Re: bbpager: bus error

[Alessandro DE LAURENZIS]

On Fri 19/02/2016 21:43, Jérémie Courrèges-Anglas wrote:
> 
> Thanks.  The problem is that the iterator is invalidated by erase(); the
> diff ensures that after erase() returns, pit points at the previous item
> in the list.
> 
> oks?
> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/x11/bbpager/Makefile,v
> retrieving revision 1.18
> diff -u -p -r1.18 Makefile
> --- Makefile  28 Oct 2014 08:18:30 -  1.18
> +++ Makefile  19 Feb 2016 20:42:34 -
> @@ -3,7 +3,7 @@
>  COMMENT =pager for the Blackbox window manager
>  
>  DISTNAME =   bbpager-0.4.7
> -REVISION=1
> +REVISION=2
>  
>  CATEGORIES = x11
>  MASTER_SITES =   ${MASTER_SITE_SOURCEFORGE:=bbtools/}
> Index: patches/patch-src_wminterface_cxx
> ===
> RCS file: patches/patch-src_wminterface_cxx
> diff -N patches/patch-src_wminterface_cxx
> --- /dev/null 1 Jan 1970 00:00:00 -
> +++ patches/patch-src_wminterface_cxx 19 Feb 2016 20:42:34 -
> @@ -0,0 +1,13 @@
> +$OpenBSD$
> +--- src/wminterface.cxx.orig Fri Feb 19 19:14:47 2016
>  src/wminterface.cxx  Fri Feb 19 19:15:04 2016
> +@@ -84,8 +84,7 @@ void WMInterface::updateWindowList(void)
> + for (; pit != pit_end; pit++) {
> + if (!(*pit)->isMarked()) {
> + delete (*pit);
> +-bbtool->pagerWindowList().erase(pit);
> +-pit--;
> ++bbtool->pagerWindowList().erase(pit--);
> + } else {
> + }
> + }
> 
> 

Confirmed: this patch solves the problem on -current, amd64.

Thanks a lot

-- 
Alessandro DE LAURENZIS
[mailto:just22@gmail.com]
LinkedIn: http://it.linkedin.com/in/delaurenzis

NEW: TeXmacs 1.99.4

[Ray Lai]

Hi,

GNU TeXmacs is a free wysiwyw (what you see is what you want) editing 
platform with special features for scientists. The software aims to 
provide a unified and user friendly framework for editing structured 
documents with different types of content (text, graphics, mathematics, 
interactive content, etc.). The rendering engine uses high-quality 
typesetting algorithms so as to produce professionally looking documents, 
which can either be printed out or presented from a laptop.

The software includes a text editor with support for mathematical 
formulas, a small technical picture editor and a tool for making 
presentations from a laptop. Moreover, TeXmacs can be used as an interface 
for many external systems for computer algebra, numerical analysis, 
statistics, etc. New presentation styles can be written by the user and 
new features can be added to the editor using the Scheme extension 
language. A native spreadsheet and tools for collaborative authoring are 
planned for later. 

Please let me know if there's anything to improve.

Ray

TeXmacs-1.99.4.tgz
Description: Binary data

truncated pointer on library return (wip net/transmission graphics/gdk-pixbuf)

[Josh Grosse]

I could use a cluestick from those with greater skills and knowledge.

I've been working on an updated port of net/transmission, preparing 
for its next release, which I expect from upstream, eventually.

I've got an odd problem with the GTK version of the app, which 
gets a segfault because a 64-bit pointer gets truncated to 32-bits
on return from gdk_pixbuf_new_from_inline, a function within the
graphics/gdk-pixbuf library.

In the library, I can see a perfectly fine 64-bit GdkPixBuf pointer 
pointing to a valid structure right before the return to the application.
But when the application assigns the variable out of the function call,
the value has been truncated, so I get segfaults.  The application is 
multithreaded, so both gdb and egdb often fail to let me step through
while running without additional aborts, but I have had some success 
obtaining data from the core files, combined with the addition of 
printf() functions in the library and the application itself to aid
with my understanding of the problem so far.

I've built both the application and library with -O0, so there 
shouldn't be any odd optimizations involved.  

The WIP port is on Jasper's openbsd-wip repository, where I've 
summarized the problem in net/transmission/TODO:

---

1. break/fix: transmission-gtk segfaults in gdk-pixbuf2 on amd64

   The segfault has been traced to a truncated pointer returning from
   gdk_pixbuf_new_from_inline, a library function of graphics/gdk-pixbuf.
   The call is from gtk/actions.c.  The problem occurs on amd64, but
   not on i386 (of course).

   The top 32-bits of the returned value will be 0x or 0x.
   It's fine in the library, and truncated upon return to the application.
   
