefile N ports/security/py-pykeepass/distinfo N ports/security/py-pykeepass/pkg/DESCR N ports/security/py-pykeepass/pkg/PLIST N ports/security/py-pykeepass/patches/patch-pykeepass_kdbx_parsing_common_py N ports/security/py-pykeepass/patches/patch-pykeepass_kdbx_parsing_twofish_py No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 09:28:37 Modified files: devel/py-construct: Makefile distinfo devel/py-construct/pkg: PLIST Log message: update to py-construct 2.10.54, ok kmos
aarch64 bulk build report
bulk build on arm64.ports.openbsd.org started on Wed Jun 30 02:23:57 MDT 2021 finished at Fri Jul 2 08:51:59 MDT 2021 lasted 2D06h28m done with kern.version=OpenBSD 6.9-current (GENERIC.MP) #1222: Tue Jun 29 15:09:54 MDT 2021 built packages:11052 Jun 30:4233 Jul 1:2414 Jul 2:4404 critical path missing pkgs: http://build-failures.rhaalovely.net/aarch64/2021-06-30/summary.log build failures: 12 http://build-failures.rhaalovely.net/aarch64/2021-06-30/cad/graywolf.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/comms/gnuradio.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/editors/micro.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/games/shockolate.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/graphics/gmic.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/lang/pcc/pcc.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/lang/pcc/pcc-libs.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/net/minio/server.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/sysutils/rundeck.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/sysutils/telegraf.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/www/chromium.log http://build-failures.rhaalovely.net/aarch64/2021-06-30/www/seamonkey,-lightning.log recurrent failures failures/comms/gnuradio.log failures/editors/micro.log failures/games/shockolate.log new failures +++ ls-failures Fri Jul 2 08:52:08 2021 +failures/cad/graywolf.log resolved failures --- ../old/aarch64/last//ls-failuresFri Jun 25 01:43:42 2021
textproc/swish-e: fix upcoming build failure
zlib 1.2.9 added an uncompress2 function that clashes with swish's
internal compress2 function. The build is fixed by the following
mechanical diff that renames the internal function to swish_uncompress2
as was done in alpine:
https://git.alpinelinux.org/aports/tree/main/swish-e/zlib-shadowing.patch?h=3.14-stable
I couldn't come up with a preprocessor trick that would allow avoiding
the bulk rename, but hopefully there's something cleverer that would
avoid a big diff.
HOMEPAGE is broken, so I removed it.
Index: Makefile
===
RCS file: /cvs/ports/textproc/swish-e/Makefile,v
retrieving revision 1.14
diff -u -p -r1.14 Makefile
--- Makefile12 Jul 2019 20:50:15 - 1.14
+++ Makefile2 Jul 2021 14:06:04 -
@@ -3,12 +3,11 @@
COMMENT = web/text search engine
DISTNAME = swish-e-2.4.7
-REVISION = 4
+REVISION = 5
SHARED_LIBS = swish-e 2.1
CATEGORIES = textproc www
-HOMEPAGE = http://swish-e.org/
-MASTER_SITES = ${HOMEPAGE}distribution/
+MASTER_SITES = http://swish-e.org/distribution/
# GPLv2
PERMIT_PACKAGE = Yes
Index: patches/patch-src_btree_c
===
RCS file: patches/patch-src_btree_c
diff -N patches/patch-src_btree_c
--- /dev/null 1 Jan 1970 00:00:00 -
