CVS: cvs.openbsd.org: ports

[Remi Pointel]

CVSROOT:/cvs
Module name:ports
Changes by: rpoin...@cvs.openbsd.org2018/03/10 22:29:59

Modified files:
devel  : Makefile 

Log message:
+ SUBDIR += py-funcy
+ SUBDIR += py-funcy,python3

CVS: cvs.openbsd.org: ports

[Remi Pointel]

CVSROOT:/cvs
Module name:ports
Changes by: rpoin...@cvs.openbsd.org2018/03/10 22:27:48

Log message:
import funcy, a fancy and practical functional tools.
ok benoit@.

Status:

Vendor Tag: rpointel
Release Tags:   rpointel_20180311

N ports/devel/py-funcy/Makefile
N ports/devel/py-funcy/distinfo
N ports/devel/py-funcy/pkg/PLIST
N ports/devel/py-funcy/pkg/DESCR

No conflicts created by this import

Re: [update] lang/go 1.10

[Joel Sing]

On Saturday 10 March 2018 13:19:06 Stuart Henderson wrote:
> On 2018/03/09 20:08, Jasper Lievisse Adriaanse wrote:
> > On Sat, Mar 10, 2018 at 03:45:36AM +1100, Joel Sing wrote:
> > > - The -x flag was previously on by default, which makes builds
> > > incredibly
> > > 
> > >   noisy. I've turned this off, but enabled it if DEBUG is specified.
> > 
> > I'm not a fan of this part. For many other ports we explicitly instruct
> > the build framework to show us the actual commands being executed, so I'd
> > like it for go to keep doing this.
> 
> I agree, having build command lines in bulk build logs has been
> invaluable in the past and we generally enable them in other ports.
> 
> On the other hand I realise it can make it harder to spot warnings in
> build output when building things by hand and I wouldn't mind having
> it controllable via a mk.conf "user settings" variable. Not necessarily
> specific to Go, this could be used in other places too. Something like
> a "PORTS_QUIETBUILD = Yes" variable perhaps?
> 
> I wouldn't expect it to ever be supported everywhere nor would I want
> to see patches to upstream files just for this, but where there's a
> simple quiet/noisy flag (V=1 or --disable-silent-rules) it could be
> controlled by this.

The motivation behind this was that while updating two of the packages that 
are included in the diff, I lost the actual build/test failure since it had 
disappeared and got lost in the noise (in one case I think it even scrolled off 
my tmux history buffer).

I can understand the benefit of seeing inside the sausage factory, however at 
least in this use case, it is IMO just too noisy (I'm not aware of anyone that 
runs go commands with -x, unless they're debugging) That said, I can keep 
this as a local diff and will drop it from the update to go.port.mk.

Re: [update] rsyslog 8.33.1

[Daniel Jakots]

On Thu, 8 Mar 2018 22:18:01 +0100, Remi Locherer
 wrote:

> +RUN_DEPENDS-elasticsearch =  ${PKGNAME-main}:${PKGPATH},-main

> + --enable-elasticsearch \

You added a subpackage for ES but shouldn't we have a ldep on
elasticsearch as it is with mysql/mariadb?

With that added (unless you think it's wrong), ok danj@

CVS: cvs.openbsd.org: ports

[Brian Callahan]

CVSROOT:/cvs
Module name:ports
Changes by: bcal...@cvs.openbsd.org 2018/03/10 19:16:43

Modified files:
multimedia/streamlink: Makefile distinfo 
multimedia/streamlink/pkg: PLIST 

Log message:
Update to streamlink-0.11.0
Changelog: https://github.com/streamlink/streamlink/releases/tag/0.11.0

CVS: cvs.openbsd.org: ports

[Juan Francisco Cantero Hurtado]

CVSROOT:/cvs
Module name:ports
Changes by: juan...@cvs.openbsd.org 2018/03/10 17:18:12

Modified files:
lang/cython: Makefile 

Log message:
Add setuptools_git as dependency. Spotted by naddy@.

CVS: cvs.openbsd.org: ports

[Frederic Cambus]

CVSROOT:/cvs
Module name:ports
Changes by: fcam...@cvs.openbsd.org 2018/03/10 16:00:33

Modified files:
www/kore   : Makefile distinfo 

Log message:
Switch from GitHub auto-generated tarballs to MASTER_SITES.

Re: UPDATE mail/s-nail

[Steffen Nurpmeso]

Hello Stuart, all.

Stuart Henderson  wrote:
 |On 2018/03/09 18:39, Steffen Nurpmeso wrote:
 |> Stuart Henderson  wrote:
 |>|On 2018/03/08 17:35, Steffen Nurpmeso wrote:
 |>|> This brings in two and a half years of development and bug fixes.
 |>|> It has really improved, though a long road is ahead still.
 |>|> Changelog etc. at: https://www.sdaoden.eu/code-nail-ann.html
 ...
 |>|It picks up idn2 in preference to idn if present at build, so
 |>|WANTLIB/LIB_DEPENDS needs fixing.
 |> 
 |> ..Ok.. hmm.  I must say i dislike the entire IDN stuff very very
 |> much (if 25 years ago the DNS had some extensions to support
 |> longer labels UTF-8 could be it; even globally), so instead i have
 |> changed the configuration to prefer IDN over IDN2, and added the
 |> changeset as a patch for the OpenBSD port.  I really would prefer
 |> this approach and have no IDN2 around, if possible. ?
 |
 |Hm - I haven't been keeping close track of IDN, but don't some
 |DNS registrars (like .de) require IDNA 2008 (i.e. libidn2)?
 |There was a CVE in curl relating to this a couple of years ago.

Yes.  They use this backward incompatible stuff and also seem to
promote DNSSEC (instead of DNS/DTLS and/or DNS/TCP/TLS, imho).
Isn't it absurd that instead of using UTF-7 for internationalized
DNS with an extension bit to support longer names they have chosen
that mysterious algorithm to provide backward compatible somewhat
shorter-than-UTF-7 ASCII representations, and then just break
compatibility with already existing names with a shrug.  No.

Heck, whatever you say.  It is much larger, of course.  It took
two hours to compile here...  And noone uses it. :)
So how about adding an "idna" FLAVOR!?!

(For the future i have implemented an any_of mechanism, so we can
say VAL_IDNA="idn2 idnkit" in the future.
Thanks for pointing this out (it tips the scale in fact,
i wondered just last week how to allow users and the test to
choose the used PRG implementation, and have added non-generic
OPT_SSL_RANDOM, this has to go again).  Credits went to you.)

