Re: CVS: cvs.openbsd.org: ports

[Antoine Jacoutot]

On Wed, Apr 20, 2016 at 12:38:41PM -0600, Robert Nagy wrote:
> CVSROOT:  /cvs
> Module name:  ports
> Changes by:   rob...@cvs.openbsd.org  2016/04/20 12:38:41
> 
> Modified files:
>   mail/zarafa: Makefile.inc 
>   mail/zarafa/zarafa: distinfo 
>   mail/zarafa/zarafa/patches: patch-ECtools_fsck_zarafa-fsck_cpp 
>   patch-ECtools_monitor_zarafa-monitor_cpp 
>   patch-Makefile_in 
>   patch-caldav_CalDAV_cpp 
>   patch-common_ECChannel_cpp 
>   patch-common_ECLogger_cpp 
>   patch-common_charset_localeutil_cpp 
>   patch-common_charset_localeutil_h 
>   patch-common_include_zarafa_ECLogger_h 
>   
> patch-common_include_zarafa_charset_traits_h 
>   patch-common_include_zarafa_platform_h 
>   
> patch-common_include_zarafa_platform_linux_h 
>   patch-common_platform_cpp 
>   patch-common_platform_linux_cpp 
>   patch-common_ustringutil_cpp 
>   patch-configure_ac 
>   patch-doc_Makefile_in 
>   patch-doc_manual_xml 
>   patch-gateway_Gateway_cpp 
>   patch-inetmapi_Makefile_in 
>   patch-installer_linux_server_cfg 
>   patch-installer_userscripts_users_common_sh 
>   patch-m4lcommon_CommonUtil_cpp 
>   
> patch-m4lcommon_include_zarafa_mapi_ptr_mapi_memory_ptr_h 
>   
> patch-m4lcommon_include_zarafa_mapi_ptr_mapi_object_ptr_h 
>   patch-php-ext_main_cpp 
>   patch-provider_client_ECChangeAdvisor_cpp 
>   patch-provider_client_ECNotifyMaster_cpp 
>   patch-provider_client_WSTransport_cpp 
>   patch-provider_contacts_Makefile_in 
>   
> patch-provider_libserver_ECDatabaseUtils_cpp 
>   patch-provider_libserver_ECSessionGroup_cpp 
>   patch-provider_libserver_ECSession_cpp 
>   patch-provider_libserver_StreamUtil_cpp 
>   patch-provider_plugins_ldappasswords_cpp 
>   patch-provider_server_ECServer_cpp 
>   
> patch-provider_server_ECSoapServerConnection_cpp 
>   patch-provider_server_ECThreadManager_cpp 
>   patch-provider_server_Makefile_in 
>   patch-spooler_DAgent_cpp 
>   patch-spooler_Spooler_cpp 
>   mail/zarafa/zarafa/pkg: PLIST-main 
> 
> Log message:
> update to 7.2.3
> 


1 out of 1 hunks failed--saving rejects to 
provider/libserver/ECStatsTables.cpp.rej


-- 
Antoine

CVS: cvs.openbsd.org: ports

[Antoine Jacoutot]

CVSROOT:/cvs
Module name:ports
Changes by: ajacou...@cvs.openbsd.org   2016/04/21 23:33:15

Modified files:
news/pan   : Makefile 
news/pan/pkg   : PLIST 

Log message:
Bring gtkspell3 support back now that it's supported.
Drop unused post-configure.
Regen PLIST+WANTLIB.

ok pea@ (maintainer)

Re: sysutils/sysclean 1.5

[Sebastien Marie]

On Thu, Apr 21, 2016 at 08:50:14PM +0200, Joerg Jung wrote:
> 
> Do you plan to remove /usr/local from defaults ignore list?
> 

It is something I considered but it would have some drawbacks too.

First, it isn't in initial scope of sysclean, which was started for
cleaning base only.

>From functionnality point of vue, it will clash a bit with pkg_check(8).

With the perl version, if the performance impact for integrate
/usr/local in the cleanup is thiner, the filesystem walk will still have
to enter in a big directory: on my host it is 20% more for walk time.

I have no specific opinion in favor or against it.
-- 
Sebastien Marie

CVS: cvs.openbsd.org: ports

[Sebastien Marie]

CVSROOT:/cvs
Module name:ports
Changes by: sema...@cvs.openbsd.org 2016/04/21 22:40:19

Modified files:
sysutils/sysclean: Makefile distinfo 

Log message:
update sysutils/sysclean to 1.5

ok jung@

Re: NEW: lua-lyaml

[Jonathan Gray]

On Fri, Apr 22, 2016 at 01:45:06PM +1000, Jonathan Gray wrote:
> On Thu, Apr 21, 2016 at 07:08:13PM +0200, Florian Stinglmayr wrote:
> > On Thu, Apr 21, 2016 at 08:47:46PM +1000, Jonathan Gray wrote:
> > > This could now drop the SHARED_ONLY part.
> > >
> > 
> > Yes.
> > 
> > > Why the need to rerun autoconf?  Could configure be used as is?  Currently
> > > the build fails here with:
> > >
> > > ===>  Configuring for lua-lyaml-6.0
> > > Using /usr/ports/pobj/lua-lyaml-6.0/config.site (generated)
> > > /usr/local/bin/autoreconf[35]: /usr/local/bin/autoreconf-2.13: not found
> > >
> > 
> > You are correct. The previous release didn't have a configure, and was
> > just a .tar.gz from git master. The appropriate lines have been removed,
> > and it still builds and works fine.
> > 
> > Regards,
> > Florian
> > 
> 
> The WRKDIST/LIBTOOL_FLAGS lines could go as well.
> 
> And libyaml should be LIB_DEPENDS with WANTLIB.
> 
> I added LUA_INCLUDE to env as below also.
> 
> I'd like to commit it with these changes unless there are any objections?

Thinking about it some more "lua-yaml" makes more sense than
"lua-lyaml".

This also fixes the distfile name to include a version.

diff -upr lua-lyaml.orig/Makefile lua-yaml/Makefile
--- lua-lyaml.orig/Makefile Fri Apr 22 13:14:20 2016
+++ lua-yaml/Makefile   Fri Apr 22 14:09:56 2016
@@ -3,9 +3,7 @@
 COMMENT =  Lua library for parsing and emitting YAML
 V =6.0
 
-DISTNAME = lua-${GH_PROJECT}
-PKGNAME =  ${DISTNAME}-${V}
-WRKDIST =  ${WRKDIR}/${GH_PROJECT}-${GH_TAGNAME}
+PKGNAME =  lua-yaml-${V}
 
 GH_ACCOUNT =   gvvaughan
 GH_PROJECT =   lyaml
@@ -18,15 +16,16 @@ MAINTAINER =Florian Stinglmayr 
 # MIT
 PERMIT_PACKAGE_CDROM = Yes
 
+WANTLIB =  yaml-0
 MODULES =  lang/lua
 
-BUILD_DEPENDS =devel/libyaml
+LIB_DEPENDS =  devel/libyaml
 
 CONFIGURE_STYLE =  gnu
-CONFIGURE_ENV =CPPFLAGS="-I${PREFIX}/include -L${PREFIX}/lib" \
-   LUA="${MODLUA_BIN}"
-
-LIBTOOL_FLAGS +=   --tag=disable-static
+CONFIGURE_ENV =CPPFLAGS="-I${PREFIX}/include" \
+   LDFLAGS="-L${PREFIX}/lib" \
+   LUA=${MODLUA_BIN} \
+   LUA_INCLUDE=-I${MODLUA_INCL_DIR}
 
 # Test requires specl, specl requires lyaml.
 NO_TEST =  Yes
diff -upr lua-lyaml.orig/distinfo lua-yaml/distinfo
--- lua-lyaml.orig/distinfo Fri Apr 22 13:14:20 2016
+++ lua-yaml/distinfo   Fri Apr 22 14:09:56 2016
@@ -1,2 +1,2 @@
-SHA256 (lua-lyaml.tar.gz) = nXz3TXdpmf9vdYxWnVIC/12h8wPG9CKdO0H3HNOj56c=
-SIZE (lua-lyaml.tar.gz) = 475288
+SHA256 (lyaml-release-v6.0.tar.gz) = 
nXz3TXdpmf9vdYxWnVIC/12h8wPG9CKdO0H3HNOj56c=
+SIZE (lyaml-release-v6.0.tar.gz) = 475288

Re: NEW: lua-lyaml

[Jonathan Gray]

On Thu, Apr 21, 2016 at 07:08:13PM +0200, Florian Stinglmayr wrote:
> On Thu, Apr 21, 2016 at 08:47:46PM +1000, Jonathan Gray wrote:
> > This could now drop the SHARED_ONLY part.
> >
> 
> Yes.
> 
> > Why the need to rerun autoconf?  Could configure be used as is?  Currently
> > the build fails here with:
> >
> > ===>  Configuring for lua-lyaml-6.0
> > Using /usr/ports/pobj/lua-lyaml-6.0/config.site (generated)
> > /usr/local/bin/autoreconf[35]: /usr/local/bin/autoreconf-2.13: not found
> >
> 
> You are correct. The previous release didn't have a configure, and was
> just a .tar.gz from git master. The appropriate lines have been removed,
> and it still builds and works fine.
> 
> Regards,
> Florian
> 

The WRKDIST/LIBTOOL_FLAGS lines could go as well.

And libyaml should be LIB_DEPENDS with WANTLIB.

I added LUA_INCLUDE to env as below also.

I'd like to commit it with these changes unless there are any objections?

diff -upr lua-lyaml.orig/Makefile lua-lyaml/Makefile
--- lua-lyaml.orig/Makefile Fri Apr 22 13:14:20 2016
+++ lua-lyaml/Makefile  Fri Apr 22 13:41:24 2016
@@ -5,7 +5,6 @@ V = 6.0
 
 DISTNAME = lua-${GH_PROJECT}
 PKGNAME =  ${DISTNAME}-${V}
-WRKDIST =  ${WRKDIR}/${GH_PROJECT}-${GH_TAGNAME}
 
 GH_ACCOUNT =   gvvaughan
 GH_PROJECT =   lyaml
@@ -18,15 +17,16 @@ MAINTAINER =Florian Stinglmayr 
 # MIT
 PERMIT_PACKAGE_CDROM = Yes
 
+WANTLIB =  yaml-0
 MODULES =  lang/lua
 
-BUILD_DEPENDS =devel/libyaml
+LIB_DEPENDS =  devel/libyaml
 
 CONFIGURE_STYLE =  gnu
-CONFIGURE_ENV =CPPFLAGS="-I${PREFIX}/include -L${PREFIX}/lib" \
-   LUA="${MODLUA_BIN}"
-
-LIBTOOL_FLAGS +=   --tag=disable-static
+CONFIGURE_ENV =CPPFLAGS="-I${PREFIX}/include" \
+   LDFLAGS="-L${PREFIX}/lib" \
+   LUA=${MODLUA_BIN} \
+   LUA_INCLUDE=-I${MODLUA_INCL_DIR}
 
 # Test requires specl, specl requires lyaml.
 NO_TEST =  Yes

Re: update sysutils/sshfs-fuse, move to GitHub

[Michael McConville]

Dmitrij D. Czarkoff wrote:
> This diff sets PKGNAME and uses explicit HOMEPAGE.  OK?

I think you forgot the diff.

In response to this:

Stuart Henderson said:
> And it changes the package name from sshfs-fuse to just sshfs.

That ship sailed when we changed DISTNAME for the GitHub mirror.
Regardless, it's probably easiest to just set HOMEPAGE.

Sorry for taking us on so many needless diversions. Here's what I have
now.