   This is from debugging with fprintf()s sprinkled through the mentioned
   source module and the gdk-pixbuf library.  gdb/egdb isn't very helpful,
   as the multithreaded operation of the application precludes clean step
   through.  

---

Thank you in advance for any recommendations -- diagnostic, corrective, or
ball peen hammer.  I know that this particular Gdk function has been
deprecated, but the upstream application still requires it.
gdk function has been deprecated, but upstream is still using it.  

Re: databases/pkglocatedb: locate(1) doesn't have -m/-s flag

[Michael Reed]

On 02/18/16 15:53, Marc Espie wrote:

On Thu, Feb 18, 2016 at 02:44:58PM -0500, Michael Reed wrote:

locate(1)'s `m' and `s' flags were deleted a while ago, but pkg_locate's man
page wasn't updated.

I read [1] and thought that this change, given it's only documentation, would
be okay to commit, but please correct me if I'm wrong

[1]: https://marc.info/?l=openbsd-ports=145460329517278=2


Since it's an always-update package, practically, this will work.
But this kind of change still feels better with a version change
(e.g., bump to 1.2)


Sure, hows this?




Index: Makefile
===
RCS file: /cvs/ports/databases/pkglocatedb/Makefile,v
retrieving revision 1.29
diff -u -p -r1.29 Makefile
--- Makefile28 Apr 2015 17:03:00 -  1.29
+++ Makefile20 Feb 2016 03:11:43 -
@@ -4,7 +4,7 @@ COMMENT =   database of packages for use
 
 # locate data is NOT MI, unfortunately

 # PKG_ARCH = ${ARCH}
-DISTNAME = pkglocatedb-1.1
+DISTNAME = pkglocatedb-1.2
 DISTFILES =
 
 CATEGORIES =		databases

Index: files/pkg_locate.1
===
RCS file: /cvs/ports/databases/pkglocatedb/files/pkg_locate.1,v
retrieving revision 1.2
diff -u -p -r1.2 pkg_locate.1
--- files/pkg_locate.1  21 Oct 2013 09:04:47 -  1.2
+++ files/pkg_locate.1  20 Feb 2016 03:11:43 -
@@ -23,7 +23,7 @@
 .Sh SYNOPSIS
 .Nm pkg_locate
 .Bk -words
-.Op Fl bcimSs
+.Op Fl bciS
 .Op Fl l Ar limit
 .Ar pattern ...
 .Ek
@@ -95,21 +95,8 @@ Suppress normal output; instead print a
 Ignore case distinctions in both the pattern and the database.
 .It Fl l Ar limit
 Limit output to a specific number of files and exit.
-.It Fl m
-Use
-.Xr mmap 2
-instead of the
-.Xr stdio 3
-library.
-This is the default behavior.
-It performs better in most cases.
 .It Fl S
 Print some statistics about the database and exit.
-.It Fl s
-Use the
-.Xr stdio 3
-library instead of
-.Xr mmap 2 .
 .El
 .Sh FILES
 .Bl -tag -compact -width /usr/local/share/pkglocatedb
Index: files/pkglocate
===
RCS file: /cvs/ports/databases/pkglocatedb/files/pkglocate,v
retrieving revision 1.2
diff -u -p -r1.2 pkglocate
--- files/pkglocate 10 Oct 2012 09:49:20 -  1.2
+++ files/pkglocate 20 Feb 2016 03:11:43 -
@@ -1,7 +1,7 @@
 #! /bin/sh
 
 if [ $# -eq 0 ]; then

-   echo "usage: ${0##*/} [-bcimSs] [-l limit] pattern ..."
+   echo "usage: ${0##*/} [-bciS] [-l limit] pattern ..."
exit 1
 fi
 

Fwd: Re: filename too long in /usr/ports ?

[Nigel Taylor]

 Forwarded Message 
Subject: Re: filename too long in /usr/ports ?
Date: Fri, 19 Feb 2016 23:11:39 +
From: Nigel Taylor 
To: Alexandre H 

On 02/19/16 22:27, Alexandre H wrote:
> A file is rejected by tar :
> 
> # cd /usr
> # tar zcf ports.tar.gz ports/
> tar: File name too long for ustar
> ports/sysutils/logstash/logstash/patches/patch-vendor_bundle_jruby_1_9_gems_logstash-core-1_5_3-java_lib_logstash_patches_stronger_openssl_defaults_rb
> 
> 
> 
Your ports tree is out of date. That's in the Attic.


date: 2015/04/26 19:02:57;  author: jasper;  state: Exp;  commitid:
Cp83CPsFNJPN6K2y;
- update to logstash-1.5.0-rc3 (from beta1)
=