  ...
 |> Running "make test" does not work because we verify $HOME is an
 |> accessible directory, which the port framework does not provide.
  ...
 |Funnily enough this came up in another list post today.
 |You can either set PORTHOME as a port Makefile variable, or set
 |HOME as an environment variable for test. Using ${WRKDIR} for
 |this is often a good choice.

Ah, ok, i did not find anything in FAQ, and then refrained
from asking.  Now am seeing it in bsd.port.mk(5).
Did the first, ok?

Fine, so with a corrected patch (was not a double free indeed,
kept on wondering), and a patch for that false printf(1) in the
test, IDN2 patch removed and IDN2 dependency added for a new idna
FLAVOR (thus not using the .xz tarball since only IDN2 brings xz;
unfortunately there is no zstd(1) packaging around yet, nor is it
in the OpenBSD base system), all tests work and we end up as
pasted below.

Thanks Stuart for the comments and suggestions, a nice weekend
i wish you, and all,
Ciao already here,

Index: Makefile
===
RCS file: /cvs/ports/mail/s-nail/Makefile,v
retrieving revision 1.19
diff -u -p -r1.19 Makefile
--- Makefile8 Dec 2017 01:04:14 -   1.19
+++ Makefile10 Mar 2018 21:56:02 -
@@ -1,45 +1,51 @@
 # $OpenBSD: Makefile,v 1.19 2017/12/08 01:04:14 espie Exp $
 
-COMMENT=   enhanced, MIME capable Berkeley Mail
+COMMENT=   SysV mail/BSD Mail/POSIX mailx: send and receive mail
 
-DISTNAME=  s-nail-14.8.16
-CATEGORIES=mail
-HOMEPAGE=  https://www.sdaoden.eu/code.html\#s-mailx
-FIX_EXTRACT_PERMISSIONS=Yes
+DISTNAME=  s-nail-14.9.9
+CATEGORIES=mail
+FLAVORS=   idna
+HOMEPAGE=  https://www.sdaoden.eu/code.html\#s-mailx
+MAINTAINER=Dmitrij D. Czarkoff 
 
+MASTER_SITES=  https://ftp.sdaoden.eu/
 # BSD
-PERMIT_PACKAGE_CDROM=  Yes
-
-WANTLIB += c crypto iconv idn ssl
-
-MASTER_SITES=  https://www.sdaoden.eu/downloads/
-MAINTAINER=Dmitrij D. Czarkoff 
+PERMIT_PACKAGE_CDROM=Yes
+# XXX remove FIX_EXTRACT_PERMISSIONS with update to version >14.9.9
+FIX_EXTRACT_PERMISSIONS=Yes
 
-MAKE_FILE= makefile
-MAKE_FLAGS=CC="${CC}" \
-   CFLAGS="${CFLAGS}" \
-   C_INCLUDE_PATH="${LOCALBASE}/include" \
-   LDFLAGS="${LDFLAGS}" \
-   LD_LIBRARY_PATH="${LOCALBASE}/lib" \
-   MANDIR="${PREFIX}/man" \
-   SYSCONFDIR="${SYSCONFDIR}" \
-   WANT_AUTOCC=0 \
-   WANT_EDITLINE=0 \
-   WANT_NCL=1 \
-   WANT_DOTLOCK=0 \
-   SID=
+FLAVOR?=
+WANTLIB=   c crypto curses iconv ssl
+LIB_DEPENDS=   converters/libiconv
+.if ${FLAVOR:Midna}

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2018/03/10 14:58:48

Modified files:
mail/dovecot   : Tag: OPENBSD_6_2 Makefile distinfo 
mail/dovecot/patches: Tag: OPENBSD_6_2 
  patch-src_master_master-settings_c 
mail/dovecot/pkg: Tag: OPENBSD_6_2 PLIST-server dovecot.rc 
mail/dovecot-pigeonhole: Tag: OPENBSD_6_2 Makefile distinfo 
mail/dovecot-pigeonhole/patches: Tag: OPENBSD_6_2 
 
patch-src_managesieve-login_Makefile_in 
Added files:
mail/dovecot/patches: Tag: OPENBSD_6_2 
  
patch-src_lib-master_master-service-ssl-settings_c 
  patch-src_plugins_fts-solr_fts-backend-solr_c 

Log message:
MFC update of Dovecot to 2.2.34, and dovecot-pigeonhole to 0.4.22 to match.

Dovecot security fixes: CVE-2017-15130 CVE-2017-14461 CVE-2017-15132

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2018/03/10 14:54:45

Modified files:
mail/dovecot-pigeonhole: Makefile distinfo 

Log message:
update dovecot-pigeonhole to match Dovecot 2.2.34

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2018/03/10 14:54:30

Modified files:
mail/dovecot   : Makefile distinfo 
mail/dovecot/pkg: PLIST-server 
Added files:
mail/dovecot/patches: 
  
patch-src_lib-master_master-service-ssl-settings_c 
  patch-src_plugins_fts-solr_fts-backend-solr_c 

Log message:
security update to Dovecot 2.2.34. while there, also fix the default TLS
protocol string to avoid using !SSLv2 which is not supported. ok juanfra@ Brad

* CVE-2017-15130: TLS SNI config lookups may lead to excessive memory usage,
causing imap-login/pop3-login VSZ limit to be reached and the process
restarted. This happens only if Dovecot config has local_name { } or local
{ } configuration blocks and attacker uses randomly generated SNI servernames.

* CVE-2017-14461: Parsing invalid email addresses may cause a crash or leak
memory contents to attacker. For example, these memory contents might contain
parts of an email from another user if the same imap process is reused for
multiple users.

* CVE-2017-15132: Aborted SASL authentication leaks memory in login process.

Re: New port: arx-libertatis

[Thomas Frohwein]

Thanks Stuart and Marc - I adjusted the port based on your comments.
pl0 -> pl1, CONFIGURE_ARGS+= -DUSE_QT4=OFF, @cwd is gone

New version attached.