Index: Makefile
===
RCS file: /cvs/ports/sysutils/sshfs-fuse/Makefile,v
retrieving revision 1.5
diff -u -p -u -r1.5 Makefile
--- Makefile25 Sep 2015 07:17:57 -  1.5
+++ Makefile22 Apr 2016 03:20:22 -
@@ -2,11 +2,13 @@
 
 COMMENT =  mount remote directories over ssh
 
-DISTNAME = sshfs-fuse-2.5
+V =2.7
+DISTNAME = sshfs-${V}
+PKGNAME =  sshfs-fuse-${V}
 CATEGORIES =   sysutils
 
-HOMEPAGE = http://fuse.sourceforge.net/sshfs.html
-MASTER_SITES = ${MASTER_SITE_SOURCEFORGE:=fuse/}
+HOMEPAGE = https://github.com/libfuse/sshfs
+MASTER_SITES = https://github.com/libfuse/sshfs/releases/download/${DISTNAME}/
 
 MAINTAINER =   Gallon Sylvestre 
 
Index: distinfo
===
RCS file: /cvs/ports/sysutils/sshfs-fuse/distinfo,v
retrieving revision 1.2
diff -u -p -u -r1.2 distinfo
--- distinfo25 Sep 2015 07:17:57 -  1.2
+++ distinfo22 Apr 2016 03:20:22 -
@@ -1,2 +1,2 @@
-SHA256 (sshfs-fuse-2.5.tar.gz) = 6RcUUuXQFQucaiFY/S4tzvtdXQO6TSCJSeAKOkbG5j4=
-SIZE (sshfs-fuse-2.5.tar.gz) = 136378
+SHA256 (sshfs-2.7.tar.gz) = zl+dOQk0/IjCh92TejnRXCtAy+ppZ7okrZPtkOACcSg=
+SIZE (sshfs-2.7.tar.gz) = 153014
Index: patches/patch-sshfs_c
===
RCS file: /cvs/ports/sysutils/sshfs-fuse/patches/patch-sshfs_c,v
retrieving revision 1.3
diff -u -p -u -r1.3 patch-sshfs_c
--- patches/patch-sshfs_c   25 Sep 2015 07:17:57 -  1.3
+++ patches/patch-sshfs_c   22 Apr 2016 03:20:22 -
@@ -1,15 +1,15 @@
 $OpenBSD: patch-sshfs_c,v 1.3 2015/09/25 07:17:57 landry Exp $
 sshfs.c.orig   Wed Jan  8 16:34:52 2014
-+++ sshfs.cFri Sep 25 09:16:48 2015
+--- sshfs.c.orig   Tue Mar  1 11:49:53 2016
 sshfs.cFri Apr 15 12:51:26 2016
 @@ -11,7 +11,6 @@
  
  #include 
  #include 
 -#include 
- #include 
- #include 
- #include 
-@@ -32,6 +31,7 @@
+ #ifdef __APPLE__
+ #  include 
+ #endif
+@@ -37,6 +36,7 @@
  #include 
  #include 
  #include 
@@ -17,7 +17,7 @@ $OpenBSD: patch-sshfs_c,v 1.3 2015/09/25
  #include 
  #include 
  #include 
-@@ -1552,7 +1552,7 @@ static int sftp_error_to_errno(uint32_t error)
+@@ -1615,7 +1615,7 @@ static int sftp_error_to_errno(uint32_t error)
case SSH_FX_NO_SUCH_FILE:  return ENOENT;
case SSH_FX_PERMISSION_DENIED: return EACCES;
case SSH_FX_FAILURE:   return EPERM;

CVS: cvs.openbsd.org: ports

[Gleydson Soares]

CVSROOT:/cvs
Module name:ports
Changes by: gsoa...@cvs.openbsd.org 2016/04/21 20:43:05

Modified files:
plan9/plan9port: Makefile distinfo 

Log message:
update to latest git commit

UPDATE: security/stunnel-5.31

[Gleydson Soares]

following diff updates stunnel to the latest release. works here @amd64
it was slightly tested... so I'm looking for test reports, someone willing? OKs?

Index: Makefile
===
RCS file: /cvs/ports/security/stunnel/Makefile,v
retrieving revision 1.82
diff -u -p -r1.82 Makefile
--- Makefile11 Mar 2016 20:28:31 -  1.82
+++ Makefile22 Apr 2016 01:06:22 -
@@ -2,7 +2,7 @@
 
 COMMENT=   SSL encryption wrapper for standard network daemons
 
-DISTNAME=  stunnel-5.29
+DISTNAME=  stunnel-5.31
 CATEGORIES=security
 
 MAINTAINER=Gleydson Soares 
Index: distinfo
===
RCS file: /cvs/ports/security/stunnel/distinfo,v
retrieving revision 1.39
diff -u -p -r1.39 distinfo
--- distinfo13 Jan 2016 12:39:55 -  1.39
+++ distinfo22 Apr 2016 01:06:22 -
@@ -1,2 +1,2 @@
-SHA256 (stunnel-5.29.tar.gz) = Q5CWJUA+pjT6fLg5nVj6+OfxHBt7KQl0kUaZUfVt9VE=
-SIZE (stunnel-5.29.tar.gz) = 637806
+SHA256 (stunnel-5.31.tar.gz) = p0a3GrPcbCPqyw2vc0JGeHDkOskzQwkF6xsdBQu64Lc=
+SIZE (stunnel-5.31.tar.gz) = 639381
cvs server: Diffing patches
Index: patches/patch-src_ctx_c
===
RCS file: /cvs/ports/security/stunnel/patches/patch-src_ctx_c,v
retrieving revision 1.2
diff -u -p -r1.2 patch-src_ctx_c
--- patches/patch-src_ctx_c 4 Dec 2015 01:38:56 -   1.2
+++ patches/patch-src_ctx_c 22 Apr 2016 01:06:22 -
@@ -1,7 +1,7 @@
 $OpenBSD: patch-src_ctx_c,v 1.2 2015/12/04 01:38:56 gsoares Exp $
 src/ctx.c.orig Mon Nov 30 11:15:43 2015
-+++ src/ctx.c  Thu Dec  3 19:26:54 2015
-@@ -351,7 +351,7 @@ NOEXPORT int ecdh_init(SERVICE_OPTIONS *section) {
+--- src/ctx.c.orig Fri Feb 19 16:18:43 2016
 src/ctx.c  Thu Apr 21 21:59:31 2016
+@@ -363,7 +363,7 @@ NOEXPORT int ecdh_init(SERVICE_OPTIONS *section) {
  / initialize OpenSSL CONF */
  
  NOEXPORT int conf_init(SERVICE_OPTIONS *section) {
Index: patches/patch-src_options_c
===
RCS file: /cvs/ports/security/stunnel/patches/patch-src_options_c,v
retrieving revision 1.5
diff -u -p -r1.5 patch-src_options_c
--- patches/patch-src_options_c 13 Jan 2016 12:39:55 -  1.5
+++ patches/patch-src_options_c 22 Apr 2016 01:06:22 -
@@ -3,9 +3,9 @@ $OpenBSD: patch-src_options_c,v 1.5 2016
 use SSLv23_client_method() required to build with libressl since that it 
haven't
 TLS_client_method()  for now.
 
 src/options.c.orig Tue Jan  5 06:20:12 2016
-+++ src/options.c  Tue Jan 12 19:57:30 2016
-@@ -2559,7 +2559,7 @@ NOEXPORT char *parse_service_option(CMD cmd, SERVICE_O
+--- src/options.c.orig Fri Jan 15 13:45:23 2016
 src/options.c  Thu Apr 21 21:59:31 2016
+@@ -2579,7 +2579,7 @@ NOEXPORT char *parse_service_option(CMD cmd, SERVICE_O
  /* sslVersion */
  switch(cmd) {
  case CMD_BEGIN:
@@ -14,7 +14,7 @@ TLS_client_method()  for now.
  section->client_method=(SSL_METHOD *)TLS_client_method();
  section->server_method=(SSL_METHOD *)TLS_server_method();
  #else
-@@ -2571,7 +2571,7 @@ NOEXPORT char *parse_service_option(CMD cmd, SERVICE_O
+@@ -2591,7 +2591,7 @@ NOEXPORT char *parse_service_option(CMD cmd, SERVICE_O
  if(strcasecmp(opt, "sslVersion"))
  break;
  if(!strcasecmp(arg, "all")) {
Index: patches/patch-src_verify_c
===
RCS file: /cvs/ports/security/stunnel/patches/patch-src_verify_c,v
retrieving revision 1.2
diff -u -p -r1.2 patch-src_verify_c
--- patches/patch-src_verify_c  14 Oct 2015 02:05:06 -  1.2
+++ patches/patch-src_verify_c  22 Apr 2016 01:06:22 -
@@ -3,8 +3,8 @@ $OpenBSD: patch-src_verify_c,v 1.2 2015/
 disable OpenSSL 1.0.2 X509_check_* functions, so it can build with libressl.
 
 
 src/verify.c.orig  Wed Sep 23 07:00:08 2015
-+++ src/verify.c   Tue Oct  6 05:52:42 2015
+--- src/verify.c.orig  Fri Feb 19 16:18:43 2016
 src/verify.c   Thu Apr 21 21:59:31 2016
 @@ -51,7 +51,7 @@ NOEXPORT int add_dir_lookup(X509_STORE *, char *);
  NOEXPORT int verify_callback(int, X509_STORE_CTX *);
  NOEXPORT int verify_checks(CLI *, int, X509_STORE_CTX *);
@@ -14,7 +14,7 @@ disable OpenSSL 1.0.2 X509_check_* funct
  NOEXPORT int cert_check_subject(CLI *, X509_STORE_CTX *);
  #endif /* OPENSSL_VERSION_NUMBER>=0x10002000L */
  NOEXPORT int cert_check_local(X509_STORE_CTX *);
-@@ -280,7 +280,7 @@ NOEXPORT int cert_check(CLI *c, X509_STORE_CTX *callba
+@@ -274,7 +274,7 @@ NOEXPORT int cert_check(CLI *c, X509_STORE_CTX *callba
  }
  
  if(depth==0) { /* additional peer certificate checks */
@@ -23,7 +23,7 @@ disable OpenSSL 1.0.2 X509_check_* funct
  if(!cert_check_subject(c, callback_ctx))
  return 0; /* reject */
  #endif /* 

CVS: cvs.openbsd.org: ports

[Gleydson Soares]

CVSROOT:/cvs
Module name:ports
Changes by: gsoa...@cvs.openbsd.org 2016/04/21 18:50:25

Modified files:
devel/cppcheck : Makefile distinfo 

Log message:
update to 1.73

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/04/21 17:46:47

Modified files:
net/rrdtool: Makefile distinfo 
net/rrdtool/patches: patch-bindings_Makefile_in patch-configure 
 patch-doc_Makefile_in 
 patch-examples_Makefile_in 
 patch-examples_rrdcached_Makefile_in 
 patch-src_Makefile_in 
net/rrdtool/pkg: PLIST-main 
Removed files:
net/rrdtool/patches: patch-src_rrd_daemon_c 

Log message:
update to rrdtool-1.6.0, librrd_th is removed,
the main library is now meant to be thread-safe

Re: Disable SSE3 in graphics/opencv

[Michael McConville]

Michael McConville wrote:
> Stuart Henderson wrote:
> > On 2016/04/21 20:58, Rafael Sadowski wrote:
> > > On Wed Apr 20, 2016 at 05:30:26AM -0400, Michael McConville wrote:
> > > > It's on by default, but can be disabled with the below cmake
> > > > argument.
> > > 
> > > Are test results with and without SSE3 same or better? If yes okay
> > > for me.
> > 
> > I think it would have to be "better" otherwise what's the point in
> > disabling this in opencv (which does runtime detection)?
> 
> Are you sure? Looking at the build logs, it seems like -msse3 is in the
> build-global C and C++ flags, so it'll be used for autovectorization
> (see build snippet below).
> 
> It looks like Debian explicitly disables it:
> 
> https://anonscm.debian.org/cgit/debian-science/packages/opencv.git/tree/debian/rules#n16
> 
> And FreeBSD treats SIMD extensions as if they weren't autodetected:
> 
> https://svnweb.freebsd.org/ports/head/graphics/opencv/Makefile?view=co

For what it's worth, this probably applies to us too:

https://sources.debian.net/src/opencv/2.4.9.1%2Bdfsg-1.5/debian/patches/optimize_i586.patch/

Re: Disable SSE3 in graphics/opencv

[Michael McConville]

Stuart Henderson wrote:
> On 2016/04/21 20:58, Rafael Sadowski wrote:
> > On Wed Apr 20, 2016 at 05:30:26AM -0400, Michael McConville wrote:
> > > It's on by default, but can be disabled with the below cmake
> > > argument.
> > 
> > Are test results with and without SSE3 same or better? If yes okay
> > for me.
> 
> I think it would have to be "better" otherwise what's the point in
> disabling this in opencv (which does runtime detection)?