+++ patches/patch-src_btree_c 2 Jul 2021 14:04:13 -
@@ -0,0 +1,119 @@
+$OpenBSD$
+
+Fix namespace collision with libz
+
+Index: src/btree.c
+--- src/btree.c.orig
src/btree.c
+@@ -321,7 +321,7 @@ unsigned char *key_k;
+ {
+ k = j + (i - j) / 2;
+ key_k = BTREE_KeyData(pg,k);
+-key_len_k = uncompress2(_k);
++key_len_k = swish_uncompress2(_k);
+ isbigger = BTREE_CompareKeys(key,key_len,key_k,key_len_k);
+ if (!isbigger)
+ break;
+@@ -358,7 +358,7 @@ sw_off_t data_pointer;
+ return 0;
+
+ *found = BTREE_KeyData(pg,k);
+-*found_len = uncompress2(found);
++*found_len = swish_uncompress2(found);
+
+ /* Solaris do not like this. Use memcpy instead
+ data_pointer = *(sw_off_t *) (*found + *found_len);
+@@ -483,7 +483,7 @@ int j, k = pos;
+
+ /* Compute length of deleted key */
+ del_key_start = q = BTREE_KeyData(pg,k);
+-q += uncompress2();
++q += swish_uncompress2();
+ q += sizeof(sw_off_t);
+ del_key_end = q;
+ del_entry_len = del_key_end - del_key_start;
+@@ -610,7 +610,7 @@ int tmp;
+ for(i = 0; i < n; i++)
+ {
+ key_data = start = BTREE_KeyData(pg, pg->n - n + i);
+-key_len = uncompress2(_data);
++key_len = swish_uncompress2(_data);
+
+ memcpy(p, start, (key_data - start) + key_len + sizeof(sw_off_t));
+ tmp = p - new_pg->data;
+@@ -628,7 +628,7 @@ int tmp;
+ for(i = 0; i < (int)pg->n ; i++)
+ {
+ key_data = start = BTREE_KeyData(pg,i);
+-key_len = uncompress2(_data);
++key_len = swish_uncompress2(_data);
+
+ memmove(p, start, (key_data - start) + key_len + sizeof(sw_off_t));
+ tmp = p - pg->data;
+@@ -687,7 +687,7 @@ int comp;
+ if(!(pg->flags & BTREE_ROOT_NODE))
+ {
+ key_data0 = BTREE_KeyData(pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 = swish_uncompress2(_data0);
+ father_pg = BTREE_ReadPage(b,b->tree[level]);
+ BTREE_InsertInPage(b,father_pg, key_data0, key_len0,
pg->page_number, level - 1, 1);
+ }
+@@ -711,7 +711,7 @@ int comp;
+ pg->next = new_pg->page_number;
+
+ key_data0 = BTREE_KeyData(new_pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 = swish_uncompress2(_data0);
+
+ /* Let's see where to put the key */
+ if(BTREE_CompareKeys(key, key_len, key_data0, key_len0) > 0)
+@@ -740,10 +740,10 @@ int comp;
+ root_page = BTREE_NewPage(b,b->page_size, BTREE_ROOT_NODE);
+
+ key_data0 = BTREE_KeyData(pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 = swish_uncompress2(_data0);
+ BTREE_AddKeyToPage(root_page, 0, key_data0, key_len0 ,
pg->page_number);
+ key_data0 = BTREE_KeyData(new_pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 = swish_uncompress2(_data0);
+ BTREE_AddKeyToPage(root_page, 1, key_data0, key_len0,
new_pg->page_number);
+
+ b->root_page = root_page->page_number;
+@@ -769,7 +769,7 @@ int comp;
+ BTREE_FreePage(b, pg);
+
+ key_data0 = BTREE_KeyData(new_pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 = swish_uncompress2(_data0);
+ BTREE_FreePage(b, BTREE_Walk(b,key_data0,key_len0));
+ }
+ else
+@@ -778,7 +778,7 @@ int comp;
+ BTREE_FreePage(b, pg);
+
+ key_data0 = BTREE_KeyData(new_pg,0);
+-key_len0 = uncompress2(_data0);
++key_len0 =
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: fcam...@cvs.openbsd.org 2021/07/02 08:43:46 Modified files: emulators/mame : Makefile distinfo emulators/mame/patches: patch-makefile Log message: Update mame to 0.233.