RCS file:
/home/cvs/ports/sysutils/logstash/logstash/patches/Attic/patch-vendor_bundle_jruby_1_9_gems_logstash-core-1_5_3-java_lib_logstash_patches_stronger_openssl_defaults_rb,v
Working file:
patches/patch-vendor_bundle_jruby_1_9_gems_logstash-core-1_5_3-java_lib_logstash_patches_stronger_openssl_defaults_rb
head: 1.2
branch:
locks: strict
access list:
symbolic names:
OPENBSD_5_8: 1.1.0.4
OPENBSD_5_8_BASE: 1.1
keyword substitution: kv
total revisions: 2; selected revisions: 2
description:

revision 1.2
date: 2015/09/07 08:09:22;  author: jasper;  state: dead;  lines: +1 -1;
 commitid: qW472YPgXs4LCooX;
should've been removed in previous

Re: alpha-1.ports.openbsd.org bulk build report

[Sebastian Reitenbach]

On 02/19/16 08:18, Sebastian Reitenbach wrote:

Hi,

Am 2/18/2016 um 9:15 PM schrieb Landry Breuil:

On Wed, Feb 17, 2016 at 10:08:40PM +0100, Matthieu Herrb wrote:

On Wed, Feb 17, 2016 at 09:46:30PM +0100, Matthieu Herrb wrote:

On Mon, Feb 15, 2016 at 03:44:02PM +, Christian Weisgerber wrote:

On 2016-02-14, Matthieu Herrb  wrote:


I've taken the list of individual optimisation from gcc(1) for both -O1
and -O2 and replaced -O2 by this in CFLAGS.

gcc-local(1) is also worth a look...


+O2= ${O1} -fthread-jumps -fcrossjumping \
+-foptimize-sibling-calls -fcse-follow-jumps -fcse-skip-blocks \
+-fgcse  -fgcse-lm -fexpensive-optimizations -frerun-cse-after-loop \
+-fcaller-saves -fpeephole2 -fschedule-insns -fschedule-insns2 \
+-fsched-interblock  -fsched-spec -fregmove -fstrict-aliasing \
+-fstrict-overflow -fdelete-null-pointer-checks -freorder-blocks \
+-freorder-functions -falign-functions -falign-jumps -falign-loops \
+-falign-labels -ftree-vrp -ftree-pre

Our -O2 does not include -fstrict-aliasing, -fstrict-overflow,
-ftree-vrp.


Here's a new patch, taking your remarks and jsg's advice on
USER_CFLAGS into account. I can't test it so please someone with an
alpha apply it and try a full xenocara build (see release(8)) with
it. Thanks.

Oops missed one important chunk. Use this one instead

Fwiw, mlarkin@ built xenocara with this patch on his alpha (thanks!),
and this successfully build libGL, libGLU, and even built games/abuse
and graphics/lsys without issues, so this is clearly an improvement.

I also finally got libGL and libGLU built with the patch. just going to build 
some ports,
now.


built some x11/gnustep/ stuff. that worked out well, and works.

Sebastian



Sebastian




Landry

filename too long in /usr/ports ?

[Alexandre H]

A file is rejected by tar :

# cd /usr
# tar zcf ports.tar.gz ports/
tar: File name too long for ustar 
ports/sysutils/logstash/logstash/patches/patch-vendor_bundle_jruby_1_9_gems_logstash-core-1_5_3-java_lib_logstash_patches_stronger_openssl_defaults_rb

Re: [5.8-stable] math/hdf5

[Adam Jensen]

On Thu, 18 Feb 2016 02:41:54 +0100
Juan Francisco Cantero Hurtado  wrote:

> On Wed, Feb 17, 2016 at 02:58:43PM -0500, Adam Jensen wrote:
> > ===> octave-4.0.0 depends on: hdf5-* - not found
> > ===>  Verifying install for hdf5-* in math/hdf5
> > ===>  Checking files for hdf5-1.8.15pl1p0
> > >> Fetch 
> > >> http://www.hdfgroup.org/ftp/HDF5/current/src/hdf5-1.8.15-patch1.tar.gz
> > ftp: Error retrieving file: 404 Not Found
> > >> Fetch 
> > >> http://ftp.openbsd.org/pub/OpenBSD/distfiles//hdf5-1.8.15-patch1.tar.gz
> > ftp: Error retrieving file: 404 Not Found
> > >> Fetch 
> > >> ftp://ftp.usa.openbsd.org/pub/OpenBSD/distfiles//hdf5-1.8.15-patch1.tar.gz
> > hdf5-1.8.15-patch1.tar.gz: No such file or directory.
> > >> Fetch 
> > >> http://ftp.fr.openbsd.org/pub/OpenBSD/distfiles//hdf5-1.8.15-patch1.tar.gz
> > ftp: Error retrieving file: 404 Not Found
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2954 
> > '/usr/ports/distfiles/hdf5-1.8.15-patch1.tar.gz': @lock=h...)
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2321 '_internal-fetch')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2575 
> > '/usr/ports/pobj/hdf5-1.8.15pl1/.extract_done')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:1952 
> > '/usr/ports/packages/i386/all/hdf5-1.8.15pl1p0.tgz')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2508 '_internal-package')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'package')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:1969 
> > '/var/db/pkg/hdf5-1.8.15pl1p0/+CONTENTS')
> > *** Error 1 in /usr/ports/math/hdf5 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'install')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2112 
> > '/usr/ports/pobj/octave-4.0.0/.dep-math-hdf5')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2575 
> > '/usr/ports/pobj/octave-4.0.0/.extract_done')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:1952 
> > '/usr/ports/packages/i386/all/octave-4.0.0.tgz')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2508 
> > '_internal-package')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'package')
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:1969 
> > '/var/db/pkg/octave-4.0.0/+CONTENTS')
> > *** Error 1 in /usr/ports/math/octave 
> > (/usr/ports/infrastructure/mk/bsd.port.mk:2488 'install')
> > 
> 
> Try this patch:
> 
> Index: Makefile
> ===
> RCS file: /cvs/ports/math/hdf5/Makefile,v
> retrieving revision 1.3
> diff -u -p -r1.3 Makefile
> --- Makefile  30 Jun 2015 09:15:32 -  1.3
> +++ Makefile  18 Feb 2016 01:35:20 -
> @@ -5,7 +5,7 @@ COMMENT = Hierarchical Data Format 5 Te
>  V =  1.8.15-patch1
>  DISTNAME =   hdf5-${V}
>  PKGNAME =hdf5-${V:S/-patch/pl/}
> -REVISION =   0
> +REVISION =   1
>  
>  SHARED_LIBS +=   hdf50.0 # 10.1
>  SHARED_LIBS +=   hdf5_cpp0.0 # 10.1
> @@ -23,7 +23,8 @@ PERMIT_PACKAGE_CDROM =  Yes
>  
>  WANTLIB += c m stdc++ z
>  
> -MASTER_SITES =   http://www.hdfgroup.org/ftp/HDF5/current/src/
> +MASTER_SITES =   https://www.hdfgroup.org/ftp/HDF5/current/src/ \
> + 
> https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-${V}/src/
>  
>  SEPARATE_BUILD = Yes
>  CONFIGURE_STYLE =gnu

Thanks, I used the link from your patch for this work around:

cd /usr/ports/distfiles
wget 
https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.8.15-patch1/src/hdf5-1.8.15-patch1.tar.gz
cd /usr/ports/math/hdf5/
make install clean

Would your patch be the kind of thing that gets pushed into -stable updates? 
(and/or) Will an upcoming -stable update include a fix for /usr/ports/math/hdf5 
(and other problems of this class)?

Re: ghostscript without cups: advice

[Matthias Kilian]

On Fri, Feb 19, 2016 at 02:52:35PM -0500, gwes wrote:
> Why install cups, avahi, and dbus if you'll never use them?

Because it doesn't harm having it *installed* but *not running*,
and because at the same time avoiding those ports to be installed
as dependencies of ghostscript would require either 6 additional
flavor combinations of ghostscript or the dlopen(3) hack to
ghostscript.

> That's three large packages, three user IDs, etc. etc.

Well, how and where does this cause any harm? Except for wasted
traffic when running pkg_add?

> I'm advocating installing less useless code.

But you're also advocating building ghostscript not only 6 but 12
times. That's a burden for the brave people building official
packages on slower architectures. (And, of couse for people testing
updates whenever I unslack).

> >Building an OS is about choice. I am advocating you buy a bigger hardware.
> >If you don't agree, then choose another OS or build you own -- it shouldn't 
> >be difficult to build you own ghostscript from ports with your favorite 
> >options.
> >Oh wait, we even have a framework for it, it's called 'mystuff'.
> >
> If I read the code correctly, gs won't queue its output for printing
> (its default) unless cups and all of its dependencies are running.

Ghostscript is a PostScript to $whatever interpreter. It doesn't
queue it's output to anywhere. IIRC, the cups driver of ghostscript
just emits the results of rendering the input to stdout, which then
is processed by cups. Or did I (again) miss that I didn't miss
something?

Ciao,
Kili

Re: bbpager: bus error

[Jérémie Courrèges-Anglas]

(cc'ing ${MAINTAINER} this time...)