Are you sure? Looking at the build logs, it seems like -msse3 is in the
build-global C and C++ flags, so it'll be used for autovectorization
(see build snippet below).

It looks like Debian explicitly disables it:

https://anonscm.debian.org/cgit/debian-science/packages/opencv.git/tree/debian/rules#n16

And FreeBSD treats SIMD extensions as if they weren't autodetected:

https://svnweb.freebsd.org/ports/head/graphics/opencv/Makefile?view=co


> -- C++ Compiler:/usr/bin/c++  (ver 4.2.1)
> -- C++ flags (Release): -O2 -pipe  -fsigned-char -W -Wall 
> -Werror=return-type -Werror=address -Werror=sequence-point -Wformat 
> -Werror=format-security -W
> undef -Winit-self -Wpointer-arith -Wshadow -Wsign-promo 
> -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse 
> -msse2 -msse3 -ffunction-sections -
> DNDEBUG  -DNDEBUG
> -- C++ flags (Debug):   -O2 -pipe  -fsigned-char -W -Wall 
> -Werror=return-type -Werror=address -Werror=sequence-point -Wformat 
> -Werror=format-security -W
> undef -Winit-self -Wpointer-arith -Wshadow -Wsign-promo 
> -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse 
> -msse2 -msse3 -ffunction-sections -
> g  -O0 -DDEBUG -D_DEBUG
> -- C Compiler:  /usr/bin/cc
> -- C flags (Release):   -O2 -pipe  -fsigned-char -W -Wall 
> -Werror=return-type -Werror=address -Werror=sequence-point -Wformat 
> -Werror=format-security -W
> missing-declarations -Wmissing-prototypes -Wstrict-prototypes -Wundef 
> -Winit-self -Wpointer-arith -Wshadow -fdiagnostics-show-option -Wno-long-long 
> -pthread -fomit-fram
> e-pointer -msse -msse2 -msse3 -ffunction-sections -DNDEBUG  -DNDEBUG
> -- C flags (Debug): -O2 -pipe  -fsigned-char -W -Wall 
> -Werror=return-type -Werror=address -Werror=sequence-point -Wformat 
> -Werror=format-security -W
> missing-declarations -Wmissing-prototypes -Wstrict-prototypes -Wundef 
> -Winit-self -Wpointer-arith -Wshadow -fdiagnostics-show-option -Wno-long-long 
> -pthread -fomit-fram
> e-pointer -msse -msse2 -msse3 -ffunction-sections -g  -O0 -DDEBUG -D_DEBUG

Re: x11/sxhkd

[ix]

I forgot to attach the port.

msmtp isn't cooperating with me at the moment, so here's a .tar.gz.
https://arcetera.moe/pub/sxhkd.tar.gz

On Thu, Apr 21, 2016 at 02:09:51PM -0400, ix wrote:
> Hi,
>
> A few weeks ago I sent in a port for sxhkd 
> (https://github.com/baskerville/sxhkd).
>
> sxhkd is a hotkey daemon for x11 that supports a simple config syntax.
>
> I've fixed the port so portcheck accepts it, and I haven't gotten any 
> negative feedback,
> so can someone commit the port?
>
> Thanks.
>

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/04/21 16:09:53

Modified files:
net/dhcpcd : Makefile distinfo 
Removed files:
net/dhcpcd/patches: patch-dhcp6_c 

Log message:
update to dhcpcd-6.10.3

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Stuart Henderson]

On 2016/04/21 23:10, Dmitrij D. Czarkoff wrote:
> Stuart Henderson said:
> > On 2016/04/21 19:35, Christian Weisgerber wrote:
> >> Dmitrij D. Czarkoff:
> >> 
>  I think you should just set
>  INSTALL_STRIP =
>  in the ports (or modules) that require it and be done with it.
>  No point in introducing yet another variable.
> >>> 
> >>> INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
> >>> /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
> >>> INSTALL_STRIP= from port.
> >> 
> >> In practice, INSTALL_STRIP has two possible values: "-s" and "".
> >> "-s" is already the default, so the only user setting that makes
> >> any sense is "".  Having some ports already set it to "" is just
> >> fine.
> >> 
> >> A few ports already do this.
> > 
> > Makes sense to me. go-bootstrap needs it too.
> 
> Well, then asking for OKs for the following patch.

Apart from logstash/forwarder these will be changing the packages, so
need bumps. Otherwise OK with me.

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Stuart Henderson said:
> On 2016/04/21 23:19, Dmitrij D. Czarkoff wrote:
> > 
> > While at it, maybe we should move description of INSTALL_STRIP from
> > mk.conf(5) to bsd.port.mk(5), as apparently it is now used this way?
> 
> Should stay in mk.conf(5), it affects src too.

To my understanding, in src it is only useful with DEBUG, which unsets
it automatically.

-- 
Dmitrij D. Czarkoff

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Jeremie Courreges-Anglas]

"Dmitrij D. Czarkoff"  writes:

> Dmitrij D. Czarkoff said:
>> Stuart Henderson said:
>> > On 2016/04/21 19:35, Christian Weisgerber wrote:
>> >> Dmitrij D. Czarkoff:
>> >> 
>>  I think you should just set
>>  INSTALL_STRIP =
>>  in the ports (or modules) that require it and be done with it.
>>  No point in introducing yet another variable.
>> >>> 
>> >>> INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
>> >>> /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
>> >>> INSTALL_STRIP= from port.
>> >> 
>> >> In practice, INSTALL_STRIP has two possible values: "-s" and "".
>> >> "-s" is already the default, so the only user setting that makes
>> >> any sense is "".  Having some ports already set it to "" is just
>> >> fine.
>> >> 
>> >> A few ports already do this.
>> > 
>> > Makes sense to me. go-bootstrap needs it too.
>> 
>> Well, then asking for OKs for the following patch.
>
> While at it, maybe we should move description of INSTALL_STRIP from
> mk.conf(5) to bsd.port.mk(5), as apparently it is now used this way?

I don't think so, INSTALL_STRIP comes from bsd.own.mk.

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Stuart Henderson]

On 2016/04/21 23:19, Dmitrij D. Czarkoff wrote:
> 
> While at it, maybe we should move description of INSTALL_STRIP from
> mk.conf(5) to bsd.port.mk(5), as apparently it is now used this way?

Should stay in mk.conf(5), it affects src too.

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/04/21 15:19:17

Modified files:
net/icinga/core2: Makefile distinfo 

Log message:
update to icinga2-2.4.7

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Dmitrij D. Czarkoff said:
> Stuart Henderson said:
> > On 2016/04/21 19:35, Christian Weisgerber wrote:
> >> Dmitrij D. Czarkoff:
> >> 
>  I think you should just set
>  INSTALL_STRIP =
>  in the ports (or modules) that require it and be done with it.
>  No point in introducing yet another variable.
> >>> 
> >>> INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
> >>> /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
> >>> INSTALL_STRIP= from port.
> >> 
> >> In practice, INSTALL_STRIP has two possible values: "-s" and "".
> >> "-s" is already the default, so the only user setting that makes
> >> any sense is "".  Having some ports already set it to "" is just
> >> fine.
> >> 
> >> A few ports already do this.
> > 
> > Makes sense to me. go-bootstrap needs it too.
> 
> Well, then asking for OKs for the following patch.

While at it, maybe we should move description of INSTALL_STRIP from
mk.conf(5) to bsd.port.mk(5), as apparently it is now used this way?

-- 
Dmitrij D. Czarkoff

Index: bsd.port.mk.5
===
RCS file: /cvs/src/share/man/man5/bsd.port.mk.5,v
retrieving revision 1.434
diff -u -p -r1.434 bsd.port.mk.5
--- bsd.port.mk.5   22 Mar 2016 16:02:02 -  1.434
+++ bsd.port.mk.5   21 Apr 2016 21:16:27 -
@@ -1743,6 +1743,14 @@ User settings.
 If set to
 .Sq Yes ,
 do not print anything when ignoring a port.
+.It Ev INSTALL_STRIP
+Flag for
+.Ev INSTALL_PROGRAM
+macro for stripping binaries.
+Default is
+.Sq -s .
+Set to empty value to disable stripping.
+.Sq
 .It Ev INSTALL_{PROGRAM,SCRIPT,DATA,MAN}[_DIR]
 Macros to use to install a program, a script, data, or a man page (or the
 corresponding directory), respectively.
Index: mk.conf.5
===
RCS file: /cvs/src/share/man/man5/mk.conf.5,v
retrieving revision 1.31
diff -u -p -r1.31 mk.conf.5
--- mk.conf.5   31 Mar 2016 15:53:25 -  1.31
+++ mk.conf.5   21 Apr 2016 21:13:55 -
@@ -64,7 +64,7 @@ The following variables are set by
 .In bsd.own.mk ,
 if they are not already defined.
 Defaults are in brackets.
-.Bl -tag -width INSTALL_STRIP
+.Bl -tag -width INSTALL_COPY
 .It Ev BINGRP
 Binary group.
 .Bq bin
@@ -110,16 +110,6 @@ This is to be used when building an inst
 system can either be installed with copies, or copy-if-different using a
 single knob.
 .Bq Fl c
-.It Ev INSTALL_STRIP
-The flag passed to the install program to cause the binary to be stripped.
-This is to be used when building an install script so that the entire
-system can be made stripped/not-stripped using a single knob.
-Note that
-.Ev INSTALL_STRIP
-is not set if
-.Ev ${DEBUG}
-is defined.
-.Bq Fl s
 .It Ev LIBDIR
 Base path for library installation.
 .Bq Pa /usr/lib

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Stuart Henderson said:
> On 2016/04/21 19:35, Christian Weisgerber wrote:
>> Dmitrij D. Czarkoff:
>> 
 I think you should just set
 INSTALL_STRIP =
 in the ports (or modules) that require it and be done with it.
 No point in introducing yet another variable.
>>> 
>>> INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
>>> /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
>>> INSTALL_STRIP= from port.
>> 
>> In practice, INSTALL_STRIP has two possible values: "-s" and "".
>> "-s" is already the default, so the only user setting that makes
>> any sense is "".  Having some ports already set it to "" is just
>> fine.
>> 
>> A few ports already do this.
> 
> Makes sense to me. go-bootstrap needs it too.

Well, then asking for OKs for the following patch.