Re: new: security/pass-import, py-pykeepass, and py-construct bump
On Thu, Jul 01, 2021 at 09:36:37PM +0100, Stuart Henderson wrote: > pass-import is an extension for password-store aka pass(1) which allows > importing from a variety of other password managers. I only wanted it > for one of the csv formats but since I was there anyway I figured it > would be useful to port pykeepass to allow that to work too, which in > turn needs a specific version of py-construct. > The warning in py-construct's Makefile seems obsolete. plaso's Makefile > no longer references py-construct (it did previously), and it doesn't > show up in the code. It is a listed dep in some of plaso's dependencies > but again doesn't seem to actually show up in the code. So I don't > think udpating it is a problem (though pykeepass does want this > specific version at the moment). > py-construct diff inline, new ports attached. Any comments/OKs? ok kmos on py-construct I'd like a comment about why NO_TEST is set for py-pykeepass. Otherwise ok kmos for import Is there a reason to not just use MODPY_EGG_VERSION directly on pass-import? Other than that ok kmos on importing pass-import --Kurt
Re: new: security/pass-import, py-pykeepass, and py-construct bump
On 2021/07/02 10:25, Kurt Mosiejczuk wrote: > On Thu, Jul 01, 2021 at 09:36:37PM +0100, Stuart Henderson wrote: > > pass-import is an extension for password-store aka pass(1) which allows > > importing from a variety of other password managers. I only wanted it > > for one of the csv formats but since I was there anyway I figured it > > would be useful to port pykeepass to allow that to work too, which in > > turn needs a specific version of py-construct. > > > The warning in py-construct's Makefile seems obsolete. plaso's Makefile > > no longer references py-construct (it did previously), and it doesn't > > show up in the code. It is a listed dep in some of plaso's dependencies > > but again doesn't seem to actually show up in the code. So I don't > > think udpating it is a problem (though pykeepass does want this > > specific version at the moment). > > > py-construct diff inline, new ports attached. Any comments/OKs? > > >> Fetch https://pypi.io/packages/source/c/construct/construct-2.10.54.tar.gz > construct-2.10.54.tar.gz 100% |*| 55318 00:00 > > >> Size does not match for construct-2.10.54.tar.gz Oh, I know what this is, I tried the github tarball to look at tests and forget to refetch afterwards - please regen distinfo.
Re: new: security/pass-import, py-pykeepass, and py-construct bump
On Thu, Jul 01, 2021 at 09:36:37PM +0100, Stuart Henderson wrote:
> pass-import is an extension for password-store aka pass(1) which allows
> importing from a variety of other password managers. I only wanted it
> for one of the csv formats but since I was there anyway I figured it
> would be useful to port pykeepass to allow that to work too, which in
> turn needs a specific version of py-construct.
> The warning in py-construct's Makefile seems obsolete. plaso's Makefile
> no longer references py-construct (it did previously), and it doesn't
> show up in the code. It is a listed dep in some of plaso's dependencies
> but again doesn't seem to actually show up in the code. So I don't
> think udpating it is a problem (though pykeepass does want this
> specific version at the moment).
> py-construct diff inline, new ports attached. Any comments/OKs?
>> Fetch https://pypi.io/packages/source/c/construct/construct-2.10.54.tar.gz
construct-2.10.54.tar.gz 100% |*| 55318 00:00
>> Size does not match for construct-2.10.54.tar.gz
--Kurt
>
> Index: Makefile
> ===
> RCS file: /cvs/ports/devel/py-construct/Makefile,v
> retrieving revision 1.8
> diff -u -p -r1.8 Makefile
> --- Makefile 10 Oct 2020 18:45:00 - 1.8
> +++ Makefile 1 Jul 2021 20:31:06 -
> @@ -2,29 +2,28 @@
>
> COMMENT =powerful declarative parser/builder for binary data
>
> -# !!! do not update blindly: security/plaso needs a specific version
> +# as of 2021/07/01, pykeepass needs this specific version
> +MODPY_EGG_VERSION = 2.10.54
> EPOCH = 0
> -MODPY_EGG_VERSION = 2.5.3
> +
> DISTNAME = construct-${MODPY_EGG_VERSION}
> PKGNAME =py-${DISTNAME}
> -REVISION = 2
>
> CATEGORIES = devel
>
> -HOMEPAGE = http://construct.readthedocs.io/
> +HOMEPAGE = https://construct.readthedocs.io/
>
> # MIT
> PERMIT_PACKAGE = Yes
>
> MODULES =lang/python
> MODPY_SETUPTOOLS = Yes
> -MODPY_PYTEST = Yes
> -
> MODPY_PI = Yes
>
> FLAVORS =python3
> FLAVOR = python3
>
> -TEST_DEPENDS = math/py-numpy${MODPY_FLAVOR}
> +# github version has tests, but they require arrow (not yet in ports)
