CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 00:53:00 Modified files: devel/ruby-rspec: Makefile.inc devel/ruby-rspec/1: Makefile devel/ruby-rspec/core: Makefile devel/ruby-rspec/expectations: Makefile devel/ruby-rspec/mocks: Makefile devel/ruby-rspec/rspec: Makefile Log message: - remove unneeded PERMIT_* lines - update HOMEPAGE
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:17:08 Modified files: devel/ruby-diff-lcs: Makefile distinfo devel/ruby-diff-lcs/pkg: PLIST Log message: - update to diff-lcs-1.2.4
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:22:57 Modified files: devel/ruby-rspec/1: Makefile Log message: point this HOMEPAGE to the old rspec-1 pages.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:31:07 Modified files: devel/ruby-jeweler: Makefile distinfo devel/ruby-jeweler/pkg: PLIST Log message: - update to jeweler-1.8.4
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:31:17 Modified files: devel/ruby-extlib: Makefile distinfo devel/ruby-extlib/pkg: PLIST Log message: - update to extlib-0.9.16
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:38:16 Modified files: devel/ruby-extlib: Makefile Removed files: devel/ruby-extlib/patches: patch-Rakefile Log message: remove unneeded patch since we can just use json_pure now that it's around
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:40:42 Modified files: devel/ruby-rspec/expectations: Makefile distinfo devel/ruby-rspec/expectations/pkg: PLIST Log message: update to rspec-expectations-2.13.0
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:40:33 Modified files: devel/ruby-rspec/core: Makefile distinfo devel/ruby-rspec/core/pkg: DESCR PLIST Log message: update to rspec-core-2.13.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:40:50 Modified files: devel/ruby-rspec/mocks: Makefile distinfo devel/ruby-rspec/mocks/pkg: DESCR PLIST Log message: update to rspec-mocks-2.13.1
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:41:17 Modified files: devel/ruby-rspec/rspec: Makefile distinfo devel/ruby-rspec/rspec/pkg: PLIST Log message: - update to rspec-2.13.0 - tighten dependencies while here
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 01:44:43 Modified files: devel/ruby-cucumber: Makefile Log message: wow, this was a useful HOMEPAGE, still, let's fix it.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2013/04/23 01:59:31 Modified files: www/tomcat/v7 : Makefile distinfo www/tomcat/v7/pkg: PLIST-examples Log message: Update to tomcat-7.0.39
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: giova...@cvs.openbsd.org2013/04/23 03:22:46 Modified files: mail/postfwd : Makefile distinfo mail/postfwd/pkg: PLIST Log message: Update to 1.35, some compatibility fixes with newer Perl
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2013/04/23 03:31:33 Modified files: net/openvpn: Makefile Removed files: net/openvpn/patches: patch-doc_openvpn_8 patch-src_openvpn_init_c patch-src_openvpn_options_c patch-src_openvpn_options_h patch-src_openvpn_socket_c patch-src_openvpn_socket_h Log message: remove the (minimal) rtable support; shortly after it was added to the OpenVPN port, exec was added to route(8) which provides the same support in a more general way. use a command like this instead: route -T table# exec openvpn -daemon ...