-- 
Dmitrij D. Czarkoff

Index: lang/go/Makefile
===
RCS file: /cvs/ports/lang/go/Makefile,v
retrieving revision 1.29
diff -u -p -r1.29 Makefile
--- lang/go/Makefile14 Apr 2016 17:43:26 -  1.29
+++ lang/go/Makefile21 Apr 2016 20:52:41 -
@@ -30,6 +30,8 @@ SUBST_VARS =  GOCFG
 WRKDIST =  ${WRKDIR}/go
 WRKSRC =   ${WRKDIST}/src
 
+INSTALL_STRIP =
+
 GOOS = openbsd
 GOARCH =   unknown
 GOROOT =   ${PREFIX}/go
Index: lang/go/go.port.mk
===
RCS file: /cvs/ports/lang/go/go.port.mk,v
retrieving revision 1.6
diff -u -p -r1.6 go.port.mk
--- lang/go/go.port.mk  28 Feb 2016 13:24:16 -  1.6
+++ lang/go/go.port.mk  21 Apr 2016 21:07:40 -
@@ -53,6 +53,10 @@ MODGO_INSTALL_TARGET =   ${INSTALL_DATA_DI
${MODGO_PACKAGE_PATH};
 .endif
 
+# Although this module does not use INSTALL_PROGRAM, unset INSTALL_STRIP to
+# prevent stripping go programs in ports with custom do-install targets.
+INSTALL_STRIP =
+
 MODGO_TEST_TARGET =${MODGO_TEST_CMD} ${TEST_TARGET}
 
 .if empty(CONFIGURE_STYLE)
Index: lang/go-bootstrap/Makefile
===
RCS file: /cvs/ports/lang/go-bootstrap/Makefile,v
retrieving revision 1.1.1.1
diff -u -p -r1.1.1.1 Makefile
--- lang/go-bootstrap/Makefile  4 Dec 2015 17:19:18 -   1.1.1.1
+++ lang/go-bootstrap/Makefile  21 Apr 2016 20:52:56 -
@@ -32,6 +32,8 @@ SUBST_VARS =  GOEXE GOCFG
 WRKDIST =  ${WRKDIR}/go
 WRKSRC =   ${WRKDIST}/src
 
+INSTALL_STRIP =
+
 GOOS = openbsd
 GOARCH =   unknown
 GOROOT =   ${PREFIX}/go/bootstrap
Index: sysutils/logstash/forwarder//Makefile
===
RCS file: /cvs/ports/sysutils/logstash/forwarder/Makefile,v
retrieving revision 1.8
diff -u -p -r1.8 Makefile
--- sysutils/logstash/forwarder//Makefile   21 Apr 2016 00:07:08 -  
1.8
+++ sysutils/logstash/forwarder//Makefile   21 Apr 2016 21:05:45 -
@@ -16,9 +16,8 @@ WANTLIB +=c pthread
 pre-configure:
${SUBST_CMD} ${WRKSRC}/logstash-forwarder.conf.example
 
-# Use INSTALL_SCRIPT to prevent stripping go binaries
 do-install:
-   ${INSTALL_SCRIPT} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
+   ${INSTALL_PROGRAM} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
${PREFIX}/sbin/
${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/logstash-forwarder/
${INSTALL_DATA} ${WRKSRC}/README.md \

Re: [lang/go] missing segments

[Juan Francisco Cantero Hurtado]

On Thu, Apr 21, 2016 at 12:20:17AM +0100, Edd Barrett wrote:
> On Thu, Jul 23, 2015 at 04:05:21PM +0200, Fritjof Bornebusch wrote:
> > Hi ports@,
> > 
> > while working on a port that uses lang/go it looks like there are some 
> > segments missing.
> > The install(1) command says the following:
> > 
> > BFD: /fake-amd64/usr/local/bin/st827dDR: warning: allocated section 
> > `.gosymtab' not in segment
> > BFD: /fake-amd64/usr/local/bin/st827dDR: warning: allocated section 
> > `.gnu.version_r' not in segment
> 
> A few of us have started seeing messages like this. I only started
> seeing these messages when I upgraded from go-1.5.3 to 1.5.4. 
> 
> E.g. 'make fake' in net/syncthing from openbsd-wip:
> 
> ---8<---
> ===>  Faking installation for syncthing-0.12.22
> install -d -m 755 /usr/ports/pobj/syncthing-0.12.22/fake-amd64
> /usr/ports/pobj/syncthing-0.12.22/bin/install -c -s  -m 755 
> /usr/ports/pobj/syncthing-0.12.22/go/src/github.com/syncthing/syncthing/bin/syncthing
>  /usr/ports/pobj/syncthing-0.12.22/fake-amd64/usr/local/bin/
> BFD: /usr/ports/pobj/syncthing-0.12.22/fake-amd64/usr/local/bin//stoBJG4T: 
> warning: allocated section `.gosymtab' not in segment
> BFD: /usr/ports/pobj/syncthing-0.12.22/fake-amd64/usr/local/bin//stoBJG4T: 
> warning: allocated section `.gnu.version_r' not in segment
> /usr/ports/pobj/syncthing-0.12.22/bin/install -c  -m 644 
> /usr/ports/pobj/syncthing-0.12.22/go/src/github.com/syncthing/syncthing/man/*.{1,5,7}
>  /usr/ports/pobj/syncthing-0.12.22/fake-amd64/usr/local/man/
> --->8---
> 
> I also noticed similar stuff when building go itself.

Can you try this patch?


diff --git lang/go/go.port.mk lang/go/go.port.mk
index 2fd3db0..92f24af 100644
--- lang/go/go.port.mk
+++ lang/go/go.port.mk
@@ -33,6 +33,9 @@ TEST_TARGET ?=${ALL_TARGET}
 SEPARATE_BUILD ?=  Yes
 WRKSRC ?=  ${MODGO_WORKSPACE}/src/${ALL_TARGET}
 
+# We never should use the standard "strip" command against go binaries
+INSTALL_STRIP=
+
 MODGO_SETUP_WORKSPACE =mkdir -p ${WRKSRC:H}; mv ${MODGO_SUBDIR} 
${WRKSRC};
 
 CATEGORIES +=  lang/go

CVS: cvs.openbsd.org: ports

[Christian Weisgerber]

CVSROOT:/cvs
Module name:ports
Changes by: na...@cvs.openbsd.org   2016/04/21 14:42:01

Modified files:
shells : Makefile 

Log message:
+yash

CVS: cvs.openbsd.org: ports

[Christian Weisgerber]

CVSROOT:/cvs
Module name:ports
Changes by: na...@cvs.openbsd.org   2016/04/21 14:41:09

Log message:
Import yash 2.41.

Yet another shell (yash) is a POSIX-compliant command line shell, featuring
more strict POSIX compliance than those of other shells, as well as powerful
command line editing.

ok sthen@

Status:

Vendor Tag: naddy
Release Tags:   naddy_20160421

N ports/shells/yash/Makefile
N ports/shells/yash/distinfo
N ports/shells/yash/pkg/DESCR
N ports/shells/yash/pkg/PLIST
N ports/shells/yash/patches/patch-variable_c

No conflicts created by this import

CVS: cvs.openbsd.org: ports

[Christian Weisgerber]

CVSROOT:/cvs
Module name:ports
Changes by: na...@cvs.openbsd.org   2016/04/21 14:28:26

Modified files:
devel/quirks   : Makefile 
devel/quirks/files: Quirks.pm 
net: Makefile 
Removed files:
net/amsn   : Makefile distinfo 
net/amsn/patches: patch-Makefile_in patch-amsn-remote 
  patch-amsn-remote-CLI patch-configure 
net/amsn/pkg   : DESCR PLIST 
net/pebrot : Makefile distinfo 
net/pebrot/pkg : DESCR PLIST 

Log message:
remove clients for discontinued Microsoft Messenger service
noticed by Frederic Cambus

Re: Disable SSE3 in graphics/opencv

[Stuart Henderson]

On 2016/04/21 20:58, Rafael Sadowski wrote:
> On Wed Apr 20, 2016 at 05:30:26AM -0400, Michael McConville wrote:
> > It's on by default, but can be disabled with the below cmake argument.
> > 
> 
> Are test results with and without SSE3 same or better? If yes okay for
> me.

I think it would have to be "better" otherwise what's the point in
disabling this in opencv (which does runtime detection)?

(And if it is better that way, then that sounds like a bug).

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Christian Weisgerber]

On 2016-04-21, Raul Miller  wrote:

[INSTALL_STRIP=""]
>> A few ports already do this.
>
>  Do those ports crash with INSTALL_STRIP=-s ?

astro/xworld attaches its data as an ELF segment.  I don't know if
it will crash, but it certainly won't work if you strip it.

emulators/gxemul, I have no idea.

emulators/qemu uses it in a debug flavor.

-- 
Christian "naddy" Weisgerber  na...@mips.inka.de

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Stuart Henderson]

On 2016/04/21 19:35, Christian Weisgerber wrote:
> Dmitrij D. Czarkoff:
> 
> > > I think you should just set
> > > INSTALL_STRIP =
> > > in the ports (or modules) that require it and be done with it.
> > > No point in introducing yet another variable.
> > 
> > INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
> > /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
> > INSTALL_STRIP= from port.
> 
> In practice, INSTALL_STRIP has two possible values: "-s" and "".
> "-s" is already the default, so the only user setting that makes
> any sense is "".  Having some ports already set it to "" is just
> fine.
> 
> A few ports already do this.
> 
> -- 
> Christian "naddy" Weisgerber  na...@mips.inka.de
> 

Makes sense to me. go-bootstrap needs it too.

Re: NEW: devel/luacheck

[Florian Stinglmayr]

On Thu, Apr 21, 2016 at 07:43:48PM +1000, Jonathan Gray wrote:
> Luacheck supports checking Lua files using syntax of Lua 5.1, Lua 5.2,
> Lua 5.3 and LuaJIT 2.0. Luacheck itself is written in Lua and runs on
> all of mentioned Lua versions.

ok from me.

Works fine for me and my projects, and is very useful.

CVS: cvs.openbsd.org: ports

[Edd Barrett]

CVSROOT:/cvs
Module name:ports
Changes by: e...@cvs.openbsd.org2016/04/21 13:04:00

Modified files:
net: Makefile 

Log message:
Link net/syncthing.

Re: tdb-1.3.8, libtalloc-2.1.6, samba-4.4.2

[Jeremie Courreges-Anglas]

Ian McWilliam  writes:

> Looks good so far.
>
> make lib-depends-check
>
> Create /usr/packages/amd64/all/samba-util-4.4.2.tgz
> LIB_DEPENDS STEM->=0.10.38:devel/gettext not needed for net/samba,-util ?
> LIB_DEPENDS converters/libiconv not needed for net/samba,-util ?

FWIW I'd prefer not to touch this for now.  I consider those warnings
are harmless, and I don't know yet what to do about the gettext module.

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

CVS: cvs.openbsd.org: ports

[Edd Barrett]

CVSROOT:/cvs
Module name:ports
Changes by: e...@cvs.openbsd.org2016/04/21 13:01:31

Log message:
Import net/syncthing. A file synchroniser.

Based on work by czarkoff@

Input from czarkoff@, awolk@ and sthen@.

OK awolk@ (and from czarkoff@ for a slightly earlier version).

Thanks!

Status:

Vendor Tag: edd
Release Tags:   edd_20160421

N ports/net/syncthing/Makefile
N ports/net/syncthing/distinfo
N ports/net/syncthing/pkg/DESCR
N ports/net/syncthing/pkg/PLIST
N ports/net/syncthing/pkg/README
N ports/net/syncthing/pkg/syncthing.rc

No conflicts created by this import

Re: Disable SSE3 in graphics/opencv

[Rafael Sadowski]

On Wed Apr 20, 2016 at 05:30:26AM -0400, Michael McConville wrote:
> It's on by default, but can be disabled with the below cmake argument.
> 

Are test results with and without SSE3 same or better? If yes okay for
me.

> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/graphics/opencv/Makefile,v
> retrieving revision 1.37
> diff -u -p -u -r1.37 Makefile
> --- Makefile  15 Apr 2016 08:53:49 -  1.37
> +++ Makefile  20 Apr 2016 04:59:34 -
> @@ -13,6 +13,7 @@ DISTNAME =  opencv-$V
>  PKGNAME-main =   ${PKGNAME}
>  PKGNAME-docs =   opencv-docs-$V
>  PKGNAME-java =   opencv-java-$V
> +REVISION =   0
>  REVISION-docs =  0
>  
>  HOMEPAGE =   http://www.opencv.org/
> @@ -122,6 +123,7 @@ CONFIGURE_ARGS =  -DBUILD_opencv_nonfree:
>   -DWITH_PNG:Bool=On \
>   -DWITH_TIFF:Bool=On \
>   -DWITH_V4L:Bool=Off \
> + -DENABLE_SSE3:Bool=Off \
>   -DPYTHON_EXECUTABLE:FILEPATH=${MODPY_BIN}
>  
>  .if ${BUILD_PACKAGES:M-java}

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Raul Miller]

On Thu, Apr 21, 2016 at 1:35 PM, Christian Weisgerber
 wrote:
> A few ports already do this.