On Mon, Mar 05, 2018 at 08:53:05AM +, Stuart Henderson wrote:
> On 2018/03/04 11:05, Thomas Frohwein wrote:
> > On Sun, Mar 04, 2018 at 05:31:31PM +1100, Jonathan Gray wrote:
> > > 1.1.2-r1 may map to 1.1.2pl1 (1.1.2 patch level 1) if I read
> > > packages-specs(7) correctly as there was a 1.1.2 release prior to r1.
> > 
> > I looked through packages. There are a few packages with pl0 (apvlv,
> > inotify-tools, and xarchiver) and since we also start with v0 and p0, I set 
> > this
> > to arx-libertatis-1.1.2pl0.
> 
> the number after pl usually maps directly to something from upstream,
> I'd go for:
> 
> GH_ACCOUNT =arx
> GH_PROJECT =ArxLibertatis
> GH_TAGNAME =1.1.2-r1
> PKGNAME = arx-libertatis-${GH_TAGNAME:S/-r/pl/}
> 
> > I updated PLIST again because update-desktop-database should really be in
> > there AFAIU. Not sure if the @cwd /usr/local is needed, but it seemed safer 
> > to
> > put it in there. Built it again and ran it without any issues.
> 
> @cwd does nothing here. It's used for rare cases where you have some
> files under one PREFIX and other files under another directory in the same
> package (most common for files installed to /var/www where there's also
> a pkg-readme file).
> 
> : -option(USE_QT4 "Use Qt 4 if available" ON)
> : +option(USE_QT4 "Use Qt 4 if available" OFF)
> 
> CONFIGURE_ARGS= -DUSE_QT4=OFF
> 


arx-libertatis.tgz
Description: Binary data

Re: UPDATE ports/plan9/plan9port MAP_STACK patch

[Aaron Poffenberger]

* Gleydson Soares  [2018-03-10 13:45:08 -0300]:

> On Fri, Mar 09, 2018 at 06:42:25PM -0600, Aaron Poffenberger wrote:
> > This patch updates libthread in plan9port to be MAP_STACK compliant,
> > replacing calls to malloc/free with mmap and munmap when allocating
> > thread stacks.
> > 
> > The most obvious effect  was sam(1), acme(1), and 9term(1) would
> > core when executed. All three run now with seemingly no issues.
> > 
> > There maybe other problems I haven't noticed yet.
> 
> it's missing a revision bump, apart from that, your diff looks good for
> me. thanks for looking into this,

Looks like I implemented a memory leak. I meant to put sizeof *t, but
left it off.

Thanks to Bryan for pinging me on it.

Revised patch follows.

--Aaron

Index: Makefile
===
RCS file: /cvs/ports/plan9/plan9port/Makefile,v
retrieving revision 1.24
diff -u -p -r1.24 Makefile
--- Makefile20 Jan 2018 03:29:56 -  1.24
+++ Makefile10 Mar 2018 16:34:49 -
@@ -6,6 +6,7 @@ BROKEN-powerpc =threading issues
 COMMENT =  Plan 9 from user space
 
 DISTNAME = plan9port-20180117
+REVISION = 0
 
 GH_ACCOUNT =   9fans
 GH_PROJECT =   plan9port
Index: patches/patch-src_libthread_thread_c
===
RCS file: patches/patch-src_libthread_thread_c
diff -N patches/patch-src_libthread_thread_c
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_libthread_thread_c10 Mar 2018 16:34:49 -
@@ -0,0 +1,26 @@
+$OpenBSD$
+
+Index: src/libthread/thread.c
+--- src/libthread/thread.c.orig
 src/libthread/thread.c
+@@ -108,8 +108,9 @@ threadalloc(void (*fn)(void*), void *arg, uint stack)
+   ulong z;
+ 
+   /* allocate the task and stack together */
+-  t = malloc(sizeof *t+stack);
+-  if(t == nil)
++  t = mmap(0, sizeof *t+stack, PROT_READ|PROT_WRITE,
++  MAP_PRIVATE|MAP_ANON|MAP_STACK, -1, 0);
++  if(t == MAP_FAILED)
+   sysfatal("threadalloc malloc: %r");
+   memset(t, 0, sizeof *t);
+   t->stk = (uchar*)(t+1);
+@@ -364,7 +365,7 @@ procscheduler(Proc *p)
+   delthreadinproc(p, t);
+   p->nthread--;
+ /*print("nthread %d\n", p->nthread); */
+-  free(t);
++  munmap(t, sizeof *t+t->stksize);
+   }
+   }
+ 
Index: patches/patch-src_libthread_threadimpl_h
===
RCS file: patches/patch-src_libthread_threadimpl_h
diff -N patches/patch-src_libthread_threadimpl_h
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_libthread_threadimpl_h10 Mar 2018 16:34:49 -
@@ -0,0 +1,14 @@
+$OpenBSD$
+
+Index: src/libthread/threadimpl.h
+--- src/libthread/threadimpl.h.orig
 src/libthread/threadimpl.h
+@@ -10,6 +10,8 @@
+ # define _XOPEN_SOURCE/* for Snow Leopard */
+ # endif
+ # include 
++#elif defined(__OpenBSD__)
++# include 
+ #endif
+ #include 
+ #include "libc.h"

Re: [New] nnn-1.7

[mitchell wodach]

change ncurses to curses in you WANTLIB.

Mitch


On Sat, Mar 10, 2018 at 12:19 PM, Klemens Nanni  wrote:
> On Thu, Mar 08, 2018 at 05:49:51PM +0100, Ljuba Nedeljkovic wrote:
>> Attached is the port of nnn, "The missing terminal file browser for X"
>>
>> Apart from development machine (amd64) it was tested on clean base
>> OpenBSD install also on amd64.  Additional tests/comments/ok would be
>> highly appreciated.
> COMMENT starts with an upper case letter.
>
> As a github project without proper release tarballs, please use GH_*.
> Can you communicate with upstream to provide stable release tarballs?
> They're already packaging for linux, so this shouldn't take much effort.
>
> The license marker is missing.
>
> WRK{DIST,SRC,BUILD} can go once GH_* is set.
>
> PLIST contains an empty line.
>
> Sources provide shell completions, what about packaging them as well?
>
> There is no need for patch-Makefile, you can pass these variables in
> various ways.
>
> For replacing the shebang please see LOCALBASE in bsd.port.mk(5). This
> also implies a missing RDEP on shells/bash.
>
> https://www.openbsd.org/faq/ports/guide.html is helpful in getting these
> fixed, but you can also look at other ports.
>

Re: [New] nnn-1.7

[Klemens Nanni]

On Thu, Mar 08, 2018 at 05:49:51PM +0100, Ljuba Nedeljkovic wrote:
> Attached is the port of nnn, "The missing terminal file browser for X"
> 
> Apart from development machine (amd64) it was tested on clean base
> OpenBSD install also on amd64.  Additional tests/comments/ok would be
> highly appreciated.
COMMENT starts with an upper case letter.

As a github project without proper release tarballs, please use GH_*.
Can you communicate with upstream to provide stable release tarballs?
They're already packaging for linux, so this shouldn't take much effort.

The license marker is missing.

WRK{DIST,SRC,BUILD} can go once GH_* is set.

PLIST contains an empty line.

Sources provide shell completions, what about packaging them as well?