Ryan Freeman  writes:

> On Fri, Feb 19, 2016 at 07:24:22PM +0100, Jérémie Courrèges-Anglas wrote:
>> Alessandro DE LAURENZIS  writes:
>> 
>> > Hello,
>> >
>> > bbpager systematically receives a SIGBUS when a window is closed:
>> >
>> > # gdb bbpager
>> > GNU gdb 6.3
>> > Copyright 2004 Free Software Foundation, Inc.
>> > GDB is free software, covered by the GNU General Public License, and you 
>> > are
>> > welcome to change it and/or distribute copies of it under certain 
>> > conditions.
>> > Type "show copying" to see the conditions.
>> > There is absolutely no warranty for GDB.  Type "show warranty" for details.
>> > This GDB was configured as "amd64-unknown-openbsd5.9"...(no debugging 
>> > symbols found)
>> >
>> > (gdb) run
>> > Starting program: /usr/local/bin/bbpager 
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> > (no debugging symbols found)
>> >
>> > Program received signal SIGBUS, Bus error.
>> > 0x14ea13f13c06 in __register_frame_info () from /usr/local/bin/bbpager
>> >
>> > The problem is 100% reproducible (I tried with different WM: fluxbox,
>> > Openbox, JWM): just launch the program, open anew window and then close
>> > it.
>> >
>> > Any further info needed? Hints?
>> 
>> Give this a try.  I don't understand the purpose of the -- though.
>> *shrug*
>
> This seems to fix the issue for me on amd64, thanks!

Thanks.  The problem is that the iterator is invalidated by erase(); the
diff ensures that after erase() returns, pit points at the previous item
in the list.

oks?

Index: Makefile
===
RCS file: /cvs/ports/x11/bbpager/Makefile,v
retrieving revision 1.18
diff -u -p -r1.18 Makefile
--- Makefile28 Oct 2014 08:18:30 -  1.18
+++ Makefile19 Feb 2016 20:42:34 -
@@ -3,7 +3,7 @@
 COMMENT =  pager for the Blackbox window manager
 
 DISTNAME = bbpager-0.4.7
-REVISION=  1
+REVISION=  2
 
 CATEGORIES =   x11
 MASTER_SITES = ${MASTER_SITE_SOURCEFORGE:=bbtools/}
Index: patches/patch-src_wminterface_cxx
===
RCS file: patches/patch-src_wminterface_cxx
diff -N patches/patch-src_wminterface_cxx
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_wminterface_cxx   19 Feb 2016 20:42:34 -
@@ -0,0 +1,13 @@
+$OpenBSD$
+--- src/wminterface.cxx.orig   Fri Feb 19 19:14:47 2016
 src/wminterface.cxxFri Feb 19 19:15:04 2016
+@@ -84,8 +84,7 @@ void WMInterface::updateWindowList(void)
+ for (; pit != pit_end; pit++) {
+ if (!(*pit)->isMarked()) {
+ delete (*pit);
+-bbtool->pagerWindowList().erase(pit);
+-pit--;
++bbtool->pagerWindowList().erase(pit--);
+ } else {
+ }
+ }


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

Re: ghostscript without cups: advice

[Matthias Kilian]

On Fri, Feb 19, 2016 at 02:31:26PM -0500, gwes wrote:
> >>>I set the default to without cups because that is the
> >>>version of least surprise. It seems to me that installing a graphics
> >>>package shouldn't install a print package, a network package,
> >>>etc. etc. so that's how it's set. I can easily reverse that.
> >>Installing cups-libs (and the packages it depends on, including
> >>avahi) shouldn't cause any harm unless they are really used. Which
> >>isn't the case unless ghostscripts cups driver is used. Correct me
> >>if I'm wrong.
> >You're not wrong. And avahi will not run if you don't enable it.
> >
> Why install cups, avahi, and dbus if you'll never use them?
> That's three large packages, three user IDs, etc. etc.
> I'm advocating installing less useless code.

So you're willing to test (i.e. build *and* test-run several
docuuments with several ghostscript drivers) all combinations of
all the flavors for the next update of ghostscript?  That would be
a collection like this:

(no flavor)
a4
a4,no_cups
gtk
gtk,a4
gtk,a4,no_cups
gtk,no_cups
no_cups
no_x11
no_x11,a4
no_x11,a4,no_cups
no_x11,no_cups

Or should we go back to the (very brittle, IIRC) approach of trying
to dlopen(3) libcups when the cups driver is used?

Ciao,
Kili

ps: i'm aware that the ghostscript port needs an update, and i hope
to find the time to work on it after the release lock phase.