> +NO_TEST =Yes
>
> .include
> Index: distinfo
> ===
> RCS file: /cvs/ports/devel/py-construct/distinfo,v
> retrieving revision 1.2
> diff -u -p -r1.2 distinfo
> --- distinfo 23 Feb 2017 13:11:20 - 1.2
> +++ distinfo 1 Jul 2021 20:31:06 -
> @@ -1,2 +1,2 @@
> -SHA256 (construct-2.5.3.tar.gz) =
> siM7x8ak1zgD9YBrwxZ+STXao+qKXNBWy6u9ucPXmjY=
> -SIZE (construct-2.5.3.tar.gz) = 688072
> +SHA256 (construct-2.10.54.tar.gz) =
> Te9rTuAm86Z6tTf+94PjLsYz11HeegADBgJfv9Mcf9g=
> +SIZE (construct-2.10.54.tar.gz) = 1184400
> Index: pkg/PLIST
> ===
> RCS file: /cvs/ports/devel/py-construct/pkg/PLIST,v
> retrieving revision 1.3
> diff -u -p -r1.3 PLIST
> --- pkg/PLIST 10 Oct 2020 18:45:00 - 1.3
> +++ pkg/PLIST 1 Jul 2021 20:31:06 -
> @@ -11,123 +11,25 @@ lib/python${MODPY_VERSION}/site-packages
> lib/python${MODPY_VERSION}/site-packages/construct/__init__.py
>
> ${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}/
>
> lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}__init__.${MODPY_PYC_MAGIC_TAG}pyc
> -lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}adapters.${MODPY_PYC_MAGIC_TAG}pyc
>
> lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}core.${MODPY_PYC_MAGIC_TAG}pyc
>
> lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}debug.${MODPY_PYC_MAGIC_TAG}pyc
> -lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}macros.${MODPY_PYC_MAGIC_TAG}pyc
> +lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}expr.${MODPY_PYC_MAGIC_TAG}pyc
>
> lib/python${MODPY_VERSION}/site-packages/construct/${MODPY_PYCACHE}version.${MODPY_PYC_MAGIC_TAG}pyc
> -lib/python${MODPY_VERSION}/site-packages/construct/adapters.py
> lib/python${MODPY_VERSION}/site-packages/construct/core.py
> lib/python${MODPY_VERSION}/site-packages/construct/debug.py
> -lib/python${MODPY_VERSION}/site-packages/construct/formats/
> -lib/python${MODPY_VERSION}/site-packages/construct/formats/__init__.py
> -${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/construct/formats/${MODPY_PYCACHE}/
> -lib/python${MODPY_VERSION}/site-packages/construct/formats/${MODPY_PYCACHE}__init__.${MODPY_PYC_MAGIC_TAG}pyc
> -lib/python${MODPY_VERSION}/site-packages/construct/formats/data/
> -lib/python${MODPY_VERSION}/site-packages/construct/formats/data/__init__.py
> -${MODPY_COMMENT}lib/python${MODPY_VERSION}/site-packages/construct/formats/data/${MODPY_PYCACHE}/
>
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 08:14:10 Modified files: graphics/py-Pillow: Makefile distinfo Log message: update to py3-Pillow-8.3.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:52 Modified files: lang/php/8.0 : Tag: OPENBSD_6_9 Makefile distinfo lang/php/8.0/patches: Tag: OPENBSD_6_9 patch-configure_ac patch-sapi_fpm_fpm_fpm_atomic_h Log message: update to PHP 8.0.8
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:48 Modified files: lang/php/7.4 : Tag: OPENBSD_6_9 Makefile distinfo Log message: update to PHP 7.4.21
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:45 Modified files: lang/php/7.3 : Tag: OPENBSD_6_9 Makefile distinfo lang/php/7.3/patches: Tag: OPENBSD_6_9 patch-ext_imap_php_imap_c Log message: update to PHP 7.3.29
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:12 Modified files: lang/php/8.0 : Makefile distinfo lang/php/8.0/patches: patch-configure_ac patch-sapi_fpm_fpm_fpm_atomic_h Log message: update to PHP 8.0.8
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:08 Modified files: lang/php/7.4 : Makefile distinfo Log message: update to PHP 7.4.21
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 07:53:04 Modified files: lang/php/7.3 : Makefile distinfo lang/php/7.3/patches: patch-ext_imap_php_imap_c Log message: update to PHP 7.3.29
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: p...@cvs.openbsd.org2021/07/02 07:52:27 Modified files: net/mattermost-server: Makefile distinfo net/mattermost-server/patches: patch-config_config_json net/mattermost-server/pkg: PLIST Log message: Update mattermost-server 5.35.1 -> 5.36.1 Changelog: https://docs.mattermost.com/administration/changelog.html
Re: Convenience changes for Go
Daniel Jakots writes:
> On Wed, 30 Jun 2021 14:24:24 -0600, Aaron Bieber
> wrote:
>
> Thanks for working on it! I see a few possible improvements:
>
>> -# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES
>> +# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES for
>> +# the latest version of a given MODGO_MODNAME if MODGO_VERSION is set to
>> +# "current". Otherwise it will fetch the MODULES/MODFILES for the presently
>> +# set MODGO_VERSION.
