[new port] getmail6

[Martin Ziemer]

This tarball contains a port of getmail6, which is a python 3 fork of
getmail.

I had to modify the getmails script to work on OpenBSD. While there i 
added a ignore for directories in the config folder, as those stopped
getmails.

Old getmail configurations still work so migration is only pkg_delete
getmail and pkg_add getmail6 for existing users.

Use it on my daily driver with no problems.


getmail6.tgz
Description: application/tar-gz

Re: UPDATE: x11/qt6

[Klemens Nanni]

On Mon, Dec 05, 2022 at 10:05:55PM +0100, Rafael Sadowski wrote:
> On Mon Dec 05, 2022 at 06:49:40PM +0100, Rafael Sadowski wrote:
> > Here is an update from Qt 6.3.1 to 6.4.1. Please test with your favorite
> > consumers.
> > 
> > Cheers,
> > 
> > Rafael
> > 
> 
> Tweaked after feedback form tb@. Build test with all consumers.

Rebuilt tdesktop runs fine with it on amd64.
OK kn

Re: [new port] getmail6

[Klemens Nanni]

On Tue, Dec 06, 2022 at 09:29:54AM +0100, Martin Ziemer wrote:
> This tarball contains a port of getmail6, which is a python 3 fork of
> getmail.
> 
> I had to modify the getmails script to work on OpenBSD. While there i 
> added a ignore for directories in the config folder, as those stopped
> getmails.
> 
> Old getmail configurations still work so migration is only pkg_delete
> getmail and pkg_add getmail6 for existing users.
> 
> Use it on my daily driver with no problems.

Just use GH_* instead of hand-rolling it with archive/refs/tags/, this
is shorter, fixes distinfo (tarball name is version only) and WRKDIST
can go as well and post-install just uses WRKSRC/GH_TAGNAME.

I'd add "(python 3 fork)" from DESCR to COMMENT as well.

mail/getmail and mail/getmail6 install the same files, so you'll need
a @conflict marker.

Rest looks OK port-wise.

commenting out `set -e' in the script seems odd, maybe something you'd
want to fix, with upstream?

Re: [new port] getmail6

[Stuart Henderson]

On 2022/12/06 14:21, Klemens Nanni wrote:
> On Tue, Dec 06, 2022 at 09:29:54AM +0100, Martin Ziemer wrote:
> > This tarball contains a port of getmail6, which is a python 3 fork of
> > getmail.
> > 
> > I had to modify the getmails script to work on OpenBSD. While there i 
> > added a ignore for directories in the config folder, as those stopped
> > getmails.
> > 
> > Old getmail configurations still work so migration is only pkg_delete
> > getmail and pkg_add getmail6 for existing users.
> > 
> > Use it on my daily driver with no problems.
> 
> Just use GH_* instead of hand-rolling it with archive/refs/tags/, this
> is shorter, fixes distinfo (tarball name is version only) and WRKDIST
> can go as well and post-install just uses WRKSRC/GH_TAGNAME.
> 
> I'd add "(python 3 fork)" from DESCR to COMMENT as well.
> 
> mail/getmail and mail/getmail6 install the same files, so you'll need
> a @conflict marker.

Or just update mail/getmail to this version.

> Rest looks OK port-wise.
> 
> commenting out `set -e' in the script seems odd, maybe something you'd
> want to fix, with upstream?
> 

Re: [new port] getmail6

[Stuart Henderson]

On 2022/12/06 15:07, Stuart Henderson wrote:
> On 2022/12/06 14:21, Klemens Nanni wrote:
> > On Tue, Dec 06, 2022 at 09:29:54AM +0100, Martin Ziemer wrote:
> > > This tarball contains a port of getmail6, which is a python 3 fork of
> > > getmail.
> > > 
> > > I had to modify the getmails script to work on OpenBSD. While there i 
> > > added a ignore for directories in the config folder, as those stopped
> > > getmails.
> > > 
> > > Old getmail configurations still work so migration is only pkg_delete
> > > getmail and pkg_add getmail6 for existing users.
> > > 
> > > Use it on my daily driver with no problems.
> > 
> > Just use GH_* instead of hand-rolling it with archive/refs/tags/, this
> > is shorter, fixes distinfo (tarball name is version only) and WRKDIST
> > can go as well and post-install just uses WRKSRC/GH_TAGNAME.

oh, it's on pypi, probably better to use that instead


> > 
> > I'd add "(python 3 fork)" from DESCR to COMMENT as well.
> > 
> > mail/getmail and mail/getmail6 install the same files, so you'll need
> > a @conflict marker.
> 
> Or just update mail/getmail to this version.
> 
> > Rest looks OK port-wise.
> > 
> > commenting out `set -e' in the script seems odd, maybe something you'd
> > want to fix, with upstream?
> > 
> 
> 

shells/bash: dlopen() of modules fails

[Christian Weisgerber]

A little-known feature of bash is that it can be extended with
loadable modules.  A number of example modules are built by default
and shipped in /usr/local/lib/bash.  However, loading some of those
modules fails because of missing symbols:

  bash$ enable finfo
  bash:/usr/local/lib/bash/finfo: undefined symbol 'sh_optind'
  bash:/usr/local/lib/bash/finfo: undefined symbol 'sh_optarg'
  bash: enable: finfo: not a shell builtin

This appears to concern bss symbols that _are_ defined in bash
itself:

  $ nm -g /usr/local/lib/bash/finfo | egrep 'sh_opt(ind|arg)'  
   U sh_optarg
   U sh_optind
  $ nm -g /usr/local/bin/bash | egrep 'sh_opt(ind|arg)'   
  00136588 B sh_optarg
  00136590 B sh_optind

I don't understand why this fails.  Note that this has been broken
for months, if not forever.  It is not a recent error.

Running bash with LD_DEBUG=1 provides no further information:

  bash$ enable finfo
  dlopen: loading: /usr/local/lib/bash/finfo
   flags /usr/local/lib/bash/finfo = 0x0
  head /usr/local/lib/bash/finfo
  obj /usr/local/lib/bash/finfo has /usr/local/lib/bash/finfo as head
  linking /usr/local/lib/bash/finfo as dlopen()ed
  head [/usr/local/lib/bash/finfo]
  examining: '/usr/local/lib/bash/finfo'
  tail /usr/local/lib/bash/finfo
  bash:/usr/local/lib/bash/finfo: undefined symbol 'sh_optind'
  bash:/usr/local/lib/bash/finfo: undefined symbol 'sh_optarg'
  unload_shlib called on /usr/local/lib/bash/finfo
  unload_shlib unloading on /usr/local/lib/bash/finfo
  dlopen: /usr/local/lib/bash/finfo: done (failed).
  dlopen: loading: finfo
  dlopen: failed to open finfo
  dlopen: finfo: done (failed).
  bash: enable: finfo: not a shell builtin

I'm at a loss.