Re: bbpager: bus error

[Ryan Freeman]

On Fri, Feb 19, 2016 at 07:24:22PM +0100, Jérémie Courrèges-Anglas wrote:
> Alessandro DE LAURENZIS  writes:
> 
> > Hello,
> >
> > bbpager systematically receives a SIGBUS when a window is closed:
> >
> > # gdb bbpager
> > GNU gdb 6.3
> > Copyright 2004 Free Software Foundation, Inc.
> > GDB is free software, covered by the GNU General Public License, and you are
> > welcome to change it and/or distribute copies of it under certain 
> > conditions.
> > Type "show copying" to see the conditions.
> > There is absolutely no warranty for GDB.  Type "show warranty" for details.
> > This GDB was configured as "amd64-unknown-openbsd5.9"...(no debugging 
> > symbols found)
> >
> > (gdb) run
> > Starting program: /usr/local/bin/bbpager 
> > (no debugging symbols found)
> > (no debugging symbols found)
> > (no debugging symbols found)
> > (no debugging symbols found)
> > (no debugging symbols found)
> > (no debugging symbols found)
> > (no debugging symbols found)
> >
> > Program received signal SIGBUS, Bus error.
> > 0x14ea13f13c06 in __register_frame_info () from /usr/local/bin/bbpager
> >
> > The problem is 100% reproducible (I tried with different WM: fluxbox,
> > Openbox, JWM): just launch the program, open anew window and then close
> > it.
> >
> > Any further info needed? Hints?
> 
> Give this a try.  I don't understand the purpose of the -- though.
> *shrug*

This seems to fix the issue for me on amd64, thanks!

> 
> Index: patches/patch-src_wminterface_cxx
> ===
> RCS file: patches/patch-src_wminterface_cxx
> diff -N patches/patch-src_wminterface_cxx
> --- /dev/null 1 Jan 1970 00:00:00 -
> +++ patches/patch-src_wminterface_cxx 19 Feb 2016 18:21:33 -
> @@ -0,0 +1,13 @@
> +$OpenBSD$
> +--- src/wminterface.cxx.orig Fri Feb 19 19:14:47 2016
>  src/wminterface.cxx  Fri Feb 19 19:15:04 2016
> +@@ -84,8 +84,7 @@ void WMInterface::updateWindowList(void)
> + for (; pit != pit_end; pit++) {
> + if (!(*pit)->isMarked()) {
> + delete (*pit);
> +-bbtool->pagerWindowList().erase(pit);
> +-pit--;
> ++bbtool->pagerWindowList().erase(pit--);
> + } else {
> + }
> + }
> 
> 
> -- 
> jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE
> 

Re: ghostscript without cups: advice

[gwes]

Why install cups, avahi, and dbus if you'll never use them?
That's three large packages, three user IDs, etc. etc.
I'm advocating installing less useless code.


Building an OS is about choice. I am advocating you buy a bigger hardware.
If you don't agree, then choose another OS or build you own -- it shouldn't be 
difficult to build you own ghostscript from ports with your favorite options.
Oh wait, we even have a framework for it, it's called 'mystuff'.


If I read the code correctly, gs won't queue its output for printing
(its default) unless cups and all of its dependencies are running.

Re: ghostscript without cups: advice

[Antoine Jacoutot]

> Why install cups, avahi, and dbus if you'll never use them?
> That's three large packages, three user IDs, etc. etc.
> I'm advocating installing less useless code.

Building an OS is about choice. I am advocating you buy a bigger hardware.
If you don't agree, then choose another OS or build you own -- it shouldn't be 
difficult to build you own ghostscript from ports with your favorite options.
Oh wait, we even have a framework for it, it's called 'mystuff'.

-- 
Antoine

Re: ghostscript without cups: advice

[gwes]

On 02/19/2016 04:12, Antoine Jacoutot wrote:

On Fri, Feb 19, 2016 at 09:21:41AM +0100, Matthias Kilian wrote:

On Thu, Feb 18, 2016 at 11:54:56PM -0500, gwes wrote:
[...]

Six addotional flavors which gain you nothing?


I set the default to without cups because that is the
version of least surprise. It seems to me that installing a graphics
package shouldn't install a print package, a network package,
etc. etc. so that's how it's set. I can easily reverse that.

Installing cups-libs (and the packages it depends on, including
avahi) shouldn't cause any harm unless they are really used. Which
isn't the case unless ghostscripts cups driver is used. Correct me
if I'm wrong.

You're not wrong. And avahi will not run if you don't enable it.


Why install cups, avahi, and dbus if you'll never use them?
That's three large packages, three user IDs, etc. etc.
I'm advocating installing less useless code.