There is no need for patch-Makefile, you can pass these variables in
various ways.

For replacing the shebang please see LOCALBASE in bsd.port.mk(5). This
also implies a missing RDEP on shells/bash.

https://www.openbsd.org/faq/ports/guide.html is helpful in getting these
fixed, but you can also look at other ports.

Re: update SDL2 to 2.0.8 and security update of sdl2-image to 2.0.3

[Thomas Frohwein]

Ping... and double-ping for the sdl2-image security fix

On Thu, Mar 01, 2018 at 07:52:03PM -0800, Thomas Frohwein wrote:
> Hi,
> 
> Here the diff for updating both sdl2 and sdl2-image. Note that sdl2-image is a
> security update for some buffer overflows and crashes by maliciously-crafted
> data (see details below).
> 
> The diff below updates sdl2 to 2.0.8. This includes the optional mapping of 
> the
> Gamecontroller API via the SDL_GAMECONTROLLERCONFIG env var. To recap, there
> are 2 joystick/gamepad APIs in SDL2. The old joystick API was not an issue, 
> but
> the newer gamecontroller API did not work before the 2.0.7 update. The 2.0.7
> updated included a workaround that mapped the gamecontroller API based on the
> fallback part that was already there for Linux, but that was set at compile
> time only. This update includes a slightly expanded workaround that checks for
> the SDL_GAMECONTROLLERCONFIG env var and maps based on that if found.
> 
> Also bump minor. Runs stone-soup, neverball, megaglest, and FNA games as
> before on brief testing, including sound, graphics, and gamepad.
> 
> The sdl2-image update to 2.0.3 was released today and contains a series of
> security patches. Grab maintainer while at it.
> NOTE: the sdl2-image update needs sdl2 version 2.0.8.
> Also tested with above ports without issues.
> 
> The official release notes:
> 
> Thanks to all the people who contributed code and feedback, SDL 2.0.8 is now 
> available!
> http://www.libsdl.org/download-2.0.php
> In addition to lots of bug fixes and build improvements, here are the major 
> changes in this release:
> General:
> Added SDL_fmod() and SDL_log10()
> Each of the SDL math functions now has the corresponding float version
> Added SDL_SetYUVConversionMode() and SDL_GetYUVConversionMode() to control 
> the formula used when converting to and from YUV colorspace. The options are 
> JPEG, BT.601, and BT.709
> Windows:
> Implemented WASAPI support on Windows UWP and removed the deprecated XAudio2 
> implementation
> Added resampling support on WASAPI on Windows 7 and above
> Windows UWP:
> Added SDL_WinRTGetDeviceFamily() to find out what type of device your 
> application is running on
> Mac OS X:
> Added support for the Vulkan SDK for Mac:
> https://www.lunarg.com/lunarg-releases-vulkan-sdk-1-0-69-0-for-mac/
> Added support for OpenGL ES using ANGLE when it's available
> Mac OS X / iOS / tvOS:
> Added a Metal 2D render implementation
> Added SDL_RenderGetMetalLayer() and SDL_RenderGetMetalCommandEncoder() to 
> insert your own drawing into SDL rendering when using the Metal implementation
> iOS:
> Added the hint SDL_HINT_IOS_HIDE_HOME_INDICATOR to control whether the home 
> indicator bar on iPhone X should be hidden. This defaults to dimming the 
> indicator for fullscreen applications and showing the indicator for windowed 
> applications.
> iOS / Android:
> Added the hint SDL_HINT_RETURN_KEY_HIDES_IME to control whether the return 
> key on the software keyboard should hide the keyboard or send a key event 
> (the default)
> Android:
> SDL now supports building with Android Studio and Gradle by default, and the 
> old Ant project is available in android-project-ant
> SDL now requires the API 19 SDK to build, but can still target devices down 
> to API 14 (Android 4.0.1)
> Added SDL_IsAndroidTV() to tell whether the application is running on Android 
> TV
> Android / tvOS:
> Added the hint SDL_HINT_TV_REMOTE_AS_JOYSTICK to control whether TV remotes 
> should be listed as joystick devices (the default) or send keyboard events.
> Linux:
> Added the hint SDL_HINT_VIDEO_X11_NET_WM_BYPASS_COMPOSITOR to control whether 
> the X server should skip the compositor for the SDL application. This 
> defaults to "1"
> Added the hint SDL_HINT_VIDEO_DOUBLE_BUFFER to control whether the Raspberry 
> Pi and KMSDRM video drivers should use double or triple buffering (the 
> default)
> 
> 
> 
> SDL_image 2.0.3 is now available:
> http://www.libsdl.org/projects/SDL_image/
> 
> This is a security update release, fixing the following security reports:
> TALOS-2017-0488
> TALOS-2017-0489
> TALOS-2017-0490
> TALOS-2017-0491
> TALOS-2017-0497
> TALOS-2017-0498
> TALOS-2017-0499
> 
> You'll need to get SDL 2.0.8 for this release of SDL_image.
> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/devel/sdl2/Makefile,v
> retrieving revision 1.20
> diff -u -p -r1.20 Makefile
> --- Makefile  12 Jan 2018 19:32:54 -  1.20
> +++ Makefile  2 Mar 2018 03:50:32 -
> @@ -4,14 +4,13 @@ COMMENT=cross-platform multimedia libra
>  BROKEN-hppa= src/atomic/SDL_spinlock.c:101:2: error: \
>   #error Please implement for your platform.
>  
> -V=   2.0.7
> +V=   2.0.8
>  DISTNAME=SDL2-${V}
>  PKGNAME= sdl2-${V}
>  CATEGORIES=  devel
>  MASTER_SITES=https://www.libsdl.org/release/
> -REVISION=0
>  
> -SHARED_LIBS= SDL20.4  

Re: UPDATE ports/plan9/plan9port MAP_STACK patch

[rob]

> This patch updates libthread in plan9port to be MAP_STACK compliant,
> replacing calls to malloc/free with mmap and munmap when allocating
> thread stacks.

Thanks for this, Aaron! It got those programs from plan9port working
again for me here.