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

Re: [new port] getmail6

[Martin Ziemer]

Am Tue, Dec 06, 2022 at 02:21:06PM + schrieb Klemens Nanni:
> On Tue, Dec 06, 2022 at 09:29:54AM +0100, Martin Ziemer wrote:
> > This tarball contains a port of getmail6, which is a python 3 fork of
> > getmail.
> > 
> > I had to modify the getmails script to work on OpenBSD. While there i 
> > added a ignore for directories in the config folder, as those stopped
> > getmails.
> > 
> > Old getmail configurations still work so migration is only pkg_delete
> > getmail and pkg_add getmail6 for existing users.
> > 
> > Use it on my daily driver with no problems.
> 
> Just use GH_* instead of hand-rolling it with archive/refs/tags/, this
> is shorter, fixes distinfo (tarball name is version only) and WRKDIST
> can go as well and post-install just uses WRKSRC/GH_TAGNAME.
> 
> I'd add "(python 3 fork)" from DESCR to COMMENT as well.
> 
> mail/getmail and mail/getmail6 install the same files, so you'll need
> a @conflict marker.
> 
> Rest looks OK port-wise.
Updated the port with GH_* and DESCR.
Added @conflict marker to this port and think, we should add the same
to the old getmail.

> commenting out `set -e' in the script seems odd, maybe something you'd
> want to fix, with upstream?
The 'set -e' breaks the script for us, as the shell exits, if no
process is running. Have no Linux System at hand to see, how it
behaves there. (But will find one at work and then think about
contacting upstream) 


getmail6.tgz
Description: application/tar-gz

Re: [UPDATE] lang/node 18.12.1

[Volker Schlecht]

On 12/6/22 06:57, Theo Buehler wrote:

I don't care deeply if OpenSSL 3 is used for this, I just wanted to make
sure the switch wasn't made blindly since newer is better (which in this
case it probably isn't).


It really wasn't - I just intend to stick as closely to upstream's choices
as possible without having to get the bundled OpenSSL dependency to build
and work.


Thanks for the explanations. That's more than good enough for me.

The patch doesn't build for me since it tries to fetch zwitch from the
net. That's on aarch64 in case it matters:


It looks like you'running 'make test', which fails because tests can't 
be run as _pbuild.

Re: [new port] getmail6

[Stuart Henderson]

ah, there's no sdist on pypi, only a wheel.  how about this instead?

I have left the getmails patch alone for now as I can't test it but the
pgrep invocation is wrong, it should probably search for something like
"^/bin/sh /usr/local/bin/getmails$" and then I expect the set -e will
work.


Index: Makefile
===
RCS file: /cvs/ports/mail/getmail/Makefile,v
retrieving revision 1.102
diff -u -p -r1.102 Makefile
--- Makefile11 Mar 2022 19:34:37 -  1.102
+++ Makefile6 Dec 2022 15:17:29 -
@@ -1,24 +1,22 @@
 COMMENT=   IMAP/POP3/SDPS mail retriever
 
-MODPY_EGG_VERSION= 5.16
-DISTNAME=  getmail-${MODPY_EGG_VERSION}
+MODPY_EGG_VERSION= 6.18.10
+GH_ACCOUNT=getmail6
+GH_PROJECT=getmail6
+GH_TAGNAME=v${MODPY_EGG_VERSION}
+PKGNAME=   getmail-${MODPY_EGG_VERSION}
 CATEGORIES=mail
 
-HOMEPAGE=  http://pyropus.ca/software/getmail/
+# https://github.com/getmail6/getmail6
+HOMEPAGE=  https://getmail6.org/
 
 MAINTAINER=Martin Ziemer 
 
 # GPLv2
 PERMIT_PACKAGE=Yes
 
-MASTER_SITES=  ${HOMEPAGE}/old-versions/
-
 MODULES=   lang/python
-
-# No Python 3 support as of 5.16
-# https://pyropus.ca/software/getmail/documentation.html#python3
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
-
+MODPY_PYBUILD= setuptools
 NO_TEST=   Yes
 
 EXDIR= ${PREFIX}/share/examples/getmail
Index: distinfo
===
RCS file: /cvs/ports/mail/getmail/distinfo,v
retrieving revision 1.81
diff -u -p -r1.81 distinfo
--- distinfo24 Nov 2021 09:06:39 -  1.81
+++ distinfo6 Dec 2022 15:17:29 -
@@ -1,2 +1,2 @@
-SHA256 (getmail-5.16.tar.gz) = auj46u+avEZQUMO2TlWjGvvc1Mbt8xl7W1m71WymZ/o=
-SIZE (getmail-5.16.tar.gz) = 180577
+SHA256 (getmail6-6.18.10.tar.gz) = DlYXz6LMh8WyWWNPWfVyjfOyVoqSyqVkdn4lb3mGatY=
+SIZE (getmail6-6.18.10.tar.gz) = 205481
Index: patches/patch-getmails
===
RCS file: patches/patch-getmails
diff -N patches/patch-getmails
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-getmails  6 Dec 2022 15:17:29 -
@@ -0,0 +1,52 @@
+# Fix for use on OpenBSD and ignore Directories in folder
+
+Index: getmails
+--- getmails.orig
 getmails
+@@ -9,7 +9,7 @@ Since getmail6 v6.18, if no --rcfile ,
+ getmail itself fetches using all config files, but serially.
+ '
+ 
+-set -e
++#set -e
+ if [ "$1" = "-p" ]; then
+ para=true
+ shift
+@@ -29,7 +29,7 @@ BASE1=${1##*/}
+ }
+ UID_BY_ID=$(id -u)
+ PID_GETMAILS=$(pgrep -U $UID_BY_ID '^getmails$')
+-if [ "x$PID_GETMAILS" != "x$$" ]; then
++if [ "x${PID_GETMAILS}x" != "xx" ]; then
+   echo "The getmails script is already running as PID=\"$PID_GETMAILS\" 
." >&2
+   exit 1
+ fi
+@@ -44,7 +44,7 @@ if [ -f $getmailrcdir/stop ]; then
+   echo "Do not run getmail ... (if not, remove $getmailrcdir/stop)" >&2
+   exit 1
+ fi
+-rcfiles="/usr/bin/getmail"
++rcfiles="/usr/local/bin/getmail"
+ # Address concerns raised by #863856
+ #  emacs backup files:   foo~ foo#
+ #  vim backup files: foo~ foo.swp
+@@ -57,7 +57,8 @@ if $para ; then
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   $rcfiles --rcfile "$file" "$@" &
+   pids="$pids $!"
+ fi
+@@ -79,7 +80,8 @@ else
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   rcfiles="$rcfiles --rcfile \"$file\""
+ fi
+ done
Index: patches/patch-setup_py
===
RCS file: /cvs/ports/mail/getmail/patches/patch-setup_py,v
retrieving revision 1.5
diff -u -p -r1.5 patch-setup_py
--- patches/patch-setup_py  11 Mar 2022 19:34:37 -  1.5
+++ patches/patch-setup_py  6 Dec 2022 15:17:29 -
@@ -1,25 +1,16 @@
 Index: setup.py
 --- setup.py.orig
 +++ setup.py
-@@ -39,12 +39,11 @@ GETMAILDOCDIR = os.path.join(
- datadir or prefix,
- 'share',
- 'doc',
--'getmail-%s' % __version__
-+'getmail'
- )
+@@ -36,10 +36,10 @@ for (pos, arg) in enumerate(args):
+ # hack hack hack hack hack hack hack
+ datadir = args[pos + 1]
  
- GETMAILMANDIR = os.path.join(
- datadir or prefix,
--'share',
- 'man',
- 'man1'
- )
-@@ -101,7 +100,6 @@ setup(
- data_files=[
- (GETMAILDOCDIR, [
- './README',
--'./getmail.spec',
- 'docs/BUGS',
- 'docs/COPYING',
- 'docs/CHANGELOG',
+-DOCDIR = os.path.join('share','doc','getmail-%s' % __version__)
++

Re: [UPDATE] lang/node 18.12.1

[Theo Buehler]

> It looks like you'running 'make test', which fails because tests can't be
> run as _pbuild.