 Do those ports crash with INSTALL_STRIP=-s ?

Thanks,

-- 
Raul

Re: sysutils/sysclean 1.5

[Joerg Jung]

On Thu, Apr 21, 2016 at 01:37:37PM +0200, Sebastien Marie wrote:
> Hi,
> 
> I just hope to not spam the list...
> 
> Version 1.5:
> - Correct perl error which occurs on empty line (or only comment) in
>   sysclean.ignore
> 
> - Reintroduces shell pattern in sysclean.ignore
> 
> - Manages files in /usr/local/lib/X11/app-defaults/ (is a symlink to
>   /etc/X11/app-defaults/)
> 
> Comments ?

ok jung@

Do you plan to remove /usr/local from defaults ignore list?

> -- 
> Sebastien Marie
> 
> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/sysutils/sysclean/Makefile,v
> retrieving revision 1.5
> diff -u -p -r1.5 Makefile
> --- Makefile  19 Apr 2016 08:32:37 -  1.5
> +++ Makefile  21 Apr 2016 11:35:30 -
> @@ -2,7 +2,7 @@
>  
>  COMMENT =help removing obsolete files between OpenBSD upgrades
>  
> -V =  1.4
> +V =  1.5
>  DISTNAME =   sysclean-${V}
>  
>  CATEGORIES = sysutils
> Index: distinfo
> ===
> RCS file: /cvs/ports/sysutils/sysclean/distinfo,v
> retrieving revision 1.5
> diff -u -p -r1.5 distinfo
> --- distinfo  19 Apr 2016 08:32:37 -  1.5
> +++ distinfo  21 Apr 2016 11:35:30 -
> @@ -1,2 +1,2 @@
> -SHA256 (sysclean-1.4.tar.gz) = B4JBu/3dlTc7MxeUl/yjVeZu9OL4Uf/AEDN/TtaZGHQ=
> -SIZE (sysclean-1.4.tar.gz) = 4504
> +SHA256 (sysclean-1.5.tar.gz) = vEX1rfijTDmXmzzYAJnyaeppPlIBuI9Avvcs52BSMuM=
> +SIZE (sysclean-1.5.tar.gz) = 4732
> 

x11/sxhkd

[ix]

Hi,

A few weeks ago I sent in a port for sxhkd 
(https://github.com/baskerville/sxhkd).

sxhkd is a hotkey daemon for x11 that supports a simple config syntax.

I've fixed the port so portcheck accepts it, and I haven't gotten any negative 
feedback,
so can someone commit the port?

Thanks.

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Christian Weisgerber]

Dmitrij D. Czarkoff:

> > I think you should just set
> > INSTALL_STRIP =
> > in the ports (or modules) that require it and be done with it.
> > No point in introducing yet another variable.
> 
> INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
> /etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
> INSTALL_STRIP= from port.

In practice, INSTALL_STRIP has two possible values: "-s" and "".
"-s" is already the default, so the only user setting that makes
any sense is "".  Having some ports already set it to "" is just
fine.

A few ports already do this.

-- 
Christian "naddy" Weisgerber  na...@mips.inka.de

Re: NEW: lua-lyaml

[Florian Stinglmayr]

On Thu, Apr 21, 2016 at 08:47:46PM +1000, Jonathan Gray wrote:
> This could now drop the SHARED_ONLY part.
>

Yes.

> Why the need to rerun autoconf?  Could configure be used as is?  Currently
> the build fails here with:
>
> ===>  Configuring for lua-lyaml-6.0
> Using /usr/ports/pobj/lua-lyaml-6.0/config.site (generated)
> /usr/local/bin/autoreconf[35]: /usr/local/bin/autoreconf-2.13: not found
>

You are correct. The previous release didn't have a configure, and was
just a .tar.gz from git master. The appropriate lines have been removed,
and it still builds and works fine.

Regards,
Florian



lua-lyaml.tar.gz
Description: application/tar-gz

devel/scons fixes

[Ingo Feinerer]

Hi,

I was working on a port which uses devel/scons. Following changes to
scons.port.mk were necessary which appear to fix general problems:

- use the proper path separator ":" as otherwise all but the first path
  are ignored
- ensure that MODSCONS_ENV is seen by MODSCONS_BIN

Comments?

Best regards,
Ingo

Index: scons.port.mk
===
RCS file: /cvs/ports/devel/scons/scons.port.mk,v
retrieving revision 1.6
diff -u -p -r1.6 scons.port.mk
--- scons.port.mk   2 Jul 2013 08:36:16 -   1.6
+++ scons.port.mk   21 Apr 2016 16:36:28 -
@@ -9,8 +9,8 @@ MODSCONS_ENV?=  CC="${CC}" \
CCFLAGS="${CFLAGS}" \
CXXFLAGS="${CXXFLAGS}" \
LINKFLAGS="${LDFLAGS}" \
-   CPPPATH="${LOCALBASE}/include ${X11BASE}/include" \
-   LIBPATH="${LOCALBASE}/lib ${X11BASE}/lib" \
+   CPPPATH="${LOCALBASE}/include:${X11BASE}/include" \
+   LIBPATH="${LOCALBASE}/lib:${X11BASE}/lib" \
PREFIX="${PREFIX}" \
debug=0
 
@@ -19,12 +19,12 @@ ALL_TARGET?=
 NO_CCACHE?=Yes
 
 MODSCONS_BUILD_TARGET = \
-   ${SETENV} ${MAKE_ENV} ${MODSCONS_BIN} -C ${WRKSRC} \
-   ${MODSCONS_ENV} ${MODSCONS_FLAGS} ${ALL_TARGET}
+   ${SETENV} ${MAKE_ENV} ${MODSCONS_ENV} ${MODSCONS_BIN} -C ${WRKSRC} \
+   ${MODSCONS_FLAGS} ${ALL_TARGET}
 
 MODSCONS_INSTALL_TARGET = \
-   ${SETENV} ${MAKE_ENV} ${MODSCONS_BIN} -C ${WRKSRC} \
-   ${MODSCONS_ENV} ${MODSCONS_FLAGS} ${INSTALL_TARGET} \
+   ${SETENV} ${MAKE_ENV} ${MODSCONS_ENV} ${MODSCONS_BIN} -C ${WRKSRC} \
+   ${MODSCONS_FLAGS} ${INSTALL_TARGET} \
DESTDIR=${WRKINST}
 
 .if !target(do-build)

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Christian Weisgerber said:
> Dmitrij D. Czarkoff:
> 
> > --- infrastructure/mk/bsd.port.mk   20 Mar 2016 20:07:20 -  1.1310
> > +++ infrastructure/mk/bsd.port.mk   21 Apr 2016 13:10:29 -
> > @@ -729,6 +729,12 @@ UNZIP ?= unzip
> >  BZIP2 ?= bzip2
> >  
> >  
> > +# disable stripping if requested
> > +INSTALL_NO_STRIP ?=
> > +.if ${INSTALL_NO_STRIP:L} == "yes"
> > +INSTALL_STRIP =
> > +.endif
> > +
> 
> I think you should just set
> 
> INSTALL_STRIP =
> 
> in the ports (or modules) that require it and be done with it.
> No point in introducing yet another variable.

INSTALL_STRIP is mentioned in mk.conf(5), so it is a user setting.
/etc/mk.conf may contain INSTALL_STRIP=-s, and that would override
INSTALL_STRIP= from port.

-- 
Dmitrij D. Czarkoff

CVS: cvs.openbsd.org: ports

[Vadim Zhukov]

CVSROOT:/cvs
Module name:ports
Changes by: z...@cvs.openbsd.org2016/04/21 09:51:33

Modified files:
infrastructure/bin: portcheck 

Log message:
Print warning when SHARED_ONLY being set in a port.

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Christian Weisgerber]

Dmitrij D. Czarkoff:

> --- infrastructure/mk/bsd.port.mk 20 Mar 2016 20:07:20 -  1.1310
> +++ infrastructure/mk/bsd.port.mk 21 Apr 2016 13:10:29 -
> @@ -729,6 +729,12 @@ UNZIP ?= unzip
>  BZIP2 ?= bzip2
>  
>  
> +# disable stripping if requested
> +INSTALL_NO_STRIP ?=
> +.if ${INSTALL_NO_STRIP:L} == "yes"
> +INSTALL_STRIP =
> +.endif
> +

I think you should just set

INSTALL_STRIP =

in the ports (or modules) that require it and be done with it.
No point in introducing yet another variable.

-- 
Christian "naddy" Weisgerber  na...@mips.inka.de

CVS: cvs.openbsd.org: ports

[Christian Weisgerber]

CVSROOT:/cvs
Module name:ports
Changes by: na...@cvs.openbsd.org   2016/04/21 09:43:23

Modified files:
x11/type1inst  : Makefile 

Log message:
use INSTALL_SCRIPT

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Adam Wolk said:
> On Thu, 21 Apr 2016 15:12:24 +0200
> "Dmitrij D. Czarkoff"  wrote:
> 
>> Dmitrij D. Czarkoff said:
>>> Edd Barrett said:  
 On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:  
> +# Use INSTALL_SCRIPT to prevent stripping go binaries  
 
 Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to
 avoid confusion?  
>>> 
>>> I'd rather go with more generic change.  Patch below adds
>>> INSTALL_NO_STRIP macro to bsd.port.mk.  Setting this macro to "Yes"
>>> makes INSTALL_STRIP empty.  go.port.mk sets INSTALL_NO_STRIP to
>>> "Yes".
>>> 
>>> Diff against bsd.port.mk.5 in src also included.  
>> 
>> Forgot to define INSTALL_NO_STRIP.
>> 
> 
> OK awolk@ on condition that INSTALL_NO_STRIP = Yes is added to
> lang/go/Makefile. The go port itself doesn't use it's module but
> installs go binaries and has the same warnings. Looks ok with your diff
> plus  the additional INSTALL_NO_STRIP.
> 
> Reverting logstash/forwarder change from yesterday will also be
> required. I can do that if you commit the go port changes.

I could do it all in one go.  But I'd rather have more OKs for diff that
changes bsd.port.mk.

(Diff below is updated with lang/go/Makefile chunk.)