Geoff

Re: bbpager: bus error

[Jérémie Courrèges-Anglas]

Alessandro DE LAURENZIS  writes:

> Hello,
>
> bbpager systematically receives a SIGBUS when a window is closed:
>
> # gdb bbpager
> GNU gdb 6.3
> Copyright 2004 Free Software Foundation, Inc.
> GDB is free software, covered by the GNU General Public License, and you are
> welcome to change it and/or distribute copies of it under certain conditions.
> Type "show copying" to see the conditions.
> There is absolutely no warranty for GDB.  Type "show warranty" for details.
> This GDB was configured as "amd64-unknown-openbsd5.9"...(no debugging symbols 
> found)
>
> (gdb) run
> Starting program: /usr/local/bin/bbpager 
> (no debugging symbols found)
> (no debugging symbols found)
> (no debugging symbols found)
> (no debugging symbols found)
> (no debugging symbols found)
> (no debugging symbols found)
> (no debugging symbols found)
>
> Program received signal SIGBUS, Bus error.
> 0x14ea13f13c06 in __register_frame_info () from /usr/local/bin/bbpager
>
> The problem is 100% reproducible (I tried with different WM: fluxbox,
> Openbox, JWM): just launch the program, open anew window and then close
> it.
>
> Any further info needed? Hints?

Give this a try.  I don't understand the purpose of the -- though.
*shrug*

Index: patches/patch-src_wminterface_cxx
===
RCS file: patches/patch-src_wminterface_cxx
diff -N patches/patch-src_wminterface_cxx
--- /dev/null   1 Jan 1970 00:00:00 -
+++ patches/patch-src_wminterface_cxx   19 Feb 2016 18:21:33 -
@@ -0,0 +1,13 @@
+$OpenBSD$
+--- src/wminterface.cxx.orig   Fri Feb 19 19:14:47 2016
 src/wminterface.cxxFri Feb 19 19:15:04 2016
+@@ -84,8 +84,7 @@ void WMInterface::updateWindowList(void)
+ for (; pit != pit_end; pit++) {
+ if (!(*pit)->isMarked()) {
+ delete (*pit);
+-bbtool->pagerWindowList().erase(pit);
+-pit--;
++bbtool->pagerWindowList().erase(pit--);
+ } else {
+ }
+ }


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

bbpager: bus error

[Alessandro DE LAURENZIS]

Hello,

bbpager systematically receives a SIGBUS when a window is closed:

# gdb bbpager
GNU gdb 6.3
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "amd64-unknown-openbsd5.9"...(no debugging symbols 
found)

(gdb) run
Starting program: /usr/local/bin/bbpager 
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)

Program received signal SIGBUS, Bus error.
0x14ea13f13c06 in __register_frame_info () from /usr/local/bin/bbpager

The problem is 100% reproducible (I tried with different WM: fluxbox,
Openbox, JWM): just launch the program, open anew window and then close
it.

Any further info needed? Hints?

-- 
Alessandro DE LAURENZIS
[mailto:just22@gmail.com]
LinkedIn: http://it.linkedin.com/in/delaurenzis

Re: devel/intellj - pty4j

[Jiri B]

On Thu, Feb 18, 2016 at 09:29:58AM +0300, Vadim Zhukov wrote:
> Thank you! I started working on pty4j and some related libs which are used
> in IDEA, but didn't finish this work. Your patch should speed up things.
> 
> The problem is that IDEA bundles quiet a few libraries, like pty4j. We
> generally don't like such bundling. I thought about replacing binaries
> within IDEA directory by symlinks and/or files from pty4j package, to avoid
> double patching...
> 
> Anyway, since we're almost locked now, I'll keep an eye on pty4j and see
> what could be done after unlock. My plan is to finish KDE5/Wayland first,
> though: one big fat port at a time. :)

FYI remote debugging via pydev and native libs is relevant for Ultimate
edition. We do community and I can't have access to Ultimate one anymore.

I'll check fsnotifier as I got a diff from JetBrains' employee.

KDE5/Wayland... Wow, thank you for your great work!

j.

Re: ghostscript without cups: advice

[Antoine Jacoutot]

On Fri, Feb 19, 2016 at 09:21:41AM +0100, Matthias Kilian wrote:
> On Thu, Feb 18, 2016 at 11:54:56PM -0500, gwes wrote:
> > This diff allows generating flavors of gs with and without cups.
> > 
> > 
> > Index: ports/print/ghostscript/Makefile
> > ===
> > RCS file: /cvs/ports/print/ghostscript/Makefile,v
> > retrieving revision 1.14
> > diff -u -p -u -r1.14 Makefile
> > --- ports/print/ghostscript/Makefile1 Feb 2010 17:00:14 -   
> > 1.14
> > +++ ports/print/ghostscript/Makefile19 Feb 2016 04:29:07 -
> > @@ -7,6 +7,14 @@
> >  SUBDIR += gnu,gtk,a4
> >  SUBDIR += gnu,no_x11
> >  SUBDIR += gnu,no_x11,a4
> > +.if defined(WITH_CUPS)
> > +SUBDIR += gnu,cups
> > +SUBDIR += gnu,a4,cups
> > +SUBDIR += gnu,gtk,cups
> > +SUBDIR += gnu,gtk,a4,cups
> > +SUBDIR += gnu,no_x11,cups
> > +SUBDIR += gnu,no_x11,a4,cups
> > +.endif
> >  SUBDIR += gnu-fonts
> >  
> >  .include 
> 
> [...]
> 
> Six addotional flavors which gain you nothing?
> 
> > I set the default to without cups because that is the
> > version of least surprise. It seems to me that installing a graphics
> > package shouldn't install a print package, a network package,
> > etc. etc. so that's how it's set. I can easily reverse that.
> 
> Installing cups-libs (and the packages it depends on, including
> avahi) shouldn't cause any harm unless they are really used. Which
> isn't the case unless ghostscripts cups driver is used. Correct me
> if I'm wrong.