Oh yes, indeed. Thanks.

Re: [new port] getmail6

[Stuart Henderson]

On 2022/12/06 15:22, Stuart Henderson wrote:
> ah, there's no sdist on pypi, only a wheel.  how about this instead?
> 
> I have left the getmails patch alone for now as I can't test it but the
> pgrep invocation is wrong, it should probably search for something like
> "^/bin/sh /usr/local/bin/getmails$" and then I expect the set -e will
> work.

...or here's an (untested) version with that proposed change.
(sorry for the spam!)

Index: Makefile
===
RCS file: /cvs/ports/mail/getmail/Makefile,v
retrieving revision 1.102
diff -u -p -r1.102 Makefile
--- Makefile11 Mar 2022 19:34:37 -  1.102
+++ Makefile6 Dec 2022 15:24:14 -
@@ -1,24 +1,22 @@
 COMMENT=   IMAP/POP3/SDPS mail retriever
 
-MODPY_EGG_VERSION= 5.16
-DISTNAME=  getmail-${MODPY_EGG_VERSION}
+MODPY_EGG_VERSION= 6.18.10
+GH_ACCOUNT=getmail6
+GH_PROJECT=getmail6
+GH_TAGNAME=v${MODPY_EGG_VERSION}
+PKGNAME=   getmail-${MODPY_EGG_VERSION}
 CATEGORIES=mail
 
-HOMEPAGE=  http://pyropus.ca/software/getmail/
+# https://github.com/getmail6/getmail6
+HOMEPAGE=  https://getmail6.org/
 
 MAINTAINER=Martin Ziemer 
 
 # GPLv2
 PERMIT_PACKAGE=Yes
 
-MASTER_SITES=  ${HOMEPAGE}/old-versions/
-
 MODULES=   lang/python
-
-# No Python 3 support as of 5.16
-# https://pyropus.ca/software/getmail/documentation.html#python3
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
-
+MODPY_PYBUILD= setuptools
 NO_TEST=   Yes
 
 EXDIR= ${PREFIX}/share/examples/getmail
Index: distinfo
===
RCS file: /cvs/ports/mail/getmail/distinfo,v
retrieving revision 1.81
diff -u -p -r1.81 distinfo
--- distinfo24 Nov 2021 09:06:39 -  1.81
+++ distinfo6 Dec 2022 15:24:14 -
@@ -1,2 +1,2 @@
-SHA256 (getmail-5.16.tar.gz) = auj46u+avEZQUMO2TlWjGvvc1Mbt8xl7W1m71WymZ/o=
-SIZE (getmail-5.16.tar.gz) = 180577
+SHA256 (getmail6-6.18.10.tar.gz) = DlYXz6LMh8WyWWNPWfVyjfOyVoqSyqVkdn4lb3mGatY=
+SIZE (getmail6-6.18.10.tar.gz) = 205481
Index: patches/patch-getmails
===
RCS file: patches/patch-getmails
diff -N patches/patch-getmails
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-getmails  6 Dec 2022 15:24:14 -
@@ -0,0 +1,43 @@
+# Fix for use on OpenBSD and ignore Directories in folder
+
+Index: getmails
+--- getmails.orig
 getmails
+@@ -28,7 +28,7 @@ BASE1=${1##*/}
+ [ "$BASE1" != "${BASE1#$2}" ] && return 0 || return 1
+ }
+ UID_BY_ID=$(id -u)
+-PID_GETMAILS=$(pgrep -U $UID_BY_ID '^getmails$')
++PID_GETMAILS=$(pgrep -U $UID_BY_ID '^/bin/sh /usr/local/bin/getmails$')
+ if [ "x$PID_GETMAILS" != "x$$" ]; then
+   echo "The getmails script is already running as PID=\"$PID_GETMAILS\" 
." >&2
+   exit 1
+@@ -44,7 +44,7 @@ if [ -f $getmailrcdir/stop ]; then
+   echo "Do not run getmail ... (if not, remove $getmailrcdir/stop)" >&2
+   exit 1
+ fi
+-rcfiles="/usr/bin/getmail"
++rcfiles="/usr/local/bin/getmail"
+ # Address concerns raised by #863856
+ #  emacs backup files:   foo~ foo#
+ #  vim backup files: foo~ foo.swp
+@@ -57,7 +57,8 @@ if $para ; then
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   $rcfiles --rcfile "$file" "$@" &
+   pids="$pids $!"
+ fi
+@@ -79,7 +80,8 @@ else
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   rcfiles="$rcfiles --rcfile \"$file\""
+ fi
+ done
Index: patches/patch-setup_py
===
RCS file: /cvs/ports/mail/getmail/patches/patch-setup_py,v
retrieving revision 1.5
diff -u -p -r1.5 patch-setup_py
--- patches/patch-setup_py  11 Mar 2022 19:34:37 -  1.5
+++ patches/patch-setup_py  6 Dec 2022 15:24:14 -
@@ -1,25 +1,16 @@
 Index: setup.py
 --- setup.py.orig
 +++ setup.py
-@@ -39,12 +39,11 @@ GETMAILDOCDIR = os.path.join(
- datadir or prefix,
- 'share',
- 'doc',
--'getmail-%s' % __version__
-+'getmail'
- )
+@@ -36,10 +36,10 @@ for (pos, arg) in enumerate(args):
+ # hack hack hack hack hack hack hack
+ datadir = args[pos + 1]
  
- GETMAILMANDIR = os.path.join(
- datadir or prefix,
--'share',
- 'man',
- 'man1'
- )
-@@ -101,7 +100,6 @@ setup(
- data_files=[
- (GETMAILDOCDIR, [
- './README',
--'./getmail.spec',
- 'docs/BUGS',
- 'docs/COPYING',
- 'docs/CHANGELOG',
+-DOCDIR = os.path.join('share','doc','getmai

Re: [new port] getmail6

[Klemens Nanni]

On Tue, Dec 06, 2022 at 03:24:51PM +, Stuart Henderson wrote:
> On 2022/12/06 15:22, Stuart Henderson wrote:
> > ah, there's no sdist on pypi, only a wheel.  how about this instead?
> > 
> > I have left the getmails patch alone for now as I can't test it but the
> > pgrep invocation is wrong, it should probably search for something like
> > "^/bin/sh /usr/local/bin/getmails$" and then I expect the set -e will
> > work.
> 
> ...or here's an (untested) version with that proposed change.
> (sorry for the spam!)

appreciated :)

Both Martin's latest tarball and your diff are OK with me.

If indeed everything works just like the py2 port, then I prefer the
straight update.

But I'll leave that decision to Martin as maintainer/user.


> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/mail/getmail/Makefile,v
> retrieving revision 1.102
> diff -u -p -r1.102 Makefile
> --- Makefile  11 Mar 2022 19:34:37 -  1.102
> +++ Makefile  6 Dec 2022 15:24:14 -
> @@ -1,24 +1,22 @@
>  COMMENT= IMAP/POP3/SDPS mail retriever
>  
> -MODPY_EGG_VERSION=   5.16
> -DISTNAME=getmail-${MODPY_EGG_VERSION}
> +MODPY_EGG_VERSION=   6.18.10
> +GH_ACCOUNT=  getmail6
> +GH_PROJECT=  getmail6
> +GH_TAGNAME=  v${MODPY_EGG_VERSION}
> +PKGNAME= getmail-${MODPY_EGG_VERSION}
>  CATEGORIES=  mail
>  
> -HOMEPAGE=http://pyropus.ca/software/getmail/
> +# https://github.com/getmail6/getmail6
> +HOMEPAGE=https://getmail6.org/
>  
>  MAINTAINER=  Martin Ziemer 
>  
>  # GPLv2
>  PERMIT_PACKAGE=  Yes
>  
> -MASTER_SITES=${HOMEPAGE}/old-versions/
> -
>  MODULES= lang/python
> -
> -# No Python 3 support as of 5.16
> -# https://pyropus.ca/software/getmail/documentation.html#python3
> -MODPY_VERSION =  ${MODPY_DEFAULT_VERSION_2}
> -
> +MODPY_PYBUILD=   setuptools
>  NO_TEST= Yes
>  
>  EXDIR=   ${PREFIX}/share/examples/getmail
> Index: distinfo
> ===
> RCS file: /cvs/ports/mail/getmail/distinfo,v
> retrieving revision 1.81
> diff -u -p -r1.81 distinfo
> --- distinfo  24 Nov 2021 09:06:39 -  1.81
> +++ distinfo  6 Dec 2022 15:24:14 -
> @@ -1,2 +1,2 @@
> -SHA256 (getmail-5.16.tar.gz) = auj46u+avEZQUMO2TlWjGvvc1Mbt8xl7W1m71WymZ/o=
> -SIZE (getmail-5.16.tar.gz) = 180577
> +SHA256 (getmail6-6.18.10.tar.gz) = 
> DlYXz6LMh8WyWWNPWfVyjfOyVoqSyqVkdn4lb3mGatY=
> +SIZE (getmail6-6.18.10.tar.gz) = 205481
> Index: patches/patch-getmails
> ===
> RCS file: patches/patch-getmails
> diff -N patches/patch-getmails
> --- /dev/null 1 Jan 1970 00:00:00 -
> +++ patches/patch-getmails6 Dec 2022 15:24:14 -
> @@ -0,0 +1,43 @@
> +# Fix for use on OpenBSD and ignore Directories in folder
> +
> +Index: getmails
> +--- getmails.orig
>  getmails
> +@@ -28,7 +28,7 @@ BASE1=${1##*/}
> + [ "$BASE1" != "${BASE1#$2}" ] && return 0 || return 1
> + }
> + UID_BY_ID=$(id -u)
> +-PID_GETMAILS=$(pgrep -U $UID_BY_ID '^getmails$')
> ++PID_GETMAILS=$(pgrep -U $UID_BY_ID '^/bin/sh /usr/local/bin/getmails$')
> + if [ "x$PID_GETMAILS" != "x$$" ]; then
> + echo "The getmails script is already running as PID=\"$PID_GETMAILS\" 
> ." >&2
> + exit 1
> +@@ -44,7 +44,7 @@ if [ -f $getmailrcdir/stop ]; then
> + echo "Do not run getmail ... (if not, remove $getmailrcdir/stop)" >&2
> + exit 1
> + fi
> +-rcfiles="/usr/bin/getmail"
> ++rcfiles="/usr/local/bin/getmail"
> + # Address concerns raised by #863856
> + #  emacs backup files:   foo~ foo#
> + #  vim backup files: foo~ foo.swp
> +@@ -57,7 +57,8 @@ if $para ; then
> +! endwith "$file" '#' && \
> +! startswith "$file" 'oldmail-' && \
> +! endwith "$file" '.swp' && \
> +-   ! endwith "$file" '.bak' ; then
> ++   ! endwith "$file" '.bak' && \
> ++   [ -f "$file" ]; then
> + $rcfiles --rcfile "$file" "$@" &
> + pids="$pids $!"
> + fi
> +@@ -79,7 +80,8 @@ else
> +! endwith "$file" '#' && \
> +! startswith "$file" 'oldmail-' && \
> +! endwith "$file" '.swp' && \
> +-   ! endwith "$file" '.bak' ; then
> ++   ! endwith "$file" '.bak' && \
> ++   [ -f "$file" ]; then
> + rcfiles="$rcfiles --rcfile \"$file\""
> + fi
> + done
> Index: patches/patch-setup_py
> ===
> RCS file: /cvs/ports/mail/getmail/patches/patch-setup_py,v
> retrieving revision 1.5
> diff -u -p -r1.5 patch-setup_py
> --- patches/patch-setup_py11 Mar 2022 19:34:37 -  1.5
> +++ patches/patch-setup_py6 Dec 2022 15:24:14 -
> @@ -1,25 +1,16 @@
>  Index: setup.py
>  --- setup.py.orig
>  +++ setup.py
> -@@ -39,12 +39,11 @@ GETMAILDOCDIR = os.path.join(
> - datadir or prefix,
> - 'share',
> - 'doc',
> --'getmail-%s' % __version__
> -+'ge

Re: [new port] getmail6

[Martin Ziemer]

Am Tue, Dec 06, 2022 at 03:24:51PM + schrieb Stuart Henderson:
> On 2022/12/06 15:22, Stuart Henderson wrote:
> > I have left the getmails patch alone for now as I can't test it but the
> > pgrep invocation is wrong, it should probably search for something like
> > "^/bin/sh /usr/local/bin/getmails$" and then I expect the set -e will
> > work.
> 
> ...or here's an (untested) version with that proposed change.
> (sorry for the spam!)

> +-PID_GETMAILS=$(pgrep -U $UID_BY_ID '^getmails$')
> ++PID_GETMAILS=$(pgrep -U $UID_BY_ID '^/bin/sh /usr/local/bin/getmails$')
Just tested the getmails change: it still exits at the pgrep line. 

Re: [new port] getmail6

[Martin Ziemer]

Am Tue, Dec 06, 2022 at 03:29:18PM + schrieb Klemens Nanni:
> On Tue, Dec 06, 2022 at 03:24:51PM +, Stuart Henderson wrote:
> > On 2022/12/06 15:22, Stuart Henderson wrote:
> > ...or here's an (untested) version with that proposed change.
> > (sorry for the spam!)
> Both Martin's latest tarball and your diff are OK with me.
> 
> If indeed everything works just like the py2 port, then I prefer the
> straight update.
> 
> But I'll leave that decision to Martin as maintainer/user.
I guess, we should use the diff from Stuart to existing getmail.
That way no Users will have to do anything and we get rid of the
python2-part without modifying another port. 

