CVS: cvs.openbsd.org: ports

[Jeremie Courreges-Anglas]

CVSROOT:/cvs
Module name:ports
Changes by: j...@cvs.openbsd.org2016/10/16 17:52:31

Modified files:
converters : Makefile 
devel  : Makefile 
net: Makefile 
security   : Makefile 
textproc   : Makefile 

Log message:
Hook up tested lua52 and lua53 FLAVORS.

CVS: cvs.openbsd.org: ports

[Jeremie Courreges-Anglas]

CVSROOT:/cvs
Module name:ports
Changes by: j...@cvs.openbsd.org2016/10/16 17:21:23

Modified files:
converters/luastruct: Makefile 
devel/coxpcall : Makefile 
devel/lpeg : Makefile 
devel/lua-cjson: Makefile 
devel/lua-cliargs: Makefile 
devel/lua-cmsgpack: Makefile 
devel/lua-lgi  : Makefile 
devel/lua-penlight: Makefile 
devel/lua-term : Makefile 
devel/luaalarm : Makefile 
devel/luabitop : Makefile 
devel/luacopas : Makefile 
devel/luadoc   : Makefile 
devel/luafs: Makefile 
devel/luapack  : Makefile 
devel/luaprofiler: Makefile 
devel/luarings : Makefile 
devel/luastdlib: Makefile 
net/luasocket  : Makefile 
security/luasec: Makefile 
textproc/lua-lustache: Makefile 
textproc/lua-markdown: Makefile 
textproc/lua-yaml: Makefile 
textproc/luaexpat: Makefile 
textproc/luasoldout: Makefile 

Log message:
Add explicit lua52/lua53 FLAVORS where supported.

Discussed on ports@, sthen@ agrees with the direction.

CVS: cvs.openbsd.org: ports

[Alexander Bluhm]

CVSROOT:/cvs
Module name:ports
Changes by: bl...@cvs.openbsd.org   2016/10/16 17:06:31

Modified files:
converters/p5-Cpanel-JSON-XS: Makefile distinfo 

Log message:
update p5-Cpanel-JSON-XS to 3.0218

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/10/16 15:26:06

Modified files:
textproc/zathura/core: Makefile 

Log message:
zathura needs gettext-tools to build, found by nigel@

Re: neomutt

[Stefan Sperling]

On Mon, Sep 26, 2016 at 01:07:30PM +0100, Stuart Henderson wrote:
> People who were seeing crashes with neomutt, please test this, and get
> me backtraces if you still see them (the port Makefile here is setup to
> build with symbols so you don't need to set DEBUG in this case).
> 

I've been running this with no issues.

CVS: cvs.openbsd.org: ports

[Jeremie Courreges-Anglas]

CVSROOT:/cvs
Module name:ports
Changes by: j...@cvs.openbsd.org2016/10/16 14:10:48

Modified files:
devel/luaprofiler: Makefile 

Log message:
Standard Makefile layout

Re: iridium

[Antoine Jacoutot]

On Sun, Oct 16, 2016 at 03:57:11PM +0200, frantisek holop wrote:
> hello fellow iridium users, if any.
> 
> recently iridium starts for me only every second time,
> dumping core on the first one.  i removed the old
> profile data and still that's the case.
> 
> any ideas?

Probably the same use after free that chromium has.

-- 
Antoine

CVS: cvs.openbsd.org: ports

[Andrew Fresh]

CVSROOT:/cvs
Module name:ports
Changes by: afre...@cvs.openbsd.org 2016/10/16 12:25:07

Modified files:
devel/p5-Scope-Upper: Makefile distinfo 

Log message:
Update devel/p5-Scope-Upper to 0.29 for perl 5.24 support

OK sthen@

CVS: cvs.openbsd.org: ports

[Andrew Fresh]

CVSROOT:/cvs
Module name:ports
Changes by: afre...@cvs.openbsd.org 2016/10/16 12:23:53

Modified files:
devel/p5-Coro  : Makefile distinfo 
Added files:
devel/p5-Coro/patches: patch-Coro_State_xs 

Log message:
Update devel/p5-Coro to 6.511

Plus additional patch for perl 5.24 and disable JIT due to W^X violations.