-- 
Dmitrij D. Czarkoff

Index: infrastructure/mk/bsd.port.mk
===
RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v
retrieving revision 1.1310
diff -u -p -r1.1310 bsd.port.mk
--- infrastructure/mk/bsd.port.mk   20 Mar 2016 20:07:20 -  1.1310
+++ infrastructure/mk/bsd.port.mk   21 Apr 2016 13:10:29 -
@@ -729,6 +729,12 @@ UNZIP ?= unzip
 BZIP2 ?= bzip2
 
 
+# disable stripping if requested
+INSTALL_NO_STRIP ?=
+.if ${INSTALL_NO_STRIP:L} == "yes"
+INSTALL_STRIP =
+.endif
+
 # copy selected info from bsd.own.mk
 MAKE_ENV += COMPILER_VERSION=${COMPILER_VERSION} \
PICFLAG="${PICFLAG}" ASPICFLAG=${ASPICFLAG} \
Index: Makefile
===
RCS file: /cvs/ports/lang/go/Makefile,v
retrieving revision 1.29
diff -u -p -r1.29 Makefile
--- Makefile14 Apr 2016 17:43:26 -  1.29
+++ Makefile21 Apr 2016 14:27:24 -
@@ -30,6 +30,8 @@ SUBST_VARS =  GOCFG
 WRKDIST =  ${WRKDIR}/go
 WRKSRC =   ${WRKDIST}/src
 
+INSTALL_NO_STRIP = Yes
+
 GOOS = openbsd
 GOARCH =   unknown
 GOROOT =   ${PREFIX}/go
Index: lang/go/go.port.mk
===
RCS file: /cvs/ports/lang/go/go.port.mk,v
retrieving revision 1.6
diff -u -p -r1.6 go.port.mk
--- lang/go/go.port.mk  28 Feb 2016 13:24:16 -  1.6
+++ lang/go/go.port.mk  21 Apr 2016 13:03:21 -
@@ -2,6 +2,8 @@
 
 ONLY_FOR_ARCHS ?=  ${GO_ARCHS}
 
+INSTALL_NO_STRIP = Yes
+
 MODGO_BUILDDEP ?=  Yes
 
 MODGO_RUN_DEPENDS =lang/go
Index: share/man/man5/bsd.port.mk.5
===
RCS file: /cvs/src/share/man/man5/bsd.port.mk.5,v
retrieving revision 1.434
diff -u -p -r1.434 bsd.port.mk.5
--- share/man/man5/bsd.port.mk.522 Mar 2016 16:02:02 -  1.434
+++ share/man/man5/bsd.port.mk.521 Apr 2016 13:02:45 -
@@ -1743,6 +1743,11 @@ User settings.
 If set to
 .Sq Yes ,
 do not print anything when ignoring a port.
+.It Ev INSTALL_NO_STRIP
+If set to
+.Sq Yes ,
+.Ev INSTALL_PROGRAM
+macro does not strip binaries.
 .It Ev INSTALL_{PROGRAM,SCRIPT,DATA,MAN}[_DIR]
 Macros to use to install a program, a script, data, or a man page (or the
 corresponding directory), respectively.

Re: update sysutils/sshfs-fuse, move to GitHub

[Dmitrij D. Czarkoff]

Stuart Henderson said:
> On 2016/04/21 10:06, Dmitrij D. Czarkoff wrote:
>> Michael McConville said:
>>> Dmitrij D. Czarkoff wrote:
 Michael McConville said:
> +GH_ACCOUNT = libfuse
> +GH_PROJECT = sshfs
 
 You don't need these any more.
>>> 
>>> Don't we need them for HOMEPAGE? It uses the explicitly set MASTER_SITES
>>> in its current form, too, so we don't need to worry about GH_*
>>> overwriting that.
>> 
>> Why don't you set HOMEPAGE explicitly then?
>> 
>> Also, CONFIGURE_STYLE=gnu would be enough here.
> 
> And it changes the package name from sshfs-fuse to just sshfs.

This diff sets PKGNAME and uses explicit HOMEPAGE.  OK?

P.S.:  sshfs does not recognize "-F" option, which is a major annoyance.

-- 
Dmitrij D. Czarkoff

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Adam Wolk]

On Thu, 21 Apr 2016 15:12:24 +0200
"Dmitrij D. Czarkoff"  wrote:

> Dmitrij D. Czarkoff said:
> > Edd Barrett said:  
> > > On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:  
> > > > +# Use INSTALL_SCRIPT to prevent stripping go binaries  
> > > 
> > > Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to
> > > avoid confusion?  
> > 
> > I'd rather go with more generic change.  Patch below adds
> > INSTALL_NO_STRIP macro to bsd.port.mk.  Setting this macro to "Yes"
> > makes INSTALL_STRIP empty.  go.port.mk sets INSTALL_NO_STRIP to
> > "Yes".
> > 
> > Diff against bsd.port.mk.5 in src also included.  
> 
> Forgot to define INSTALL_NO_STRIP.
> 

OK awolk@ on condition that INSTALL_NO_STRIP = Yes is added to
lang/go/Makefile. The go port itself doesn't use it's module but
installs go binaries and has the same warnings. Looks ok with your diff
plus  the additional INSTALL_NO_STRIP.

Reverting logstash/forwarder change from yesterday will also be
required. I can do that if you commit the go port changes.

Patch to revert logstash:
Index: Makefile
===
RCS file: /cvs/ports/sysutils/logstash/forwarder/Makefile,v
retrieving revision 1.8
diff -u -p -r1.8 Makefile
--- Makefile21 Apr 2016 00:07:08 -  1.8
+++ Makefile21 Apr 2016 14:11:54 -
@@ -5,7 +5,7 @@ COMMENT=collect logs locally in prepara
 GH_ACCOUNT=elastic
 GH_PROJECT=logstash-forwarder
 GH_TAGNAME=v0.4.0
-REVISION=  2
+REVISION=  3
 
 MODULES=   lang/go
 
@@ -16,9 +16,8 @@ WANTLIB +=c pthread
 pre-configure:
${SUBST_CMD} ${WRKSRC}/logstash-forwarder.conf.example
 
-# Use INSTALL_SCRIPT to prevent stripping go binaries
 do-install:
-   ${INSTALL_SCRIPT} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
+   ${INSTALL_PROGRAM} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
${PREFIX}/sbin/
${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/logstash-forwarder/
${INSTALL_DATA} ${WRKSRC}/README.md \

CVS: cvs.openbsd.org: ports

[Robert Nagy]

CVSROOT:/cvs
Module name:ports
Changes by: rob...@cvs.openbsd.org  2016/04/21 07:52:06

Modified files:
www/chromium   : Makefile distinfo 

Log message:
update to 50.0.2661.86

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Dmitrij D. Czarkoff said:
> Edd Barrett said:
> > On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:
> > > +# Use INSTALL_SCRIPT to prevent stripping go binaries
> > 
> > Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to avoid
> > confusion?
> 
> I'd rather go with more generic change.  Patch below adds
> INSTALL_NO_STRIP macro to bsd.port.mk.  Setting this macro to "Yes"
> makes INSTALL_STRIP empty.  go.port.mk sets INSTALL_NO_STRIP to "Yes".
> 
> Diff against bsd.port.mk.5 in src also included.

Forgot to define INSTALL_NO_STRIP.

-- 
Dmitrij D. Czarkoff

Index: infrastructure/mk/bsd.port.mk
===
RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v
retrieving revision 1.1310
diff -u -p -r1.1310 bsd.port.mk
--- infrastructure/mk/bsd.port.mk   20 Mar 2016 20:07:20 -  1.1310
+++ infrastructure/mk/bsd.port.mk   21 Apr 2016 13:10:29 -
@@ -729,6 +729,12 @@ UNZIP ?= unzip
 BZIP2 ?= bzip2
 
 
+# disable stripping if requested
+INSTALL_NO_STRIP ?=
+.if ${INSTALL_NO_STRIP:L} == "yes"
+INSTALL_STRIP =
+.endif
+
 # copy selected info from bsd.own.mk
 MAKE_ENV += COMPILER_VERSION=${COMPILER_VERSION} \
PICFLAG="${PICFLAG}" ASPICFLAG=${ASPICFLAG} \
Index: lang/go/go.port.mk
===
RCS file: /cvs/ports/lang/go/go.port.mk,v
retrieving revision 1.6
diff -u -p -r1.6 go.port.mk
--- lang/go/go.port.mk  28 Feb 2016 13:24:16 -  1.6
+++ lang/go/go.port.mk  21 Apr 2016 13:03:21 -
@@ -2,6 +2,8 @@
 
 ONLY_FOR_ARCHS ?=  ${GO_ARCHS}
 
+INSTALL_NO_STRIP = Yes
+
 MODGO_BUILDDEP ?=  Yes
 
 MODGO_RUN_DEPENDS =lang/go
Index: share/man/man5/bsd.port.mk.5
===
RCS file: /cvs/src/share/man/man5/bsd.port.mk.5,v
retrieving revision 1.434
diff -u -p -r1.434 bsd.port.mk.5
--- share/man/man5/bsd.port.mk.522 Mar 2016 16:02:02 -  1.434
+++ share/man/man5/bsd.port.mk.521 Apr 2016 13:02:45 -
@@ -1743,6 +1743,11 @@ User settings.
 If set to
 .Sq Yes ,
 do not print anything when ignoring a port.
+.It Ev INSTALL_NO_STRIP
+If set to
+.Sq Yes ,
+.Ev INSTALL_PROGRAM
+macro does not strip binaries.
 .It Ev INSTALL_{PROGRAM,SCRIPT,DATA,MAN}[_DIR]
 Macros to use to install a program, a script, data, or a man page (or the
 corresponding directory), respectively.

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Dmitrij D. Czarkoff]

Edd Barrett said:
> On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:
> > +# Use INSTALL_SCRIPT to prevent stripping go binaries
> 
> Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to avoid
> confusion?

I'd rather go with more generic change.  Patch below adds
INSTALL_NO_STRIP macro to bsd.port.mk.  Setting this macro to "Yes"
makes INSTALL_STRIP empty.  go.port.mk sets INSTALL_NO_STRIP to "Yes".

Diff against bsd.port.mk.5 in src also included.

-- 
Dmitrij D. Czarkoff

Index: infrastructure/mk/bsd.port.mk
===
RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v
retrieving revision 1.1310
diff -u -p -r1.1310 bsd.port.mk
--- infrastructure/mk/bsd.port.mk   20 Mar 2016 20:07:20 -  1.1310
+++ infrastructure/mk/bsd.port.mk   21 Apr 2016 12:57:11 -
@@ -729,6 +729,11 @@ UNZIP ?= unzip
 BZIP2 ?= bzip2
 
 
+# disable stripping if requested
+.if ${INSTALL_NO_STRIP:L} == "yes"
+INSTALL_STRIP =
+.endif
+
 # copy selected info from bsd.own.mk
 MAKE_ENV += COMPILER_VERSION=${COMPILER_VERSION} \
PICFLAG="${PICFLAG}" ASPICFLAG=${ASPICFLAG} \
Index: lang/go/go.port.mk
===
RCS file: /cvs/ports/lang/go/go.port.mk,v
retrieving revision 1.6
diff -u -p -r1.6 go.port.mk
--- lang/go/go.port.mk  28 Feb 2016 13:24:16 -  1.6
+++ lang/go/go.port.mk  21 Apr 2016 13:03:21 -
@@ -2,6 +2,8 @@
 
 ONLY_FOR_ARCHS ?=  ${GO_ARCHS}
 
+INSTALL_NO_STRIP = Yes
+
 MODGO_BUILDDEP ?=  Yes
 
 MODGO_RUN_DEPENDS =lang/go
Index: share/man/man5/bsd.port.mk.5
===
RCS file: /cvs/src/share/man/man5/bsd.port.mk.5,v
retrieving revision 1.434
diff -u -p -r1.434 bsd.port.mk.5
--- share/man/man5/bsd.port.mk.522 Mar 2016 16:02:02 -  1.434
+++ share/man/man5/bsd.port.mk.521 Apr 2016 13:02:45 -
@@ -1743,6 +1743,11 @@ User settings.
 If set to
 .Sq Yes ,
 do not print anything when ignoring a port.
+.It Ev INSTALL_NO_STRIP
+If set to
+.Sq Yes ,
+.Ev INSTALL_PROGRAM
+macro does not strip binaries.
 .It Ev INSTALL_{PROGRAM,SCRIPT,DATA,MAN}[_DIR]
 Macros to use to install a program, a script, data, or a man page (or the
 corresponding directory), respectively.