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2013/04/23 03:40:58 Modified files: www/webkit : Makefile Added files: www/webkit/patches: patch-Source_WTF_wtf_StackBounds_cpp Log message: Add a patch to fix stack bounds computation, from David Hill (also pushed upstream as https://bugs.webkit.org/show_bug.cgi?id=114978) Seems to fix xombrero/surf crashes on amd64, and seems to magically make webkit sort-of usable (ie dead-slow, but js seems to work) on powerpc. (symptom : ** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded.) Note : webkit still fails to build on mips64* and hppa. sigh. ok jasper@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gonz...@cvs.openbsd.org 2013/04/23 06:00:21 Modified files: www/aria2 : Makefile distinfo Log message: Update for Aria2 to 1.17.0: * Save options directly specified for download in --save-session * Save URI returned only from FileEntry::getRemainingUris() * Use info level log for system trusted ca imports failure * Implement simple Happy Eyeballs for HTTP/FTP downloads Ok benoit@
Re: CVS: cvs.openbsd.org: ports
On Tue, Apr 23, 2013 at 03:40:58AM -0600, Landry Breuil wrote: CVSROOT: /cvs Module name: ports Changes by: lan...@cvs.openbsd.org 2013/04/23 03:40:58 Modified files: www/webkit : Makefile Added files: www/webkit/patches: patch-Source_WTF_wtf_StackBounds_cpp Log message: Add a patch to fix stack bounds computation, from David Hill (also pushed upstream as https://bugs.webkit.org/show_bug.cgi?id=114978) Seems to fix xombrero/surf crashes on amd64, and seems to magically make webkit sort-of usable (ie dead-slow, but js seems to work) on powerpc. (symptom : ** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded.) Note : webkit still fails to build on mips64* and hppa. sigh. ok jasper@ So the logic was not checked/fixed where the bounds are actually checked on architectures that have upward growing stacks? Where m_bound and m_origin have their meanings swapped? I only see MACHINE_STACK_GROWS_UP defined on hppa and hppa64. Are these really the only archs with upward growing stacks? Ken
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 07:10:46 Modified files: audio/xmp : Makefile Added files: audio/xmp/patches: patch-src_loaders_masi_load_c Log message: security fix for CVE-2013-1980, libxmp MASI Parsing Buffer Overflow Vulnerability note that this port could use an update by someone who cares..
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 07:17:07 Modified files: devel/srecord : Makefile distinfo devel/srecord/pkg: PLIST Removed files: devel/srecord/pkg: PFRAG.shared Log message: update to srecord-1.61
Re: CVS: cvs.openbsd.org: ports
On Tue, Apr 23, 2013 at 03:10:04PM +0200, Landry Breuil wrote:
On Tue, Apr 23, 2013 at 08:57:31AM -0400, Kenneth R Westerback wrote:
On Tue, Apr 23, 2013 at 03:40:58AM -0600, Landry Breuil wrote:
CVSROOT: /cvs
Module name: ports
Changes by: lan...@cvs.openbsd.org 2013/04/23 03:40:58
Modified files:
www/webkit : Makefile
Added files:
www/webkit/patches: patch-Source_WTF_wtf_StackBounds_cpp
Log message:
Add a patch to fix stack bounds computation, from David Hill (also
pushed upstream as https://bugs.webkit.org/show_bug.cgi?id=114978)
Seems to fix xombrero/surf crashes on amd64, and seems to magically make
webkit sort-of usable (ie dead-slow, but js seems to work) on powerpc.
(symptom : ** Message: console message: undefined @0: RangeError: Maximum
call stack size exceeded.)
Note : webkit still fails to build on mips64* and hppa. sigh.
ok jasper@
So the logic was not checked/fixed where the bounds are actually
checked on architectures that have upward growing stacks? Where
m_bound and m_origin have their meanings swapped?
The previous logic was doing (note the lovely FIXME) :
67 // FIXME: remove this! - this code unsafely guesses at stack sizes!
68 static const ptrdiff_t estimatedStackSize = 128 * sizeof(void*) * 1024;
69 // This method assumes the stack is growing downwards.
70 static void* estimateStackBound(void* origin)
71 {
72 return static_castchar*(origin) - estimatedStackSize;
73 }
122 void StackBounds::initialize()
123 {
124 pthread_t thread = pthread_self();
125 stack_t stack;
126 pthread_stackseg_np(thread, stack);
127 m_origin = stack.ss_sp;
128 m_bound = estimateStackBound(m_origin);
129 }
The new logic does :
122 void StackBounds::initialize()
123 {
124 pthread_t thread = pthread_self();
125 stack_t stack;
126 pthread_stackseg_np(thread, stack);
127 m_origin = stack.ss_sp;
128 #if defined(__hppa__) || defined(__hppa64__)
129 // hppa's stack grows up
130 m_bound = static_castchar*(m_origin) + stack.ss_size;
131 #else
132 m_bound = static_castchar*(m_origin) - stack.ss_size;
133 #endif
134 }
I am familiar with the patch. My concern is that on upward growing
stack machines, if the code is doing things like
if (new_stack_pos m_bound)
life_is_good()
else
stack_go_boom()
having a better value for m_bound doesn't fully fix the problem. If the
code doesn't make such assumptions then I guess things are fine.