OK sthen@

CVS: cvs.openbsd.org: ports

[Andrew Fresh]

CVSROOT:/cvs
Module name:ports
Changes by: afre...@cvs.openbsd.org 2016/10/16 12:22:11

ports/devel/p5-Coro/patches

Update of /cvs/ports/devel/p5-Coro/patches
In directory cvs.openbsd.org:/tmp/cvs-serv39179/patches

Log Message:
Directory /cvs/ports/devel/p5-Coro/patches added to the repository

CVS: cvs.openbsd.org: ports

[Jeremie Courreges-Anglas]

CVSROOT:/cvs
Module name:ports
Changes by: j...@cvs.openbsd.org2016/10/16 12:15:22

Modified files:
editors/emacs  : Makefile 
editors/emacs/patches: patch-lisp_vc_vc-cvs_el 

Log message:
Merge better fix for #24082

This one lets unregistered files appear in vc-dir.

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Jeremie Courreges-Anglas]

Simon Mages  writes:

> Hi,
>
> 2016-10-16 12:14 GMT+02:00, Stuart Henderson :
>> On 2016/10/16 12:09, Simon Mages wrote:
>>> Ok, thanks for the feedback.
>>>
>>> But i really don't get the llvm dependency.
>>> I used the llvm in base on current to build
>>> this port. Why shouldi suddenly use this
>>> one?
>>
>> LLVM is not enabled in base. Anyway, setting MODULES=lang/clang will
>> add the dep automatically.
> Finally i got it, i was working with wrong assumptions, sorry for that.
>
> Attached the new port.
>
> # tar zcvf /tmp/cbmc.tar.gz devel/cbmc
> devel/cbmc
> devel/cbmc/Makefile
> devel/cbmc/patches
> devel/cbmc/patches/patch-src_big-int_bigint-test_cc
> devel/cbmc/patches/patch-src_common~
> devel/cbmc/patches/patch-src_big-int_bigint_cc
> devel/cbmc/patches/patch-src_common
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_core_Solver_cc
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_core_SolverTypes_h
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_mtl_IntTypes_h
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_mtl_Vec_h
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_simp_SimpSolver_cc
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_utils_Options_h
> devel/cbmc/patches/patch-minisat-2_2_1_minisat_utils_ParseUtils_h
> devel/cbmc/distinfo
> devel/cbmc/pkg
> devel/cbmc/pkg/PLIST
> devel/cbmc/pkg/DESCR
>
> Lets see if now everything is alright :)

Looks nicer, here are a few improvements:
- some spacing nits
- use MASTER_SITE_DEBIAN
- as Stuart pointed out, no need for p5-libwww any more
- respect CFLAGS, CXXFLAGS and LDFLAGS, using MAKE_FLAGS; this helps
  a lot for debug builds
- instead of putting the minisat2 patches in the port patches/
  directory, we could also just use upstream's patch in post-extract.
  From my POV those patches aren't my concern, so I don't care about
  them being tracked by CVS.
- patch the include file that provides the "compatibility" goo for
  alloca, instead of patching consumers.

Updated tarball below.

Note:  you create a dedicated section where to set CP_CXXFLAGS, LINKLIB,
etc on OpenBSD, an alternative could be to just amend the FreeBSD
section:

-else ifeq ($(filter-out FreeBSD,$(BUILD_ENV_)),)
+else ifeq ($(filter-out FreeBSD OpenBSD,$(BUILD_ENV_)),)

I doubt that it matters much, upstream may have an opinion already.



cbmc.3.tar.gz
Description: Binary data

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

CVS: cvs.openbsd.org: ports

[Frederic Cambus]

CVSROOT:/cvs
Module name:ports
Changes by: fcam...@cvs.openbsd.org 2016/10/16 10:29:44

Modified files:
net/libmaxminddb: Makefile distinfo 

Log message:
Update GeoLite2 databases to latest version.