Best,
Rob

Re: UPDATE ports/plan9/plan9port MAP_STACK patch

[Gleydson Soares]

On Fri, Mar 09, 2018 at 06:42:25PM -0600, Aaron Poffenberger wrote:
> This patch updates libthread in plan9port to be MAP_STACK compliant,
> replacing calls to malloc/free with mmap and munmap when allocating
> thread stacks.
> 
> The most obvious effect  was sam(1), acme(1), and 9term(1) would
> core when executed. All three run now with seemingly no issues.
> 
> There maybe other problems I haven't noticed yet.

it's missing a revision bump, apart from that, your diff looks good for
me. thanks for looking into this,
Index: Makefile
===
RCS file: /cvs/ports/plan9/plan9port/Makefile,v
retrieving revision 1.24
diff -u -p -r1.24 Makefile
--- Makefile20 Jan 2018 03:29:56 -  1.24
+++ Makefile10 Mar 2018 16:34:49 -
@@ -6,6 +6,7 @@ BROKEN-powerpc =threading issues
 COMMENT =  Plan 9 from user space
 
 DISTNAME = plan9port-20180117
+REVISION = 0
 
 GH_ACCOUNT =   9fans
 GH_PROJECT =   plan9port
Index: patches/patch-src_libthread_thread_c
===
RCS file: patches/patch-src_libthread_thread_c
diff -N patches/patch-src_libthread_thread_c
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_libthread_thread_c10 Mar 2018 16:34:49 -
@@ -0,0 +1,26 @@
+$OpenBSD$
+
+Index: src/libthread/thread.c
+--- src/libthread/thread.c.orig
 src/libthread/thread.c
+@@ -108,8 +108,9 @@ threadalloc(void (*fn)(void*), void *arg, uint stack)
+   ulong z;
+ 
+   /* allocate the task and stack together */
+-  t = malloc(sizeof *t+stack);
+-  if(t == nil)
++  t = mmap(0, sizeof *t+stack, PROT_READ|PROT_WRITE,
++  MAP_PRIVATE|MAP_ANON|MAP_STACK, -1, 0);
++  if(t == MAP_FAILED)
+   sysfatal("threadalloc malloc: %r");
+   memset(t, 0, sizeof *t);
+   t->stk = (uchar*)(t+1);
+@@ -364,7 +365,7 @@ procscheduler(Proc *p)
+   delthreadinproc(p, t);
+   p->nthread--;
+ /*print("nthread %d\n", p->nthread); */
+-  free(t);
++  munmap(t, t->stksize);
+   }
+   }
+ 
Index: patches/patch-src_libthread_threadimpl_h
===
RCS file: patches/patch-src_libthread_threadimpl_h
diff -N patches/patch-src_libthread_threadimpl_h
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_libthread_threadimpl_h10 Mar 2018 16:34:49 -
@@ -0,0 +1,14 @@
+$OpenBSD$
+
+Index: src/libthread/threadimpl.h
+--- src/libthread/threadimpl.h.orig
 src/libthread/threadimpl.h
+@@ -10,6 +10,8 @@
+ # define _XOPEN_SOURCE/* for Snow Leopard */
+ # endif
+ # include 
++#elif defined(__OpenBSD__)
++# include 
+ #endif
+ #include 
+ #include "libc.h"

Re: [New] nnn-1.7

[mitchell wodach]

I get this error when I run "make test"

openbsdcurrent# make test
===>  Regression tests for nnn-1.7
gmake: *** No rule to make target 'test'.  Stop.
*** Error 2 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2777
'/usr/ports/pobj/nnn-1.7/.test_done')
*** Error 1 in /usr/ports/myshit/sysutils/nnn
(/usr/ports/infrastructure/mk/bsd.port.mk:2419 'test')

Try adding NO_TEST=Yes. to your Makefile

On Thu, Mar 8, 2018 at 10:49 AM, Ljuba Nedeljkovic  wrote:
> Hello,
>
> Attached is the port of nnn, "The missing terminal file browser for X"
>
> Apart from development machine (amd64) it was tested on clean base
> OpenBSD install also on amd64.  Additional tests/comments/ok would be
> highly appreciated.
>
> Cheers,
> -ljuba

CVS: cvs.openbsd.org: ports

[Remi Pointel]

CVSROOT:/cvs
Module name:ports
Changes by: rpoin...@cvs.openbsd.org2018/03/10 07:46:40

Modified files:
devel  : Makefile 

Log message:
+ SUBDIR += jadx

CVS: cvs.openbsd.org: ports

[Remi Pointel]

CVSROOT:/cvs
Module name:ports
Changes by: rpoin...@cvs.openbsd.org2018/03/10 07:45:43

Log message:
import jadx: a dex to java decompiler.
ok benoit@.

Status:

Vendor Tag: rpointel
Release Tags:   rpointel_20180310

N ports/devel/jadx/Makefile
N ports/devel/jadx/distinfo
N ports/devel/jadx/patches/patch-bin_jadx
N ports/devel/jadx/pkg/DESCR
N ports/devel/jadx/pkg/PLIST

No conflicts created by this import

UPDATE sysutils/ddrescue

[Björn Ketelaars]

Enclosed a diff for bringing ddrescue to 1.23. Changelog can be found at
https://fossies.org/linux/privat/ddrescue-1.23.tar.lz/ddrescue-1.23/ChangeLog

Make test completes successfully. No issues found while reading small
USB-stick to an image.

OK?


Index: Makefile
===
RCS file: /cvs/ports/sysutils/ddrescue/Makefile,v
retrieving revision 1.17
diff -u -p -r1.17 Makefile
--- Makefile26 Jul 2017 22:45:30 -  1.17
+++ Makefile10 Mar 2018 13:41:46 -
@@ -2,8 +2,7 @@
 
 COMMENT =  data recovery tool
 
-DISTNAME = ddrescue-1.22
-REVISION = 1
+DISTNAME = ddrescue-1.23
 
 CATEGORIES =   sysutils
 
Index: distinfo
===
RCS file: /cvs/ports/sysutils/ddrescue/distinfo,v
retrieving revision 1.8
diff -u -p -r1.8 distinfo
--- distinfo30 Apr 2017 14:36:00 -  1.8
+++ distinfo10 Mar 2018 13:41:46 -
@@ -1,2 +1,2 @@
-SHA256 (ddrescue-1.22.tar.lz) = CYV7LoB0gTrBnaXSYokPci5feQDlIaTGA1TO+V7qEKc=
-SIZE (ddrescue-1.22.tar.lz) = 80050
+SHA256 (ddrescue-1.23.tar.lz) = qa4t1EWSvzhsnBVqXayu65AVc8mGeto2CPiH1AEzjY0=
+SIZE (ddrescue-1.23.tar.lz) = 82451

CVS: cvs.openbsd.org: ports

[Bjorn Ketelaars]

CVSROOT:/cvs
Module name:ports
Changes by: b...@cvs.openbsd.org2018/03/10 06:27:36

Modified files:
net/librsync   : Makefile distinfo 
Removed files:
net/librsync/patches: patch-tests_hashtable_test_c 

Log message:
Update to librsync-2.0.2.