>
> Can we use "latest" instead of current? That's the terminology upstream
> uses:
> "The string latest, which selects the highest available release
> version."
>
> https://golang.org/ref/mod#version-queries
>
>> modgo-gen-modules:
>> .if empty(MODGO_MODNAME)
>> @${ECHO_MSG} "No MODGO_MODNAME set"
>> @exit 1
>> .else
>> +.if ${MODGO_VERSION} == "current"
>> @${_PERLSCRIPT}/modgo-gen-modules-helper ${MODGO_MODNAME}
>> +.else
>> +@${_PERLSCRIPT}/modgo-gen-modules-helper ${MODGO_MODNAME}
>> ${MODGO_VERSION} .endif
>> +.endif
>
> We're nesting ifs here. We can put the new one outside easily, but if
> you don't want to do that, the second one should be indented like the
> one L152.
>
> My main disagreement is that if you don't have any MODGO_VERSION set,
> it fails with
> $ make modgo-gen-modules
> *** Parse error in /usr/ports/security/age: Malformed conditional
> (${MODGO_VERSION} == "current") (/usr/ports/lang/go/go.port.mk:189)
> *** Parse error: Need an operator in '"current"'
> (/usr/ports/lang/go/go.port.mk:189)
>
> I tried setting something like
> MODGO_VERSION ?= "current" but make(1) didn't like it.
>
> Here's an updated diff for the two first points. It doesn't improve the
> empty MODGO_VERSION tho :(
>
> BTW, go-module(5) will need an update.
>
> Index: go.port.mk
> ===
> RCS file: /cvs/ports/lang/go/go.port.mk,v
> retrieving revision 1.42
> diff -u -p -r1.42 go.port.mk
> --- go.port.mk23 Mar 2021 13:19:08 - 1.42
> +++ go.port.mk30 Jun 2021 21:23:38 -
> @@ -71,7 +71,7 @@ DISTNAME_ESC = ${DISTNAME${_s}}
> EXTRACT_SUFX ?= .zip
> PKGNAME ?= ${DISTNAME:S/-v/-/}
> ALL_TARGET ?=${MODGO_MODNAME}
> -MODGO_FLAGS += -modcacherw
> +MODGO_FLAGS += -modcacherw -trimpath
> DISTFILES +=
> ${DISTNAME_ESC}${EXTRACT_SUFX}{${MODGO_VERSION}${EXTRACT_SUFX}}
> EXTRACT_ONLY = ${DISTNAME_ESC}${EXTRACT_SUFX}
> MASTER_SITES ?= ${MASTER_SITE_ATHENS}${MODGO_MODNAME_ESC}/@v/
> @@ -177,11 +177,17 @@ do-test:
> . endif
> .endif
>
> -# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES
> +# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES for
> +# the latest version of a given MODGO_MODNAME if MODGO_VERSION is set to
> +# "latest". Otherwise it will fetch the MODULES/MODFILES for the presently
> +# set MODGO_VERSION.