You're not wrong. And avahi will not run if you don't enable it.

-- 
Antoine

CVS: cvs.openbsd.org: ports

[Tobias Ulmer]

CVSROOT:/cvs
Module name:ports
Changes by: tobi...@cvs.openbsd.org 2016/02/19 02:07:38

Modified files:
graphics/libraw: Makefile 

Log message:
mark broken on sparc, missing atomic ops

Re: prosody with mysql clarification

[Giovanni Bechis]

On 02/18/16 20:39, Henrik Friedrichsen wrote:
> Stephen has just pointed out that prosodyctl also requires the
> workaround (obviously).
> 
> Should we include this?
> 
I think this should go in, ok giovanni@
 Cheers
  Giovanni
> Index: Makefile
> ===
> RCS file: /cvs/ports/net/prosody/Makefile,v
> retrieving revision 1.40
> diff -u -p -r1.40 Makefile
> --- Makefile  16 Feb 2016 21:42:43 -  1.40
> +++ Makefile  18 Feb 2016 19:38:07 -
> @@ -4,7 +4,7 @@ SHARED_ONLY=  Yes
>  
>  COMMENT= communications server for Jabber/XMPP written in Lua
>  DISTNAME=prosody-0.9.10
> -REVISION=0
> +REVISION=1
>  CATEGORIES=  net
>  MASTER_SITES=http://prosody.im/downloads/source/
>  
> Index: pkg/README
> ===
> RCS file: /cvs/ports/net/prosody/pkg/README,v
> retrieving revision 1.9
> diff -u -p -r1.9 README
> --- pkg/README16 Feb 2016 21:42:43 -  1.9
> +++ pkg/README18 Feb 2016 19:38:07 -
> @@ -46,3 +46,7 @@ Using Prosody in an OpenBSD environment
> For MySQL/MariaDB, also add the following to /etc/login.conf:
>  
> prosody:setenv=LD_PRELOAD=/usr/lib/libpthread.so:tc=daemon:
> +
> +   A similar workaround is required for prosodyctl, e.g. to add an account 
> use:
> +
> +   LD_PRELOAD=/usr/lib/libpthread.so prosodyctl adduser joe
> 

Re: ghostscript without cups: advice

[Matthias Kilian]

On Thu, Feb 18, 2016 at 11:54:56PM -0500, gwes wrote:
> This diff allows generating flavors of gs with and without cups.
> 
> 
> Index: ports/print/ghostscript/Makefile
> ===
> RCS file: /cvs/ports/print/ghostscript/Makefile,v
> retrieving revision 1.14
> diff -u -p -u -r1.14 Makefile
> --- ports/print/ghostscript/Makefile  1 Feb 2010 17:00:14 -   1.14
> +++ ports/print/ghostscript/Makefile  19 Feb 2016 04:29:07 -
> @@ -7,6 +7,14 @@
>  SUBDIR += gnu,gtk,a4
>  SUBDIR += gnu,no_x11
>  SUBDIR += gnu,no_x11,a4
> +.if defined(WITH_CUPS)
> +SUBDIR += gnu,cups
> +SUBDIR += gnu,a4,cups
> +SUBDIR += gnu,gtk,cups
> +SUBDIR += gnu,gtk,a4,cups
> +SUBDIR += gnu,no_x11,cups
> +SUBDIR += gnu,no_x11,a4,cups
> +.endif
>  SUBDIR += gnu-fonts
>  
>  .include 

[...]

Six addotional flavors which gain you nothing?

> I set the default to without cups because that is the
> version of least surprise. It seems to me that installing a graphics
> package shouldn't install a print package, a network package,
> etc. etc. so that's how it's set. I can easily reverse that.

Installing cups-libs (and the packages it depends on, including
avahi) shouldn't cause any harm unless they are really used. Which
isn't the case unless ghostscripts cups driver is used. Correct me
if I'm wrong.

Ciao,
Kili