CVS: cvs.openbsd.org: ports

[James Turner]

CVSROOT:/cvs
Module name:ports
Changes by: jtur...@cvs.openbsd.org 2016/04/21 07:08:01

Modified files:
www/kcaldav: Makefile 
www/kcaldav/pkg: README 

Log message:
Remove stray ' in README, bump for safety. Spotted by sthen@

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Adam Wolk]

On Thu, 21 Apr 2016 12:35:41 +0100
Edd Barrett  wrote:

> On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:
> > +# Use INSTALL_SCRIPT to prevent stripping go binaries  
> 
> Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to avoid
> confusion?
> 
> I note that I saw such weird errors when building lang/go itself, so
> perhaps we should be applying similar changes there?
> 

I sent a patch for lang/go yesterday for verification to jsing@ & sthen@

Regards,
Adam

CVS: cvs.openbsd.org: ports

[David Coppa]

CVSROOT:/cvs
Module name:ports
Changes by: dco...@cvs.openbsd.org  2016/04/21 05:42:29

Modified files:
x11/mlterm : Makefile distinfo 
x11/mlterm/patches: patch-configure_in patch-mkf_lib_Makefile_in 
patch-mlterm_Makefile_in 
patch-xwindow_x_screen_c 
x11/mlterm/pkg : PLIST 

Log message:
Update to mlterm-3.7.0

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Edd Barrett]

On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:
> +# Use INSTALL_SCRIPT to prevent stripping go binaries

Is it worth adding a INSTALL_GO_PROGRAM in the lang/go MODULE to avoid
confusion?

I note that I saw such weird errors when building lang/go itself, so
perhaps we should be applying similar changes there?

-- 
Best Regards
Edd Barrett

http://www.theunixzoo.co.uk

sysutils/sysclean 1.5

[Sebastien Marie]

Hi,

I just hope to not spam the list...

Version 1.5:
- Correct perl error which occurs on empty line (or only comment) in
  sysclean.ignore

- Reintroduces shell pattern in sysclean.ignore

- Manages files in /usr/local/lib/X11/app-defaults/ (is a symlink to
  /etc/X11/app-defaults/)

Comments ?
-- 
Sebastien Marie


Index: Makefile
===
RCS file: /cvs/ports/sysutils/sysclean/Makefile,v
retrieving revision 1.5
diff -u -p -r1.5 Makefile
--- Makefile19 Apr 2016 08:32:37 -  1.5
+++ Makefile21 Apr 2016 11:35:30 -
@@ -2,7 +2,7 @@
 
 COMMENT =  help removing obsolete files between OpenBSD upgrades
 
-V =1.4
+V =1.5
 DISTNAME = sysclean-${V}
 
 CATEGORIES =   sysutils
Index: distinfo
===
RCS file: /cvs/ports/sysutils/sysclean/distinfo,v
retrieving revision 1.5
diff -u -p -r1.5 distinfo
--- distinfo19 Apr 2016 08:32:37 -  1.5
+++ distinfo21 Apr 2016 11:35:30 -
@@ -1,2 +1,2 @@
-SHA256 (sysclean-1.4.tar.gz) = B4JBu/3dlTc7MxeUl/yjVeZu9OL4Uf/AEDN/TtaZGHQ=
-SIZE (sysclean-1.4.tar.gz) = 4504
+SHA256 (sysclean-1.5.tar.gz) = vEX1rfijTDmXmzzYAJnyaeppPlIBuI9Avvcs52BSMuM=
+SIZE (sysclean-1.5.tar.gz) = 4732

Re: NEW: lua-lyaml

[Jonathan Gray]

On Tue, Jan 19, 2016 at 06:03:36PM +0100, Florian Stinglmayr wrote:
> Since I haven't received any suggestions for fixes or improvements I
> assume it is ready to be imported, no? ;-)
> 
> Regards,
> Florian
> 

This could now drop the SHARED_ONLY part.

Why the need to rerun autoconf?  Could configure be used as is?  Currently
the build fails here with:

===>  Configuring for lua-lyaml-6.0
Using /usr/ports/pobj/lua-lyaml-6.0/config.site (generated)
/usr/local/bin/autoreconf[35]: /usr/local/bin/autoreconf-2.13: not found

UPDATE: devel/luaposix 33.4.0

[Jonathan Gray]

Update from an old archived version of luaposix to the latest.

This requires the lua-bit32 port I just posted.

Index: Makefile
===
RCS file: /cvs/ports/devel/luaposix/Makefile,v
retrieving revision 1.20
diff -u -p -r1.20 Makefile
--- Makefile18 Mar 2016 20:49:28 -  1.20
+++ Makefile21 Apr 2016 10:21:35 -
@@ -1,29 +1,28 @@
 # $OpenBSD: Makefile,v 1.20 2016/03/18 20:49:28 naddy Exp $
 
 COMMENT=   posix library for the lua language
-DISTNAME=  luaposix-5.1.4
+V= 33.4.0
+DISTNAME=  luaposix-${V}
 EPOCH= 0
-REVISION=  1
 CATEGORIES=devel
-MASTER_SITES=  http://luaforge.net/frs/download.php/3572/
 
-HOMEPAGE=  http://luaforge.net/projects/luaposix/
+GH_ACCOUNT=luaposix
+GH_PROJECT=luaposix
+GH_TAGNAME=release-v${V}
 
 # Public domain
 PERMIT_PACKAGE_CDROM=  Yes
+NO_TEST=   Yes
 
-MODULES=   lang/lua
-TEST_DEPENDS=${BUILD_PKGPATH}
+MODULES=   lang/lua
 
-do-build:
-   ${CC} ${CFLAGS} -fPIC -I${MODLUA_INCL_DIR} -shared \
-   ${WRKDIST}/lposix.c -o ${WRKDIST}/posix.so
-
-do-install:
-   ${INSTALL_DATA_DIR} ${MODLUA_LIBDIR}
-   ${INSTALL_DATA} ${WRKSRC}/posix.so ${MODLUA_LIBDIR}
+.if "${MODLUA_VERSION}" == "5.1"
+RUN_DEPENDS=   devel/lua-bit32
+.endif
 
-do-test:
-   cd ${WRKSRC} && ${MODLUA_BIN} *.lua
+CONFIGURE_STYLE=   gnu
+
+CONFIGURE_ENV+=LUA=${MODLUA_BIN} \
+   LUA_INCLUDE=-I${MODLUA_INCL_DIR}
 
 .include 
Index: distinfo
===
RCS file: /cvs/ports/devel/luaposix/distinfo,v
retrieving revision 1.7
diff -u -p -r1.7 distinfo
--- distinfo18 Jan 2015 03:13:18 -  1.7
+++ distinfo21 Apr 2016 08:11:35 -
@@ -1,2 +1,2 @@
-SHA256 (luaposix-5.1.4.tar.gz) = D3j+nIisdZWSKaWdqw2IZbP1xXMHBag+dMJG9HUigWc=
-SIZE (luaposix-5.1.4.tar.gz) = 13390
+SHA256 (luaposix-33.4.0.tar.gz) = 5mJi9bf+HDLGXxel71/7McTRh3AZtIcKXTc+KrZSaiE=
+SIZE (luaposix-33.4.0.tar.gz) = 643523
Index: patches/patch-test_lua
===
RCS file: patches/patch-test_lua
diff -N patches/patch-test_lua
--- patches/patch-test_lua  6 Sep 2009 15:39:21 -   1.1
+++ /dev/null   1 Jan 1970 00:00:00 -
@@ -1,12 +0,0 @@
-$OpenBSD: patch-test_lua,v 1.1 2009/09/06 15:39:21 jolan Exp $
 test.lua.orig  Thu Jan 24 08:15:06 2008
-+++ test.lua   Sun Sep  6 10:36:17 2009
-@@ -83,7 +83,7 @@ if pid==0 then
-   ppid=ox.getpid"ppid"
-   io.write("in child process ",pid," from ",ppid,".\nnow executing 
date... ")
-   io.flush()
--  assert(ox.exec("date","+[%c]"))
-+  assert(ox.exec("/bin/date","+[%c]"))
-   print"should not get here"
- else
-   io.write("process ",ox.getpid"pid"," forked child process ",pid,". 
waiting...\n")
Index: pkg/PLIST
===
RCS file: /cvs/ports/devel/luaposix/pkg/PLIST,v
retrieving revision 1.3
diff -u -p -r1.3 PLIST
--- pkg/PLIST   11 May 2012 20:51:04 -  1.3
+++ pkg/PLIST   21 Apr 2016 09:02:57 -
@@ -1,2 +1,59 @@
 @comment $OpenBSD: PLIST,v 1.3 2012/05/11 20:51:04 jasper Exp $
+lib/lua/${MODLUA_VERSION}/posix.a
 lib/lua/${MODLUA_VERSION}/posix.so
+share/doc/luaposix/
+share/doc/luaposix/examples/
+share/doc/luaposix/examples/dir.lua.html
+share/doc/luaposix/examples/fork.lua.html
+share/doc/luaposix/examples/fork2.lua.html
+share/doc/luaposix/examples/getopt.lua.html
+share/doc/luaposix/examples/glob.lua.html
+share/doc/luaposix/examples/limit.lua.html
+share/doc/luaposix/examples/lock.lua.html
+share/doc/luaposix/examples/netlink-uevent.lua.html
+share/doc/luaposix/examples/ping.lua.html
+share/doc/luaposix/examples/poll.lua.html
+share/doc/luaposix/examples/signal.lua.html
+share/doc/luaposix/examples/socket.lua.html
+share/doc/luaposix/examples/termios.lua.html
+share/doc/luaposix/examples/tree.lua.html
+share/doc/luaposix/index.html
+share/doc/luaposix/ldoc.css
+share/doc/luaposix/modules/
+share/doc/luaposix/modules/posix.ctype.html
+share/doc/luaposix/modules/posix.dirent.html
+share/doc/luaposix/modules/posix.errno.html
+share/doc/luaposix/modules/posix.fcntl.html
+share/doc/luaposix/modules/posix.fnmatch.html
+share/doc/luaposix/modules/posix.getopt.html
+share/doc/luaposix/modules/posix.glob.html
+share/doc/luaposix/modules/posix.grp.html
+share/doc/luaposix/modules/posix.html
+share/doc/luaposix/modules/posix.libgen.html
+share/doc/luaposix/modules/posix.poll.html
+share/doc/luaposix/modules/posix.pwd.html
+share/doc/luaposix/modules/posix.sched.html
+share/doc/luaposix/modules/posix.signal.html
+share/doc/luaposix/modules/posix.stdio.html
+share/doc/luaposix/modules/posix.stdlib.html
+share/doc/luaposix/modules/posix.sys.msg.html
+share/doc/luaposix/modules/posix.sys.resource.html
+share/doc/luaposix/modules/posix.sys.socket.html

NEW: devel/lua-bit32

[Jonathan Gray]

This provides the lua 5.2 style bit32 bitop interface to lua 5.1.
Required to use newer versions of luaposix with lua 5.1.


lua-bit32.tgz
Description: application/tar-gz

CVS: cvs.openbsd.org: ports

[Alexandr Shadchin]

CVSROOT:/cvs
Module name:ports
Changes by: shadc...@cvs.openbsd.org2016/04/21 04:03:11

Modified files:
devel/py-babel : Makefile distinfo 
devel/py-babel/pkg: PLIST 

Log message:
Update to py-babel 2.3.3 and take maintainer.

ok danj@

NEW: devel/luacheck

[Jonathan Gray]

Luacheck is a static analyzer and a linter for Lua. Luacheck detects
various issues such as usage of undefined global variables, unused
variables and values, accessing uninitialized variables, unreachable
code and more. Most aspects of checking are configurable: there are
options for defining custom project-related globals, for selecting set
of standard globals (version of Lua standard library), for filtering
warnings by type and name of related variable, etc. The options can be
used on the command line, put into a config or directly into checked
files as Lua comments.

Luacheck supports checking Lua files using syntax of Lua 5.1, Lua 5.2,
Lua 5.3 and LuaJIT 2.0. Luacheck itself is written in Lua and runs on
all of mentioned Lua versions.


luacheck.tgz
Description: application/tar-gz

CVS: cvs.openbsd.org: ports

[Sebastien Marie]

CVSROOT:/cvs
Module name:ports
Changes by: sema...@cvs.openbsd.org 2016/04/21 02:56:07

Modified files:
sysutils/pstree: Makefile 
sysutils/pstree/patches: patch-pstree_c 

Log message:
pledge sysutils/pstree

initial patch from steve latif

ok sthen@

Re: update sysutils/sshfs-fuse, move to GitHub

[Stuart Henderson]

On 2016/04/21 10:06, Dmitrij D. Czarkoff wrote:
> Michael McConville said:
> > Dmitrij D. Czarkoff wrote:
> > > Michael McConville said:
> > > > +GH_ACCOUNT =   libfuse
> > > > +GH_PROJECT =   sshfs
> > > 
> > > You don't need these any more.
> > 
> > Don't we need them for HOMEPAGE? It uses the explicitly set MASTER_SITES
> > in its current form, too, so we don't need to worry about GH_*
> > overwriting that.
> 
> Why don't you set HOMEPAGE explicitly then?
> 
> Also, CONFIGURE_STYLE=gnu would be enough here.