OK sthen@

Re: [www/rt] A flavor for LDAP?

[Stuart Henderson]

On 2018/03/08 13:08, William Leuschner wrote:
> Here's a patch that adds a short message to the README explaining how to
> add LDAP support to RT.  Also, when I tested the packaging, I noticed
> that the configure script was being given some GNU-specific arguments
> that it didn't understand, so I changed the configure script style from
> GNU to simple in addition to bumping the revision.

This shouldn't be CONFIGURE_STYLE=simple - CONFIGURE_STYLE=gnu is
correct already.

> diff -u -p -r1.40 Makefile
> --- Makefile  11 Jan 2018 19:27:12 -  1.40
> +++ Makefile  8 Mar 2018 17:50:50 -
> @@ -4,7 +4,7 @@ COMMENT=  industrial-grade ticketing sys
> 
>  DISTNAME=rt-4.4.2
>  CATEGORIES=  www
> -REVISION=0
> +REVISION=1
> 
>  MAINTAINER=  Robert Nagy 
> 
> @@ -107,7 +107,7 @@ NO_BUILD= Yes
> 
>  PKG_ARCH=*
> 
> -CONFIGURE_STYLE= gnu
> +CONFIGURE_STYLE= simple
> 
>  CONFIGURE_ENV+=  LDFLAGS="-L${LOCALBASE}/lib"
> 
> Index: pkg/README
> ===
> RCS file: /cvs/ports/www/rt/pkg/README,v
> retrieving revision 1.1
> diff -u -p -r1.1 README
> --- pkg/README14 Jun 2011 13:19:33 -  1.1
> +++ pkg/README8 Mar 2018 17:50:50 -
> @@ -20,3 +20,8 @@ ${PREFIX}/share/doc/rt3/README
>  If you are upgrading your installation, don't forget to run:
>  # rt-setup-database --dba  --prompt-for-dba-password \
>   --action upgrade
> +
> +If you plan to connect your rt installation to an LDAP domain,
> +install p5-ldap and then run
> +
> +# rt-test-dependencies

Is it actually required to run rt-test-dependencies? I haven't used rt
for some years but from what I remember it's just a checker and doesn't
make any actual changes?

Re: stack-register checking issue in Emacs?

[Stuart Henderson]

On 2018/03/10 14:34, Timo Myyrä wrote:
> Hi,
> 
> I noticed that I get following lines in dmesg when I start emacs:
> syscall: sp c420028730 not inside 7f7fff78a000-7f7da000
> 
> Seems the above doesn't trigger on clean config and tracked it so far to
> initializing go oracle:
> 
> (let ((oracle-file (concat-path (getenv "GOPATH") 
> "src/code.google.com/p/go.tools/cmd/oracle/oracle.el")))
>   (when (file-exists-p oracle-file)
> (load oracle-file)
> (setq go-oracle-command (concat (getenv "GOPATH") "/bin/oracle"
> 
> When the above part of my emacs init file is commented the dmesg line doesn't
> appear. Don't know how to proceed with this though.
> 
> Timo
> 

This is go-related rather than emacs-related, go has not been made to
work with stack checking yet.

Re: [update] lang/go 1.10

[Stuart Henderson]

On 2018/03/09 20:08, Jasper Lievisse Adriaanse wrote:
> On Sat, Mar 10, 2018 at 03:45:36AM +1100, Joel Sing wrote:
> > 
> > - The -x flag was previously on by default, which makes builds incredibly
> >   noisy. I've turned this off, but enabled it if DEBUG is specified.
> I'm not a fan of this part. For many other ports we explicitly instruct the 
> build
> framework to show us the actual commands being executed, so I'd like it for go
> to keep doing this.

I agree, having build command lines in bulk build logs has been
invaluable in the past and we generally enable them in other ports.

On the other hand I realise it can make it harder to spot warnings in
build output when building things by hand and I wouldn't mind having
it controllable via a mk.conf "user settings" variable. Not necessarily
specific to Go, this could be used in other places too. Something like
a "PORTS_QUIETBUILD = Yes" variable perhaps?

I wouldn't expect it to ever be supported everywhere nor would I want
to see patches to upstream files just for this, but where there's a
simple quiet/noisy flag (V=1 or --disable-silent-rules) it could be
controlled by this.

CVS: cvs.openbsd.org: ports

[Jasper Lievisse Adriaanse]

CVSROOT:/cvs
Module name:ports
Changes by: jas...@cvs.openbsd.org  2018/03/10 05:50:18

Modified files:
devel/quirks   : Makefile 
devel/quirks/files: Quirks.pm 

Log message:
register xerces removal

CVS: cvs.openbsd.org: ports

[Jasper Lievisse Adriaanse]

CVSROOT:/cvs
Module name:ports
Changes by: jas...@cvs.openbsd.org  2018/03/10 05:49:14

Removed files:
textproc/xerces: Makefile distinfo 
textproc/xerces/pkg: DESCR PLIST 

Log message:
remove xerces; an unused and dormant/dead xml parser is never a good idea to 
keep around

ok aja@

CVS: cvs.openbsd.org: ports

[Jasper Lievisse Adriaanse]

CVSROOT:/cvs
Module name:ports
Changes by: jas...@cvs.openbsd.org  2018/03/10 05:48:11

Modified files:
textproc   : Makefile 

Log message:
-xerces

stack-register checking issue in Emacs?

[Timo Myyrä]

Hi,

I noticed that I get following lines in dmesg when I start emacs:
syscall: sp c420028730 not inside 7f7fff78a000-7f7da000

Seems the above doesn't trigger on clean config and tracked it so far to
initializing go oracle:

(let ((oracle-file (concat-path (getenv "GOPATH") 
"src/code.google.com/p/go.tools/cmd/oracle/oracle.el")))
  (when (file-exists-p oracle-file)
(load oracle-file)
(setq go-oracle-command (concat (getenv "GOPATH") "/bin/oracle"

When the above part of my emacs init file is commented the dmesg line doesn't
appear. Don't know how to proceed with this though.

Timo

UPDATE net/librsync

[Björn Ketelaars]

Enclosed a diff for bringing librsync to 2.0.2, which is a minor point
release that includes mainly code cleanups. As a result there is one
patch less to maintain.

Instead of using MASTER_SITES Use GH_* (as explained in
Makefile.template). While here update HOMEPAGE.

Output 'make test':

100% tests passed, 0 tests failed out of 12

Build tested all consumers without any issue.

OK?