Index: Makefile
===
RCS file: /cvs/ports/mail/getmail/Makefile,v
retrieving revision 1.102
diff -u -p -r1.102 Makefile
--- Makefile11 Mar 2022 19:34:37 -  1.102
+++ Makefile6 Dec 2022 15:17:29 -
@@ -1,24 +1,22 @@
 COMMENT=   IMAP/POP3/SDPS mail retriever
 
-MODPY_EGG_VERSION= 5.16
-DISTNAME=  getmail-${MODPY_EGG_VERSION}
+MODPY_EGG_VERSION= 6.18.10
+GH_ACCOUNT=getmail6
+GH_PROJECT=getmail6
+GH_TAGNAME=v${MODPY_EGG_VERSION}
+PKGNAME=   getmail-${MODPY_EGG_VERSION}
 CATEGORIES=mail
 
-HOMEPAGE=  http://pyropus.ca/software/getmail/
+# https://github.com/getmail6/getmail6
+HOMEPAGE=  https://getmail6.org/
 
 MAINTAINER=Martin Ziemer 
 
 # GPLv2
 PERMIT_PACKAGE=Yes
 
-MASTER_SITES=  ${HOMEPAGE}/old-versions/
-
 MODULES=   lang/python
-
-# No Python 3 support as of 5.16
-# https://pyropus.ca/software/getmail/documentation.html#python3
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
-
+MODPY_PYBUILD= setuptools
 NO_TEST=   Yes
 
 EXDIR= ${PREFIX}/share/examples/getmail
Index: distinfo
===
RCS file: /cvs/ports/mail/getmail/distinfo,v
retrieving revision 1.81
diff -u -p -r1.81 distinfo
--- distinfo24 Nov 2021 09:06:39 -  1.81
+++ distinfo6 Dec 2022 15:17:29 -
@@ -1,2 +1,2 @@
-SHA256 (getmail-5.16.tar.gz) = auj46u+avEZQUMO2TlWjGvvc1Mbt8xl7W1m71WymZ/o=
-SIZE (getmail-5.16.tar.gz) = 180577
+SHA256 (getmail6-6.18.10.tar.gz) = DlYXz6LMh8WyWWNPWfVyjfOyVoqSyqVkdn4lb3mGatY=
+SIZE (getmail6-6.18.10.tar.gz) = 205481
Index: patches/patch-getmails
===
RCS file: patches/patch-getmails
diff -N patches/patch-getmails
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-getmails  6 Dec 2022 15:17:29 -
@@ -0,0 +1,52 @@
+# Fix for use on OpenBSD and ignore Directories in folder
+
+Index: getmails
+--- getmails.orig
 getmails
+@@ -9,7 +9,7 @@ Since getmail6 v6.18, if no --rcfile ,
+ getmail itself fetches using all config files, but serially.
+ '
+ 
+-set -e
++#set -e
+ if [ "$1" = "-p" ]; then
+ para=true
+ shift
+@@ -29,7 +29,7 @@ BASE1=${1##*/}
+ }
+ UID_BY_ID=$(id -u)
+ PID_GETMAILS=$(pgrep -U $UID_BY_ID '^getmails$')
+-if [ "x$PID_GETMAILS" != "x$$" ]; then
++if [ "x${PID_GETMAILS}x" != "xx" ]; then
+   echo "The getmails script is already running as PID=\"$PID_GETMAILS\" 
." >&2
+   exit 1
+ fi
+@@ -44,7 +44,7 @@ if [ -f $getmailrcdir/stop ]; then
+   echo "Do not run getmail ... (if not, remove $getmailrcdir/stop)" >&2
+   exit 1
+ fi
+-rcfiles="/usr/bin/getmail"
++rcfiles="/usr/local/bin/getmail"
+ # Address concerns raised by #863856
+ #  emacs backup files:   foo~ foo#
+ #  vim backup files: foo~ foo.swp
+@@ -57,7 +57,8 @@ if $para ; then
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   $rcfiles --rcfile "$file" "$@" &
+   pids="$pids $!"
+ fi
+@@ -79,7 +80,8 @@ else
+! endwith "$file" '#' && \
+! startswith "$file" 'oldmail-' && \
+! endwith "$file" '.swp' && \
+-   ! endwith "$file" '.bak' ; then
++   ! endwith "$file" '.bak' && \
++   [ -f "$file" ]; then
+   rcfiles="$rcfiles --rcfile \"$file\""
+ fi
+ done
Index: patches/patch-setup_py
===
RCS file: /cvs/ports/mail/getmail/patches/patch-setup_py,v
retrieving revision 1.5
diff -u -p -r1.5 patch-setup_py
--- patches/patch-setup_py  11 Mar 2022 19:34:37 -  1.5
+++ patches/patch-setup_py  6 Dec 2022 15:17:29 -
@@ -1,25 +1,16 @@
 Index: setup.py
 --- setup.py.orig
 +++ setup.py
-@@ -39,12 +39,11 @@ GETMAILDOCDIR = os.path.join(
- datadir or prefix,
- 'share',
- 'doc',
--'getmail-%s' % __version__
-+'getmail'
- )
+@@ -36,10 +36,10 @@ for (pos, arg) in enumerate(args):
+ # hack hack hack hack hack hack hack
+ datadir = args[pos + 1]
  
- GETMAILMANDIR 

[new port] audio/d11amp

[Thomas Dettbarn]

Hello.


So, I started writing a port for my retro MP3 player d11amp.
You can find the project's homepage here: https://dettus.net/d11amp


The port is what I have attached to this email. I tested it by
putting it into the ports tree and running "make install".

It would be great if it could become part of the official OpenBSD
ports tree one day, but only if it meets your standards. 


Thus far, I am not 100% happy, since I was forced to include a
do-build target. Now, its very existence seems odd to me. May I 
bother one of you more experienced porters to have a look? Is it
really necessary or did I do something wrong?
(see below)


Thomas 


-x-x-x-x-x-x-x-x-x-x (from the Makefile)
MAKE_FLAGS= CC="${CC}" CFLAGS="${CFLAGS}"

do-build:
cd ${WRKBUILD} && pwd && make

do-install:
...
-x-x-x-x-x-x-x-x-x-x

audio_d11amp_059.tgz
Description: application/compressed-tar

Re: [new port] audio/d11amp

[Stuart Henderson]

Hi, untested but I think you are getting bitten by the default
"ALL_TARGET=all"; try just setting ALL_TARGET= in the port.


On 2022/12/06 17:40, Thomas Dettbarn wrote:
> Hello.
> 
> 
> So, I started writing a port for my retro MP3 player d11amp.
> You can find the project's homepage here: https://dettus.net/d11amp
> 
> 
> The port is what I have attached to this email. I tested it by
> putting it into the ports tree and running "make install".
> 
> It would be great if it could become part of the official OpenBSD
> ports tree one day, but only if it meets your standards. 
> 
> 
> Thus far, I am not 100% happy, since I was forced to include a
> do-build target. Now, its very existence seems odd to me. May I 
> bother one of you more experienced porters to have a look? Is it
> really necessary or did I do something wrong?
> (see below)
> 
> 
> Thomas 
> 
> 
> -x-x-x-x-x-x-x-x-x-x (from the Makefile)
> MAKE_FLAGS=   CC="${CC}" CFLAGS="${CFLAGS}"
> 
> do-build:
>   cd ${WRKBUILD} && pwd && make
> 
> do-install:
> ...
> -x-x-x-x-x-x-x-x-x-x

Re: [UPDATE] lang/node 18.12.1

[Volker Schlecht]

Someone in nodejs' github found an embarrassing typo (thanks barrykn!). 
Updated patch is attached!


On 12/6/22 16:23, Theo Buehler wrote:

It looks like you'running 'make test', which fails because tests can't be
run as _pbuild.