And it changes the package name from sshfs-fuse to just sshfs.

CVS: cvs.openbsd.org: ports

[David Coppa]

CVSROOT:/cvs
Module name:ports
Changes by: dco...@cvs.openbsd.org  2016/04/21 02:13:22

Modified files:
security/ccid  : Makefile distinfo 
security/ccid/patches: patch-src_Makefile_in 

Log message:
Bugfixing update to ccid-1.4.23

Re: update sysutils/sshfs-fuse, move to GitHub

[Dmitrij D. Czarkoff]

Michael McConville said:
> Dmitrij D. Czarkoff wrote:
> > Michael McConville said:
> > > +GH_ACCOUNT = libfuse
> > > +GH_PROJECT = sshfs
> > 
> > You don't need these any more.
> 
> Don't we need them for HOMEPAGE? It uses the explicitly set MASTER_SITES
> in its current form, too, so we don't need to worry about GH_*
> overwriting that.

Why don't you set HOMEPAGE explicitly then?

Also, CONFIGURE_STYLE=gnu would be enough here.

-- 
Dmitrij D. Czarkoff

CVS: cvs.openbsd.org: ports

[David Coppa]

CVSROOT:/cvs
Module name:ports
Changes by: dco...@cvs.openbsd.org  2016/04/21 02:03:33

Modified files:
databases/hs-hedis: Makefile distinfo 

Log message:
Update to hedis-0.8.0

CVS: cvs.openbsd.org: ports

[David Coppa]

CVSROOT:/cvs
Module name:ports
Changes by: dco...@cvs.openbsd.org  2016/04/21 02:01:23

Modified files:
textproc   : Makefile 

Log message:
hook hs-scanner to the build

CVS: cvs.openbsd.org: ports

[David Coppa]

CVSROOT:/cvs
Module name:ports
Changes by: dco...@cvs.openbsd.org  2016/04/21 01:54:59

Log message:
Import hs-scanner, new dependency of an upcoming update to hedis-0.8.0.

$ cat pkg/DESCR
scanner is a fast non-backtracking incremental combinator parsing
library for bytestrings.
It is often convenient to use backtracking to parse some sophisticated
input. Unfortunately it kills performance, so usually you should
avoid backtracking.
Often (actually always, but it could be too hard sometimes) you can
implement your parser without any backtracking. It that case all
the bookkeeping usual parser combinators do becomes unnecessary.
The scanner library is designed for such cases. It is often 2 times
faster than attoparsec.

OK kili@

Status:

Vendor Tag: dcoppa
Release Tags:   dcoppa_20160421

N ports/textproc/hs-scanner/Makefile
N ports/textproc/hs-scanner/distinfo
N ports/textproc/hs-scanner/pkg/DESCR
N ports/textproc/hs-scanner/pkg/PLIST

No conflicts created by this import

Re: pledge() sysutils/pstree

[Sebastien Marie]

On Wed, Apr 20, 2016 at 10:59:44PM -0700, steve latif wrote:
> I made an attempt at adding pledge to pstree.
> steve
> 

Hi Steve,

Thanks for your contribution !

> $OpenBSD: patch-pstree_c,v 1.5 2015/05/22 13:45:30 schwarze Exp $
> --- pstree.c.orig   Wed May 13 05:24:47 2015
> +++ pstree.cWed Apr 20 22:21:07 2016
> @@ -872,6 +872,11 @@ int main(int argc, char **argv) {
>  #ifdef ZOMBIES_HAVE_PID_0
>FixZombies();
>  #endif
> +
> +  if (pledge("stdio getpw ioctl proc", NULL) == -1){
> +fprintf(stderr, "%s: pledge\n", Progname);
> +   exit(1);
> +  }

the "ioctl" promise should be replaced by "tty".

I haven't checked if the pledge(2) call could be placed
somewhere else in the code, but the promises looks sane.

below is a full port update.
-- 
Sebastien Marie


Index: Makefile
===
RCS file: /cvs/ports/sysutils/pstree/Makefile,v
retrieving revision 1.22
diff -u -p -r1.22 Makefile
--- Makefile22 May 2015 13:45:30 -  1.22
+++ Makefile21 Apr 2016 06:36:08 -
@@ -4,10 +4,12 @@ COMMENT = list processes as a tree
 DISTNAME = pstree-2.39
 CATEGORIES =   sysutils
 HOMEPAGE = http://www.thp.uni-duisburg.de/pstree/
+REVISION = 0
 
 # GPL
 PERMIT_PACKAGE_CDROM = Yes
 
+# uses pledge()
 WANTLIB =  c 
 
 MASTER_SITES = ftp://ftp.thp.uni-duisburg.de/pub/source/
Index: patches/patch-pstree_c
===
RCS file: /cvs/ports/sysutils/pstree/patches/patch-pstree_c,v
retrieving revision 1.5
diff -u -p -r1.5 patch-pstree_c
--- patches/patch-pstree_c  22 May 2015 13:45:30 -  1.5
+++ patches/patch-pstree_c  21 Apr 2016 06:36:08 -
@@ -1,6 +1,6 @@
 $OpenBSD: patch-pstree_c,v 1.5 2015/05/22 13:45:30 schwarze Exp $
 pstree.c.orig  Thu May 21 11:23:56 2015
-+++ pstree.c   Thu May 21 11:24:29 2015
+--- pstree.c.orig  Wed May 13 14:24:47 2015
 pstree.c   Thu Apr 21 08:23:00 2016
 @@ -77,7 +77,7 @@ extern getargs(struct ProcInfo *, int, char *, int);
   * (Net|Open|Free)BSD & Darwin merged by Ralf Meyer 
   */
@@ -10,3 +10,17 @@ $OpenBSD: patch-pstree_c,v 1.5 2015/05/2
  #  define PSFORMAT"%s %ld %ld %ld %[^\n]"
  #  define PSVARS  P[i].name, [i].pid, [i].ppid, [i].pgid, P[i].cmd
  #  define PSVARSN 5
+@@ -872,7 +872,12 @@ int main(int argc, char **argv) {
+ #ifdef ZOMBIES_HAVE_PID_0
+   FixZombies();
+ #endif
+-  
++
++  if (pledge("stdio getpw proc tty", NULL) == -1) {
++fprintf(stderr, "%s: pledge\n", Progname);
++exit(1);
++  }
++
+   if (NProc == 0) {
+ fprintf(stderr, "%s: No processes read.\n", Progname);
+ exit(1);

Re: sysutils/logstash/forwarder should not strip lang/go binaries

[Jasper Lievisse Adriaanse]

On Thu, Apr 21, 2016 at 01:38:54AM +0200, Adam Wolk wrote:
> Hi,
> 
> While testing a port for net/syncthing sthen@ noticed the following
> warnings:
> 
> 00:44 < sthen> hm, there are some funny messages from that port in
> 'make fake' 00:45 < sthen>
> BFD: /usr/obj/ports/syncthing-0.12.22/fake-amd64/usr/local/bin//stVnev2l:
> warning: allocated section `.gosymtab' not in segment 00:45 < sthen>
> BFD: /usr/obj/ports/syncthing-0.12.22/fake-amd64/usr/local/bin//stVnev2l:
> warning: allocated section `.gnu.version_r' not in segment
> 
> the cause for those errors is trying to strip a Go binary which doesn't
> output all expected segments. Upstream officially doesn't support
> stripping and it might lead to undefined behavior.
> 
> Debian Bug report logs - #717172
> Stripping golang binaries causes crashes:
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=717172
> 
> go get ... fails with SIGILL on armhf
> https://bugs.launchpad.net/ubuntu/+source/golang/+bug/1200255
> 
> docker: Stripping Binaries
> https://github.com/docker/docker/blob/2a95488f7843a773de2b541a47d9b971a635bfff/project/PACKAGERS.md#stripping-binaries
> 
> relevant blog post:
> https://blog.filippo.io/shrink-your-go-binaries-with-this-one-weird-trick/
> 
> I went through existing go ports to see how many of them actually strip
> the binaries.
> 
> https://gist.github.com/mulander/a0dd5d5b49ca81001628d690b70ba111
> 
> Apparently from the existing ports only sysutils/logstash/forwarder is
> affected.
> This issue was initially reported by Fritjof Bornebusch (thanks!) in
> July last year: http://marc.info/?l=openbsd-ports=143766040011892=2
> 
> I'm inlining a patch for sysutils/logstash/forwarder replacing the
> usage of INSTALL_PROGRAM with INSTALL_SCRIPT and a comment informing
> why this choice is made. Any OK's?
OK with me; go binaries can still be stripped but that's a compile/link option
and not something strip(1) should get involved with.

> Index: Makefile
> ===
> RCS file: /cvs/ports/sysutils/logstash/forwarder/Makefile,v
> retrieving revision 1.7
> diff -u -p -r1.7 Makefile
> --- Makefile  6 Jan 2016 20:02:19 -   1.7
> +++ Makefile  20 Apr 2016 23:37:17 -
> @@ -5,7 +5,7 @@ COMMENT=  collect logs locally in prepara
>  GH_ACCOUNT=  elastic
>  GH_PROJECT=  logstash-forwarder
>  GH_TAGNAME=  v0.4.0
> -REVISION=1
> +REVISION=2
>  
>  MODULES= lang/go
>  
> @@ -16,8 +16,9 @@ WANTLIB +=  c pthread
>  pre-configure:
>   ${SUBST_CMD} ${WRKSRC}/logstash-forwarder.conf.example
>  
> +# Use INSTALL_SCRIPT to prevent stripping go binaries
>  do-install:
> - ${INSTALL_PROGRAM} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
> + ${INSTALL_SCRIPT} ${MODGO_WORKSPACE}/bin/logstash-forwarder \
>   ${PREFIX}/sbin/
>   ${INSTALL_DATA_DIR} ${PREFIX}/share/{doc,examples}/logstash-forwarder/
>   ${INSTALL_DATA} ${WRKSRC}/README.md \
> 

-- 
jasper

pledge() sysutils/pstree

[steve latif]

I made an attempt at adding pledge to pstree.
steve


$OpenBSD: patch-pstree_c,v 1.5 2015/05/22 13:45:30 schwarze Exp $
--- pstree.c.orig   Wed May 13 05:24:47 2015
+++ pstree.cWed Apr 20 22:21:07 2016
@@ -77,7 +77,7 @@ extern getargs(struct ProcInfo *, int, char *, int);
  * (Net|Open|Free)BSD & Darwin merged by Ralf Meyer 
  */
 #  define HAS_PGID
-#  define PSCMD"ps -axwwo user,pid,ppid,pgid,command"
+#  define PSCMD"ps -kaxwwo user,pid,ppid,pgid,command"
 #  define PSFORMAT "%s %ld %ld %ld %[^\n]"
 #  define PSVARS   P[i].name, [i].pid, [i].ppid, [i].pgid, P[i].cmd
 #  define PSVARSN  5
@@ -872,6 +872,11 @@ int main(int argc, char **argv) {
 #ifdef ZOMBIES_HAVE_PID_0
   FixZombies();
 #endif
+
+  if (pledge("stdio getpw ioctl proc", NULL) == -1){
+fprintf(stderr, "%s: pledge\n", Progname);
+   exit(1);
+  }

   if (NProc == 0) {
 fprintf(stderr, "%s: No processes read.\n", Progname);