Ken
I only see MACHINE_STACK_GROWS_UP defined on hppa and hppa64. Are
these really the only archs with upward growing stacks?
It seems so. Do you mean we should pull that #define from a header
instead of hardcoding hppa* here ?
Landry
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: gsoa...@cvs.openbsd.org 2013/04/23 08:32:26 Modified files: mail/msmtp : Makefile distinfo mail/msmtp/patches: patch-scripts_msmtpq_msmtpq Log message: update to 1.4.31; OK dcoppa@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2013/04/23 08:50:19 Modified files: comms/rtl-sdr : Makefile comms/rtl-sdr/pkg: DESCR PLIST Added files: comms/rtl-sdr/patches: patch-src_rtl_fm_c Log message: add synchronous mode and hack around mutex problems, allowing rtl_fm to work.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jas...@cvs.openbsd.org 2013/04/23 10:41:27 Modified files: www/puppet-dashboard: Makefile www/puppet-dashboard/pkg: README Log message: use the full path to rake in the crontab lines, as /usr/local isn't in the PATH ok aja@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2013/04/23 10:47:01 Modified files: www/p5-Apache-Reload: Makefile www/p5-Apache-Reload/pkg: PLIST Log message: Apache2::Reload is included in ap2-mod_perl, remove it from p5-Apache-Reload which is only intended for use with the old version of mod_perl. Fixes problem pointed out by Chris Bennett.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2013/04/23 12:58:20 Log message: Import p5-autobox 2.77 from Chris Bennett. The autobox pragma allows methods to be called on integers, floats, strings, arrays, hashes, and code references in exactly the same manner as blessed references. The autoboxing is transparent: boxed values are not blessed into their (user-defined) implementation class (unless the method elects to bestow such a blessing) - they simply use its methods as though they are. ok sthen@ Status: Vendor Tag: cbennett Release Tags: landry_20130423 N ports/devel/p5-autobox/distinfo N ports/devel/p5-autobox/Makefile N ports/devel/p5-autobox/pkg/PLIST N ports/devel/p5-autobox/pkg/DESCR No conflicts created by this import
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: lan...@cvs.openbsd.org 2013/04/23 12:58:57 Modified files: devel : Makefile Log message: +p5-autobox
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: dco...@cvs.openbsd.org 2013/04/23 12:59:53 Modified files: devel/libev: Makefile distinfo devel/libev/patches: patch-Makefile_in patch-configure devel/libev/pkg: PLIST Removed files: devel/libev/patches: patch-ev_c devel/libev/pkg: PFRAG.shared Log message: Update to libev-4.15. While here, merge PFRAG.shared into PLIST. OK sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: st...@cvs.openbsd.org 2013/04/23 14:34:16 Modified files: security/clamav: Makefile distinfo security/clamav/patches: patch-libclamav_Makefile_in Log message: SECURITY update to ClamAV 0.97.8 ClamAV 0.97.8 addresses several reported potential security bugs. Thanks to Felix Groebert of the Google Security Team for finding and reporting these issues. - fixes in PE and autoit decoding - off by one malloc in SIS format handler - minor changes to freshclam - add error checking in md5 (memcpy etc.)
Re: CVS: cvs.openbsd.org: ports
On 2013/04/23 14:34, Stuart Henderson wrote: CVSROOT: /cvs Module name: ports Changes by: st...@cvs.openbsd.org 2013/04/23 14:34:16 Modified files: security/clamav: Makefile distinfo ... - add error checking in md5 (memcpy etc.) the latter being for Windows, it would appear.
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: jtur...@cvs.openbsd.org 2013/04/23 17:04:45 Modified files: sysutils/entr : Makefile distinfo sysutils/entr/pkg: DESCR Removed files: sysutils/entr/patches: patch-regress_sh Log message: Update entr to 1.9. MAINTAINER had similar diff. ok sthen@
CVS: cvs.openbsd.org: ports
CVSROOT:/cvs Module name:ports Changes by: b...@cvs.openbsd.org2013/04/23 20:21:20 Modified files: x11/mplayer/files: ai_sndio.c Log message: Add missing license. ok ratchov@