Oh yes, indeed. Thanks.
Index: Makefile
===
RCS file: /cvs/ports/lang/node/Makefile,v
retrieving revision 1.111
diff -u -p -r1.111 Makefile
--- Makefile	13 Nov 2022 15:28:44 -	1.111
+++ Makefile	6 Dec 2022 15:52:37 -
@@ -5,12 +5,11 @@ USE_WXNEEDED =		Yes
 
 COMMENT = JavaScript runtime built on Chrome's V8 JavaScript engine
 
-NODE_VERSION =		v16.17.1
-PLEDGE_VER =		1.1.2
+NODE_VERSION =		v18.12.1
+PLEDGE_VER =		1.1.3
 DISTFILES =		node-pledge-{}${PLEDGE_VER}.tar.gz:0 \
 			${DISTNAME}-headers.tar.gz \
 			${DISTNAME}.tar.xz
-REVISION =		2
 
 DISTNAME =		node-${NODE_VERSION}
 PKGNAME =		${DISTNAME:S/v//g}
@@ -34,7 +33,7 @@ MODULES =		lang/python
 WANTLIB += c execinfo m pthread ${COMPILER_LIBCXX}
 WANTLIB += z brotlienc brotlidec
 WANTLIB += icudata icui18n icuuc cares nghttp2 uv
-WANTLIB += lib/eopenssl11/ssl lib/eopenssl11/crypto
+WANTLIB += lib/eopenssl30/ssl lib/eopenssl30/crypto
 
 COMPILER =		base-clang ports-gcc
 
@@ -71,16 +70,16 @@ SUBST_VARS +=		WRKDIST
 SUBST_VARS +=		NODE_VERSION
 SUBST_VARS +=		EOPENSSL_LIB
 
-# OpenSSL used: {X,Ed}25519 via EVP, SSL_CIPHER_standard_name and 5-10 other missing symbols/defines.
+# uses a wide range of OpenSSL API and only really supports boring/openssl
 LIB_DEPENDS +=		archivers/brotli \
 			devel/libuv \
 			net/libcares \
 			textproc/icu4c \
 			www/nghttp2 \
-			security/openssl/1.1
+			security/openssl/3.0
 
-EOPENSSL_LIB =		${LOCALBASE}/lib/eopenssl11
-EOPENSSL_INC =		${LOCALBASE}/include/eopenssl11
+EOPENSSL_LIB =		${LOCALBASE}/lib/eopenssl30
+EOPENSSL_INC =		${LOCALBASE}/include/eopenssl30
 
 post-extract:
 	mv ${WRKDIR}/node-pledge-${PLEDGE_VER} \
Index: distinfo
===
RCS file: /cvs/ports/lang/node/distinfo,v
retrieving revision 1.65
diff -u -p -r1.65 distinfo
--- distinfo	1 Nov 2022 12:01:49 -	1.65
+++ distinfo	6 Dec 2022 15:52:37 -
@@ -1,6 +1,6 @@
-SHA256 (node-pledge-1.1.2.tar.gz) = zY/JcbZ32mmtqWXXNn3/9aTh7Y3F6fAAaADDA8SYwEk=
-SHA256 (node-v16.17.1-headers.tar.gz) = Ncy5XK8CzaO9aA2kNQqK5dZmp6nq46/lwqGz7ymu8Qg=
-SHA256 (node-v16.17.1.tar.xz) = ZyH+tBUtVtLGs1jOOXq9Wn8drwnuLiXFAhubTT+GozA=
-SIZE (node-pledge-1.1.2.tar.gz) = 3155
-SIZE (node-v16.17.1-headers.tar.gz) = 568068
-SIZE (node-v16.17.1.tar.xz) = 35661452
+SHA256 (node-pledge-1.1.3.tar.gz) = fEaXvLg6hYEJ69K+mgQFizf8DiJY2/DtyFJB/pEanVU=
+SHA256 (node-v18.12.1-headers.tar.gz) = nVXuByum1aFB2wks7xoPcV99P8k4KFptknodCgx0Qvc=
+SHA256 (node-v18.12.1.tar.xz) = T6QGRRvFJlmikOUs/bIWKnYL1UnaS4u+vmop8pbZON8=
+SIZE (node-pledge-1.1.3.tar.gz) = 3167
+SIZE (node-v18.12.1-headers.tar.gz) = 8563785
+SIZE (node-v18.12.1.tar.xz) = 38454588
Index: patches/patch-Makefile
===
RCS file: /cvs/ports/lang/node/patches/patch-Makefile,v
retrieving revision 1.15
diff -u -p -r1.15 patch-Makefile
--- patches/patch-Makefile	18 Mar 2022 19:35:16 -	1.15
+++ patches/patch-Makefile	6 Dec 2022 15:52:37 -
@@ -1,7 +1,7 @@
 Index: Makefile
 --- Makefile.orig
 +++ Makefile
-@@ -163,7 +163,7 @@ config.gypi: configure configure.py src/node_version.h
+@@ -185,7 +185,7 @@ config.gypi: configure configure.py src/node_version.h
  	fi
  
  .PHONY: install
@@ -10,7 +10,7 @@ Index: Makefile
  	$(PYTHON) tools/install.py $@ '$(DESTDIR)' '$(PREFIX)'
  
  .PHONY: uninstall
-@@ -394,6 +394,12 @@ test/addons/.buildstamp: $(ADDONS_PREREQS) \
+@@ -416,6 +416,12 @@ test/addons/.buildstamp: $(ADDONS_PREREQS) \
  # Just goes to show that recursive make really is harmful...
  # TODO(bnoordhuis) Force rebuild after gyp update.
  build-addons: | $(NODE_EXE) test/addons/.buildstamp
Index: patches/patch-common_gypi
===
RCS file: /cvs/ports/lang/node/patches/patch-common_gypi,v
retrieving revision 1.23
diff -u -p -r1.23 patch-common_gypi
--- patches/patch-common_gypi	1 Sep 2022 20:42:56 -	1.23
+++ patches/patch-common_gypi	6 Dec 2022 15:52:37 -
@@ -1,7 +1,7 @@
 Index: common.gypi
 --- common.gypi.orig
 +++ common.gypi
-@@ -416,7 +416,9 @@
+@@ -413,7 +413,9 @@
}],
['OS=="openbsd"', {
  'cflags': [ '-I/usr/local/include' ],
Index: patches/patch-deps_npm_node_modules_node-gyp_gyp_pylib_gyp_generator_make_py
===
RCS file: /cvs/ports/lang/node/patches/patch-deps_npm_node_modules_node-gyp_gyp_pylib_gyp_generator_make_py,v
retrieving revision 1.10
diff -u -p -r1.10 patch-deps_npm_node_modules_node-gyp_gyp_pylib_gyp_generator_make_py
--- patches/patch-deps_npm_node_modules_node-gyp_gyp_pylib_gyp_generator_make_py	11 Mar 2022 19:29:08 -	1.10
+++ patches/p

Re: [new port] audio/d11amp

[Jeremie Courreges-Anglas]

On Tue, Dec 06 2022, Thomas Dettbarn  wrote:
> Hello.
>
>
> So, I started writing a port for my retro MP3 player d11amp.
> You can find the project's homepage here: https://dettus.net/d11amp
>
>
> The port is what I have attached to this email. I tested it by
> putting it into the ports tree and running "make install".
>
> It would be great if it could become part of the official OpenBSD
> ports tree one day, but only if it meets your standards. 
>
>
> Thus far, I am not 100% happy, since I was forced to include a
> do-build target. Now, its very existence seems odd to me. May I 
> bother one of you more experienced porters to have a look? Is it
> really necessary or did I do something wrong?
> (see below)
>
> Thomas 
>
>
> -x-x-x-x-x-x-x-x-x-x (from the Makefile)
> MAKE_FLAGS=   CC="${CC}" CFLAGS="${CFLAGS}"
> 
> do-build:
>   cd ${WRKBUILD} && pwd && make

You're setting CFLAGS on the make command line because you spotted that
its value wasn't in control of the ports framework.  But passing CFLAGS
on the make command-line means that the CFLAGS assignement and
subsequent appends (pkg-config --cflags ...) in upstream's Makefile are
ignored, so the build fails.  Your do-build target doesn't respect
CFLAGS as set in the port Makefile, so the build succeeds.