Index: Makefile
===
RCS file: /cvs/ports/net/librsync/Makefile,v
retrieving revision 1.19
diff -u -p -r1.19 Makefile
--- Makefile8 Feb 2018 22:35:19 -   1.19
+++ Makefile10 Mar 2018 11:59:51 -
@@ -2,22 +2,21 @@
 
 COMMENT=   library for delta compression of streams
 
-V= 2.0.1
-DISTNAME=  librsync-$V
-DISTFILES= ${DISTNAME}{v$V}.tar.gz
+# XXX: https://github.com/librsync/librsync/issues/146
+GH_ACCOUNT=librsync
+GH_PROJECT=librsync
+GH_TAGNAME=v2.0.2
 
 SHARED_LIBS += rsync4.0  # 2.0
 
 CATEGORIES=net
 
-HOMEPAGE=  https://github.com/librsync/librsync
+HOMEPAGE=  https://librsync.github.io/
 
 # LGPLv2.1+
 PERMIT_PACKAGE_CDROM=  Yes
 
 WANTLIB += b2 c popt
-
-MASTER_SITES=  https://github.com/librsync/librsync/archive/
 
 MODULES=   devel/cmake
 
Index: distinfo
===
RCS file: /cvs/ports/net/librsync/distinfo,v
retrieving revision 1.5
diff -u -p -r1.5 distinfo
--- distinfo8 Feb 2018 22:35:19 -   1.5
+++ distinfo10 Mar 2018 11:59:51 -
@@ -1,2 +1,2 @@
-SHA256 (librsync-2.0.1.tar.gz) = b85pBBqk/HKiHxqygKcpm4LfKx+gol2Glf1SfmdSYl4=
-SIZE (librsync-2.0.1.tar.gz) = 169765
+SHA256 (librsync-2.0.2.tar.gz) = 5nuVIO6E9yOb5ulIeVgDvZVJUJHMAL9tDoxpdgMqSvE=
+SIZE (librsync-2.0.2.tar.gz) = 174480
Index: patches/patch-tests_hashtable_test_c
===
RCS file: patches/patch-tests_hashtable_test_c
diff -N patches/patch-tests_hashtable_test_c
--- patches/patch-tests_hashtable_test_c8 Feb 2018 22:35:19 -   
1.1
+++ /dev/null   1 Jan 1970 00:00:00 -
@@ -1,13 +0,0 @@
-$OpenBSD: patch-tests_hashtable_test_c,v 1.1 2018/02/08 22:35:19 sthen Exp $
-
-Index: tests/hashtable_test.c
 tests/hashtable_test.c.orig
-+++ tests/hashtable_test.c
-@@ -27,7 +27,6 @@
- #include "hashtable.h"
- 
- /* Key type for the hashtable. */
--typedef int key_t;
- void key_init(key_t *k, int i)
- {
- /* This is chosen to cause bad key collisions and clustering. */

UPDATE: telephony/baresip/re

[Ingo Feinerer]

Hi,

update telephony/baresip/re 0.5.5 -> 0.5.7

A few functions were removed and added so I bumped SHARED_LIBS:

$ /usr/src/lib/check_sym libre.so.4.0 libre.so.5.0
libre.so.4.0 --> libre.so.5.0
Dynamic export changes:
added:
dtls_recv_packet
http_chunk_decode
http_req_set_conn_handler
icem_cand_type
icem_selected_lcand
icem_selected_rcand
mem_seccmp

removed:
http_req_tcp
http_req_tls

PLT added:
http_chunk_decode
http_req_set_conn_handler
icem_lcand_addr

PLT removed:
http_req_tcp
http_req_tls

baresip still works fine with this update for me on amd64.

I could not reach czarkoff@ (MAINTAINER) so I am asking on ports@ for OKs.

Best regards,
Ingo

Index: Makefile
===
RCS file: /cvs/ports/telephony/baresip/re/Makefile,v
retrieving revision 1.17
diff -u -p -r1.17 Makefile
--- Makefile22 Sep 2017 07:25:26 -  1.17
+++ Makefile10 Mar 2018 11:43:59 -
@@ -2,9 +2,9 @@
 
 COMMENT =  portable library for real-time communications
 
-V =0.5.5
+V =0.5.7
 DISTNAME = re-${V}
-SHARED_LIBS =  re  4.0
+SHARED_LIBS =  re  5.0
 CATEGORIES +=  devel
 
 WANTLIB = crypto pthread ssl z
Index: distinfo
===
RCS file: /cvs/ports/telephony/baresip/re/distinfo,v
retrieving revision 1.13
diff -u -p -r1.13 distinfo
--- distinfo22 Sep 2017 07:25:26 -  1.13
+++ distinfo10 Mar 2018 11:43:59 -
@@ -1,2 +1,2 @@
-SHA256 (re-0.5.5.tar.gz) = kJF6Fz3pYtOyCrX5h1rTBRt7MH2krLgMGEty5sK6e7Q=
-SIZE (re-0.5.5.tar.gz) = 299147
+SHA256 (re-0.5.7.tar.gz) = XcwVYQwo/x3xR9KCZrKbk0rc/0O/w/2sWHZ/14kQEDk=
+SIZE (re-0.5.7.tar.gz) = 299678
Index: patches/patch-Makefile
===
RCS file: /cvs/ports/telephony/baresip/re/patches/patch-Makefile,v
retrieving revision 1.7
diff -u -p -r1.7 patch-Makefile
--- patches/patch-Makefile  22 Sep 2017 07:25:26 -  1.7
+++ patches/patch-Makefile  10 Mar 2018 11:43:59 -
@@ -2,9 +2,9 @@ $OpenBSD: patch-Makefile,v 1.7 2017/09/2
 Index: Makefile
 --- Makefile.orig
 +++ Makefile
-@@ -12,10 +12,6 @@ VER_PATCH := 5
+@@ -12,10 +12,6 @@ VER_PATCH := 7
  PROJECT   := re
- VERSION   := 0.5.5
+ VERSION   := 0.5.7
  
 -MK:= mk/re.mk
 -

Re: [New] pkgconf 1.4.2

[Adam Steen]

On March 10, 2018 10:30 AM, Stuart Henderson  wrote:

> ​​
> 
> On 2018/03/09 22:00, Adam Steen wrote:
> 
> > On March 7, 2018 10:49 PM, Klemens Nanni k...@openbsd.org wrote:
> > 
> > > On Tue, Mar 06, 2018 at 09:16:11PM -0500, Adam Steen wrote:
> > > 
> > > > On February 28, 2018 12:27 PM, Adam Steen a...@adamsteen.com.au wrote:
> > > > 
> > > > > On February 28, 2018 12:26 PM, Adam Steen a...@adamsteen.com.au wrote:
> > > > > 
> > > > > > Here is a new port "devel/pkgconf", it is needed are part of my 
> > > > > > long term
> > > > > > 
> > > > > > goal to port MirageOS/Solo5 to OpenBSD/vmm.
> > > 
> > > Glad to hear that, thanks for your efforts!
> > > 
> > > > I would like to get this added, is there anything more I need to do?
> > > 
> > > Please use a single space before and tabs after = in Makefile.
> > > 
> > > VERSION can be zapped and put into DISTNAME directly.
> > > 
> > > The port looks good to me so far except for tests, you're currently
> > > 
> > > missing devel/kyua-cli as TDEP. With it, the following error occurs:
> > > 
> > > kyua --config=none test --kyuafile='./Kyuafile' --build-root='.'
> > > 
> > > kyua: E: Failed to create directory /pkgconf-1.4.2\_writes\_to_HOME: 
> > > Permission denied.
> > 
> > Hi Klemens
> > 
> > I am new to ports, I have completed the fixes you suggested and now am 
> > working on ensuring the tests complete.
> > 
> > I take it the ports infrastructure prevents ports from writing to $HOME, 
> > and this where the error is coming from. To fix the problem do I need to 
> > configure the tests to write to, somewhere ? /tmp? instead? where is the 
> > preferred location? or am I miss understanding something?
> 
> Easiest is probably to set the PORTHOME make variable, something
> 
> like "PORTHOME= ${WRKDIR}" should do.
> 
> Ports infrastructure sets a distinctive HOME by default so that ports
> 
> writing there are more easily identified. This used to be more of a
> 
> problem when the "fake-install" stage was run as root, but nowadays
> 
> the build is more likely to fail due to lack of permissions rather
> 
> than quietly run and create this /xxx\_writes\_to_HOME directory.

Hi Stuart

That worked a treat, i will run through the setup again and make sure things 
work from "wo to go" and then come back with a complete port.

Cheers
Adam

Re: UPDATE mail/s-nail

[Stuart Henderson]

On 2018/03/09 18:39, Steffen Nurpmeso wrote:
> Hello Stuart!
> 
> Thanks for looking into this.
> 
> Stuart Henderson  wrote:
>  |On 2018/03/08 17:35, Steffen Nurpmeso wrote:
>  |> This brings in two and a half years of development and bug fixes.
>  |> It has really improved, though a long road is ahead still.
>  |> Changelog etc. at: https://www.sdaoden.eu/code-nail-ann.html
>  ...
>  |> +FIX_EXTRACT_PERMISSIONS=Yes
>  |
>  |You're upstream aren't you - is there a reason for the unusual
>  |permissions in the tar file?
> 
> In a future release i will thus create the balls with a different
> umask(1) so this should then no longer be necessary.
> I gave the credit William (finally) and you (thanks).
> 
>  |It picks up idn2 in preference to idn if present at build, so
>  |WANTLIB/LIB_DEPENDS needs fixing.
> 
> ..Ok.. hmm.  I must say i dislike the entire IDN stuff very very
> much (if 25 years ago the DNS had some extensions to support
> longer labels UTF-8 could be it; even globally), so instead i have
> changed the configuration to prefer IDN over IDN2, and added the
> changeset as a patch for the OpenBSD port.  I really would prefer
> this approach and have no IDN2 around, if possible. ?

Hm - I haven't been keeping close track of IDN, but don't some
DNS registrars (like .de) require IDNA 2008 (i.e. libidn2)?
There was a CVE in curl relating to this a couple of years ago.


>  |$ make port-lib-depends-check
>  |s-nail-14.9.9(mail/s-nail):
>  |Missing lib: idn2.1 (/usr/local/bin/nail) (NOT REACHABLE)
>  |Extra:  idn.17
> 
> Should work now.
> 
> Also i add a fix for a double-free that sneaked into the codebase,
> unfortunately.  We have builtin double-free detection, terrible ...
> 
> Running "make test" does not work because we verify $HOME is an
> accessible directory, which the port framework does not provide.
> (The message is "nail: $HOME is not a directory or not accessible:
> /s-nail-14.9.9_writes_to_HOME", but we do not write, we only
> verify because $HOME is indeed target for several standardized
> files which we possibly would need to address.)  Due to that the
> error messages invalidate the test checksums.

Funnily enough this came up in another list post today.
You can either set PORTHOME as a port Makefile variable, or set
HOME as an environment variable for test. Using ${WRKDIR} for
this is often a good choice.

Re: [New] pkgconf 1.4.2

[Stuart Henderson]

On 2018/03/09 22:00, Adam Steen wrote:
> On March 7, 2018 10:49 PM, Klemens Nanni  wrote:
> 
> > ​​
> > 
> > On Tue, Mar 06, 2018 at 09:16:11PM -0500, Adam Steen wrote:
> > 
> > > On February 28, 2018 12:27 PM, Adam Steen a...@adamsteen.com.au wrote:
> > > 
> > > > On February 28, 2018 12:26 PM, Adam Steen a...@adamsteen.com.au wrote:
> > > > 
> > > > > Here is a new port "devel/pkgconf", it is needed are part of my long 
> > > > > term
> > > > > 
> > > > > goal to port MirageOS/Solo5 to OpenBSD/vmm.
> > 
> > Glad to hear that, thanks for your efforts!
> > 
> > > I would like to get this added, is there anything more I need to do?
> > 
> > Please use a single space before and tabs after = in Makefile.
> > 
> > VERSION can be zapped and put into DISTNAME directly.
> > 
> > The port looks good to me so far except for tests, you're currently
> > 
> > missing devel/kyua-cli as TDEP. With it, the following error occurs:
> > 
> > kyua --config=none test --kyuafile='./Kyuafile' --build-root='.'
> > 
> > kyua: E: Failed to create directory /pkgconf-1.4.2\_writes\_to_HOME: 
> > Permission denied.
> 
> Hi Klemens
> 
> I am new to ports, I have completed the fixes you suggested and now am 
> working on ensuring the tests complete.
> 
> I take it the ports infrastructure prevents ports from writing to $HOME, and 
> this where the error is coming from. To fix the problem do I need to 
> configure the tests to write to, somewhere ? /tmp? instead? where is the 
> preferred location? or am I miss understanding something?

Easiest is probably to set the PORTHOME make variable, something
like "PORTHOME= ${WRKDIR}" should do.

Ports infrastructure sets a distinctive HOME by default so that ports
writing there are more easily identified. This used to be more of a
problem when the "fake-install" stage was run as root, but nowadays
the build is more likely to fail due to lack of permissions rather
than quietly run and create this /xxx_writes_to_HOME directory.