> modgo-gen-modules:
> .if empty(MODGO_MODNAME)
> @${ECHO_MSG} "No MODGO_MODNAME set"
> @exit 1
> -.else
> +.endif
> +.if ${MODGO_VERSION} == "latest"
> @${_PERLSCRIPT}/modgo-gen-modules-helper ${MODGO_MODNAME}
> +.else
> + @${_PERLSCRIPT}/modgo-gen-modules-helper ${MODGO_MODNAME}
> ${MODGO_VERSION}
> .endif
K, here is an update that works with the empty MODGO_VERSION.
diff 6f40cd6bc7cac1369cff9b00b9162c62b1d82490 /usr/ports
blob - 60b84fb0ae310786ee5c71a2e4e4e741f0904da9
file + infrastructure/bin/modgo-gen-modules-helper
--- infrastructure/bin/modgo-gen-modules-helper
+++ infrastructure/bin/modgo-gen-modules-helper
@@ -17,6 +17,7 @@
# OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
use 5.012;
+use strict;
use warnings;
my ($portdir);
@@ -31,7 +32,11 @@ use lib ( "$portdir/infrastructure/lib", "$FindBin::Bi
use OpenBSD::PortGen::Port::Go;
-my ( $module ) = @ARGV;
+die if @ARGV < 1;
+my $module = shift;
+if (@ARGV == 1) {
+ $module .= '@'. shift;
+}
my $port = OpenBSD::PortGen::Port::Go->new;
my $portinfo = $port->get_dist_info($module);
blob - 924c64a51696a59fd6dc5456b1f3f7abda7e38d6
file + lang/go/go.port.mk
--- lang/go/go.port.mk
+++ lang/go/go.port.mk
@@ -71,7 +71,7 @@ DISTNAME_ESC = ${DISTNAME${_s}}
EXTRACT_SUFX ?= .zip
PKGNAME ?= ${DISTNAME:S/-v/-/}
ALL_TARGET ?= ${MODGO_MODNAME}
-MODGO_FLAGS += -modcacherw
+MODGO_FLAGS += -modcacherw -trimpath
DISTFILES += ${DISTNAME_ESC}${EXTRACT_SUFX}{${MODGO_VERSION}${EXTRACT_SUFX}}
EXTRACT_ONLY = ${DISTNAME_ESC}${EXTRACT_SUFX}
MASTER_SITES ?= ${MASTER_SITE_ATHENS}${MODGO_MODNAME_ESC}/@v/
@@ -177,11 +177,18 @@ do-test:
. endif
.endif
-# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES
+# modgo-gen-modules will output MODGO_MODULES and MODGO_MODFILES for
+# the latest version of a given MODGO_MODNAME if MODGO_VERSION is set to
+# "latest". Otherwise it will fetch the MODULES/MODFILES for the presently
+# set MODGO_VERSION.
modgo-gen-modules:
.if
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ben...@cvs.openbsd.org 2021/07/02 06:34:30 Modified files: net/arping : Makefile distinfo Log message: Update to arping-2.22.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: o...@cvs.openbsd.org2021/07/02 05:33:31 Modified files: net/powerdns_recursor: Makefile distinfo Log message: Update to PowerDNS Recursor 4.5.4 With the new Boost the CXXFLAGS are not longer needed
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: rob...@cvs.openbsd.org 2021/07/02 04:53:06 Modified files: net/zabbix : Makefile distinfo net/zabbix/patches: patch-conf_zabbix_server_conf Log message: update to 5.0.13; from Mark Patruck
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 04:33:30 Modified files: net/dhcpcd : Makefile distinfo Removed files: net/dhcpcd/patches: patch-src_route_c patch-src_script_c Log message: update dhcpcd to a newer git checkout, rolling in the patches, and adding some improvements for control socket handling with privsep (one specific thing this fixes is if you run "dhcpcd -U6 badger", i.e. invalid interface name, previously dhcpcd would not accept any more control connections, now it works as expected).
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2021/07/02 04:27:11 Modified files: net/dhcpcd : Makefile net/dhcpcd/patches: patch-src_route_c patch-src_script_c net/dhcpcd/pkg : README Log message: dhcpcd port tweaks; update comments in patches which were committed upstream a while ago add a note to readme based on a user comment
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: ben...@cvs.openbsd.org 2021/07/02 03:21:39 Modified files: graphics/gifsicle: Makefile distinfo Log message: Update to gifsicle-1.93.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: an...@cvs.openbsd.org 2021/07/02 00:34:03 Modified files: mail/mdsort: Makefile distinfo Log message: update to mdsort-10.0.0 - stop requiring explicit new lines in mdsort.conf - support more than one path per maildir declaration - do not confuse macros and date scalars