In this kind of situation I think it's fair to patch upstream's Makefile
in order to satisfy the needs of both the ports framework and upstream's
Makefile, introducing a new variable PORTS_CFLAGS.

Here's a list of what I attempted to fix:
- zap REVISION, this is a new port
- COMMENT starts with lowercase
- move V closer to where it's used
- usual casing for PERMIT_PACKAGE
- generate WANTLIB from make port-lib-depends-check
- reorder slightly WANTLIB/MASTER_SITES (see Makefile.template)
- use PORTS_CFLAGS and patch upstream's Makefile to respect ports CFLAGS,
  zap do-build target
- override INSTALLMAN and zap do-install target
- zap empty line in PLIST - see portcheck(1)
- zap trailing whitespace in DESCR - portcheck(1).
- trim double word in DESCR.

Makefile diff below and updated tarball attached.
IMHO we could do without the licence sentence in DESCR.

The updated tarball looks ready to import, if someone wants to review
and commit it: ok jca@


--- Makefile.orig   Tue Dec  6 17:26:43 2022
+++ MakefileWed Dec  7 00:13:18 2022
@@ -1,7 +1,6 @@
-V= 0.59
-REVISION=  1
-COMMENT=   Simple MP3 player
+COMMENT=   simple MP3 player
 
+V= 0.59
 DISTNAME=  d11amp_${V}
 PKGNAME=   d11amp-${V}
 EXTRACT_SUFX=  .tar.bz2
@@ -11,30 +10,24 @@
 MAINTAINER=Thomas Dettbarn 
 
 # BSD
-PERMIT_PACKAGE=YES
+PERMIT_PACKAGE=Yes
 
+WANTLIB += c cairo cairo-gobject gdk_pixbuf-2.0 gio-2.0 glib-2.0
+WANTLIB += gobject-2.0 graphene-1.0 gtk-4 harfbuzz intl m mpg123
+WANTLIB += pango-1.0 pangocairo-1.0 portaudio pthread zip
 
+MASTER_SITES=  https://www.dettus.net/d11amp/
+
 LIB_DEPENDS=   graphics/gdk-pixbuf2\
-   x11/gtk+4   \
+   x11/gtk+4   \
archivers/libzip\
audio/mpg123\
-   audio/portaudio-svn 
-   
-WANTLIB = c gtk-4 gdk_pixbuf-2.0 zip mpg123 portaudio
+   audio/portaudio-svn
 
-MASTER_SITES=  https://www.dettus.net/d11amp/
-MAKE_FLAGS=CC="${CC}" CFLAGS="${CFLAGS}"
-
+MAKE_FLAGS=CC="${CC}" PORTS_CFLAGS="${CFLAGS}"
+FAKE_FLAGS=INSTALLMAN="${WRKINST}${PREFIX}/man"
 TEST_ENV=  SHA256_CMD=sha256 TMP_DIR=/tmp/d11amp/
 TEST_TARGET=   check
-
-
-do-build:
-   cd ${WRKBUILD} && pwd && make
-
-do-install:
-   ${INSTALL_PROGRAM} ${WRKSRC}/d11amp ${PREFIX}/bin/d11amp
-   ${INSTALL_MAN} ${WRKSRC}/d11amp.1 ${PREFIX}/man/man1
 
 .include 
 



d11amp.v2.tgz
Description: Binary data
-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

aarch64 bulk build report

[phessler]

bulk build on arm64.ports.openbsd.org
started on  Sun Dec 4 11:45:51 MST 2022
finished at Tue Dec 6 16:57:36 MST 2022
lasted 2D05h11m
done with kern.version=OpenBSD 7.2-current (GENERIC.MP) #1908: Sun Dec  4 
06:58:23 MST 2022

built packages:11388
Dec 4:3156
Dec 5:1633
Dec 6:6598


critical path missing pkgs:  
http://build-failures.rhaalovely.net/aarch64/2022-12-04/summary.log

build failures: 8
http://build-failures.rhaalovely.net/aarch64/2022-12-04/cad/openscad.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/games/lwjgl3.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/www/sope.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/x11/gnustep/dbuskit.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/x11/gnustep/gui.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/x11/gnustep/sqlclient.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/x11/gnustep/webservices.log
http://build-failures.rhaalovely.net/aarch64/2022-12-04/x11/kde-applications/step.log

recurrent failures
new failures
+++ ls-failures Tue Dec  6 16:57:48 2022
+failures/cad/openscad.log
+failures/games/lwjgl3.log
+failures/www/sope.log
+failures/x11/gnustep/dbuskit.log
+failures/x11/gnustep/gui.log
+failures/x11/gnustep/sqlclient.log
+failures/x11/gnustep/webservices.log
+failures/x11/kde-applications/step.log
resolved failures
--- ../old/aarch64/last//ls-failuresFri Dec  2 10:44:32 2022
-failures/sysutils/uefitool.log
-failures/www/ttyd.log

alpn/sem_init: OpenBSD does not support pshared

[Michel Belleau]

Hi Stuart, greetings all.

As per https://man.openbsd.org/sem_init.3

sem_init(sem_t *sem, int pshared, unsigned int value)
A non-zero value for pshared specifies a shared semaphore that can be used by 
multiple processes, which this implementation is not capable of.

Please consider having this patch in file-attach (patch-ualpn_c) applied so the 
alpn daemon can at least be used with a single worker (default is to use 2 
workers) on openbsd. Tested on 7.2 .

This fixes the EPERM error message on startup:
frontend# /tmp/ualpn -v -d -u nobody:nogroup -b 
127.0.0.1@10443 -c 
127.0.0.1@8443 -S 666
Dec 06 18:29:39 ualpn/53393: [NOTICE] version 1.7.1 starting
Dec 06 18:29:39 ualpn/53393: [NOTICE] control interface listening to 
unix:///var/run/ualpn.sock
Dec 06 18:29:39 ualpn/53393: [ERR] sem_init failed: Operation not permitted
Dec 06 18:29:39 ualpn/88515: [ERR] daemon failed to start

Once applied:
frontend# /tmp/ualpn -v -d -u nobody:nogroup -b 
127.0.0.1@10443 -c 
127.0.0.1@8443 -S 666 -n 1
Dec 06 18:29:42 ualpn/41102: [NOTICE] version 1.7.1 starting
Dec 06 18:29:42 ualpn/41102: [NOTICE] control interface listening to 
unix:///var/run/ualpn.sock
Dec 06 18:29:42 ualpn/41102: [NOTICE] frontend listening to 127.0.0.1:10443
Dec 06 18:29:42 ualpn/41102: [NOTICE] changing group (nogroup)
Dec 06 18:29:42 ualpn/41102: [NOTICE] changing user (nobody)
Dec 06 18:29:42 ualpn/33466: [NOTICE] daemon started (pid 41102)

Thank you


patch-ualpn_c
Description: patch-ualpn_c

[UPDATE] mail/aerc needs fixing

[Thim Cederlund]

Hi op@ and ports@,

I noticed that some of my directories were missing in the sidebar.

It turned out to be a bug in the maildir worker that was later patched
upstream [1]. It seems it would be appropriate to bump up the version
to a newer commit. I'm on v0.0.0-20221206102608-c1784b624aea which
released today (December 6th) and I have noticed no issues thus far.

[1]: https://git.sr.ht/~rjarry/aerc/commit/150aa0f498b9780a40fd

Now, op@ is the maintainer here so I guess he has the final say :)


Best Regards,

Thim Cederlund


Index: Makefile
===
RCS file: /home/cvs/ports/mail/aerc/Makefile,v
retrieving revision 1.6
diff -u -r1.6 Makefile
--- Makefile21 Oct 2022 11:06:08 -  1.6
+++ Makefile6 Dec 2022 20:16:07 -
@@ -1,9 +1,9 @@
 COMMENT =  terminal email client
 
 MODGO_MODNAME =git.sr.ht/~rjarry/aerc
-MODGO_VERSION =v0.0.0-20221020202109-8e37d16a32ee
+MODGO_VERSION =v0.0.0-20221206102608-c1784b624aea
 
-V =0.13.0
+V =0.13.0pl20221206
 DISTNAME = aerc-$V
 
 CATEGORIES =   mail
Index: distinfo
===
RCS file: /home/cvs/ports/mail/aerc/distinfo,v
retrieving revision 1.6
diff -u -r1.6 distinfo
--- distinfo21 Oct 2022 11:06:08 -  1.6
+++ distinfo6 Dec 2022 20:13:15 -
@@ -1,4 +1,4 @@
-SHA256 (aerc-0.13.0.zip) = IbscULse7msY7PSjGwPJqIjlpzq2HRyrMKcTcAreJzw=
+SHA256 (aerc-0.13.0pl20221206.zip) = 
hPzR4ZOY3CYE+2uj1X91YVjbZ9AJ5qystXUGsYswQrM=
 SHA256 (go_modules/4d63.com/gochecknoglobals/@v/v0.1.0.mod) = 
CoiCfB7g1EMsWIt68xomcx8db1/KEFaNOnjdCKsq+sc=
 SHA256 (go_modules/4d63.com/gochecknoglobals/@v/v0.1.0.zip) = 
9pEmszxXw/fffUClxOF+ny60YpSOF+EKe6VEKxSsBvk=
 SHA256 (go_modules/bitbucket.org/creachadair/shell/@v/v0.0.6.mod) = 
jrU1IxbKOpVIw7+aTsm0X55+XJLMrJ/mN2yuoWyj364=
@@ -325,8 +325,8 @@
 SHA256 (go_modules/github.com/fullstorydev/grpcurl/@v/v1.6.0.zip) = 
iUB9Zs3pEkK2fxGEoMjqsbGIk91IBQswSKRopSEZiyk=
 SHA256 (go_modules/github.com/fzipp/gocyclo/@v/v0.6.0.mod) = 
aeViKYd5ycYD7KmvDNjeKrTt87qUImJUMytvUtA3cVA=
 SHA256 (go_modules/github.com/fzipp/gocyclo/@v/v0.6.0.zip) = 
kdYOuR86MJcR5G1ER4KTpVj+uWV8agQ/C0BEkcivqME=
-SHA256 (go_modules/github.com/gatherstars-com/jwz/@v/v1.3.0.mod) = 
x16PWCRIxs9oZw/hlw4YvOzEpR07kesRHrcAJUJvroM=
-SHA256 (go_modules/github.com/gatherstars-com/jwz/@v/v1.3.0.zip) = 
QP+EyqnY9SMSfwAW5esPECtyIrgUZAVT1CNYR/YgLZQ=
+SHA256 
(go_modules/github.com/gatherstars-com/jwz/@v/v1.3.2-0.20221104050604-3da8c59aef0a.mod)
 = x16PWCRIxs9oZw/hlw4YvOzEpR07kesRHrcAJUJvroM=
+SHA256 
(go_modules/github.com/gatherstars-com/jwz/@v/v1.3.2-0.20221104050604-3da8c59aef0a.zip)
 = ViA6neZgYqIYapuRWCdpUXow3JcllsOfLEHr3GiLWWc=
 SHA256 (go_modules/github.com/gdamore/encoding/@v/v1.0.0.mod) = 
pJgRJVFfDy3yU8LeOjrBlwx0Q+W2adlH6HTnezaBtuU=
 SHA256 (go_modules/github.com/gdamore/encoding/@v/v1.0.0.zip) = 
Y4qYMuL2LRGNfFEdhr2uFiKlHzMd5IoB2Sn9JOvmoqY=
 SHA256 (go_modules/github.com/gdamore/tcell/v2/@v/v2.3.3.mod) = 
rQRC7QKpjBhrhj0Ba0Q4tkLQo0JEsS3neTa6ebHtXbE=
@@ -1902,7 +1902,7 @@
 SHA256 (go_modules/sigs.k8s.io/yaml/@v/v1.1.0.mod) = 
JC1yDRxq7+SjBcM0GoBXXhmlRRelO2IPjOKMKW5R06o=
 SHA256 (go_modules/sigs.k8s.io/yaml/@v/v1.2.0.mod) = 
8ikTVRI4fWZj4fCZmkkiq3CN7UvcgExxzpPj1gK6ld0=
 SHA256 (go_modules/sigs.k8s.io/yaml/@v/v1.2.0.zip) = 
Ve0Ixd9EigM79+LCkS1NqoW4VqBchUsMh8zIXH8/v8c=
-SIZE (aerc-0.13.0.zip) = 430963
+SIZE (aerc-0.13.0pl20221206.zip) = 470925
 SIZE (go_modules/4d63.com/gochecknoglobals/@v/v0.1.0.mod) = 105
 SIZE (go_modules/4d63.com/gochecknoglobals/@v/v0.1.0.zip) = 13354
 SIZE (go_modules/bitbucket.org/creachadair/shell/@v/v0.0.6.mod) = 48
@@ -2229,8 +2229,8 @@
 SIZE (go_modules/github.com/fullstorydev/grpcurl/@v/v1.6.0.zip) = 130685
 SIZE (go_modules/github.com/fzipp/gocyclo/@v/v0.6.0.mod) = 41
 SIZE (go_modules/github.com/fzipp/gocyclo/@v/v0.6.0.zip) = 16339
-SIZE (go_modules/github.com/gatherstars-com/jwz/@v/v1.3.0.mod) = 202
-SIZE (go_modules/github.com/gatherstars-com/jwz/@v/v1.3.0.zip) = 5007390
+SIZE 
(go_modules/github.com/gatherstars-com/jwz/@v/v1.3.2-0.20221104050604-3da8c59aef0a.mod)
 = 202
+SIZE 
(go_modules/github.com/gatherstars-com/jwz/@v/v1.3.2-0.20221104050604-3da8c59aef0a.zip)
 = 5154128
 SIZE (go_modules/github.com/gdamore/encoding/@v/v1.0.0.mod) = 77
 SIZE (go_modules/github.com/gdamore/encoding/@v/v1.0.0.zip) = 19867
 SIZE (go_modules/github.com/gdamore/tcell/v2/@v/v2.3.3.mod) = 308
Index: modules.inc
===
RCS file: /home/cvs/ports/mail/aerc/modules.inc,v
retrieving revision 1.4
diff -u -r1.4 modules.inc
--- modules.inc 21 Oct 2022 11:06:08 -  1.4
+++ modules.inc 6 Dec 2022 20:12:15 -
@@ -1,3 +1,5 @@
+MODGO_VERSION =v0.0.0-20221206102608-c1784b624aea
+
 MODGO_MODULES =\
4d63.com/gochecknoglobalsv0.1.0 \
bitbucket.org/creachadair/shell