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Simon Mages]

Hi,

2016-10-16 12:14 GMT+02:00, Stuart Henderson :
> On 2016/10/16 12:09, Simon Mages wrote:
>> Ok, thanks for the feedback.
>>
>> But i really don't get the llvm dependency.
>> I used the llvm in base on current to build
>> this port. Why shouldi suddenly use this
>> one?
>
> LLVM is not enabled in base. Anyway, setting MODULES=lang/clang will
> add the dep automatically.
Finally i got it, i was working with wrong assumptions, sorry for that.

Attached the new port.

# tar zcvf /tmp/cbmc.tar.gz devel/cbmc
devel/cbmc
devel/cbmc/Makefile
devel/cbmc/patches
devel/cbmc/patches/patch-src_big-int_bigint-test_cc
devel/cbmc/patches/patch-src_common~
devel/cbmc/patches/patch-src_big-int_bigint_cc
devel/cbmc/patches/patch-src_common
devel/cbmc/patches/patch-minisat-2_2_1_minisat_core_Solver_cc
devel/cbmc/patches/patch-minisat-2_2_1_minisat_core_SolverTypes_h
devel/cbmc/patches/patch-minisat-2_2_1_minisat_mtl_IntTypes_h
devel/cbmc/patches/patch-minisat-2_2_1_minisat_mtl_Vec_h
devel/cbmc/patches/patch-minisat-2_2_1_minisat_simp_SimpSolver_cc
devel/cbmc/patches/patch-minisat-2_2_1_minisat_utils_Options_h
devel/cbmc/patches/patch-minisat-2_2_1_minisat_utils_ParseUtils_h
devel/cbmc/distinfo
devel/cbmc/pkg
devel/cbmc/pkg/PLIST
devel/cbmc/pkg/DESCR

Lets see if now everything is alright :)

BR
Simon


cbmc.tar.gz
Description: GNU Zip compressed data

CVS: cvs.openbsd.org: ports

[Frederic Cambus]

CVSROOT:/cvs
Module name:ports
Changes by: fcam...@cvs.openbsd.org 2016/10/16 08:20:11

Modified files:
textproc   : Makefile 

Log message:
Hook lua52 flavors to the build for the Lua ports I maintain.

iridium

[frantisek holop]

hello fellow iridium users, if any.

recently iridium starts for me only every second time,
dumping core on the first one.  i removed the old
profile data and still that's the case.

any ideas?

OpenBSD 6.0-current (obj) #0: Sat Oct 15 08:45:10 MDT 2016

iridium-51.1p2  Iridium browser

-f
-- 
if you don't think women are explosive, just drop one!

Re: CVS: cvs.openbsd.org: ports

[Stuart Henderson]

On 2016/10/16 14:54, Stuart Henderson wrote:
> Oops, I had intended to commit Makefile.inc with that comment, and then
> commit the others separately. Here are the commit logs I would have used.

Oops2: build breakage with plugins/mupdf was spotted by nigel@ following the
mupdf update.

Re: CVS: cvs.openbsd.org: ports

[Stuart Henderson]

Oops, I had intended to commit Makefile.inc with that comment, and then
commit the others separately. Here are the commit logs I would have used.

plugins/cb: update to 0.1.5
plugins/djvu:   update to 0.2.5
plugins/mupdf:  update to 0.3.0, remove old mupdf-1.8 patches, add mupdf-1.9 
patches
plugins/ps: update to 0.2.3
plugins/poppler: update to 0.2.6


On 2016/10/16 07:51, Stuart Henderson wrote:
> CVSROOT:  /cvs
> Module name:  ports
> Changes by:   st...@cvs.openbsd.org   2016/10/16 07:51:21
> 
> Modified files:
>   textproc/zathura/plugins: Makefile.inc 
>   textproc/zathura/plugins/cb: Makefile distinfo 
>   textproc/zathura/plugins/djvu: Makefile distinfo 
>   textproc/zathura/plugins/mupdf: Makefile distinfo 
>   textproc/zathura/plugins/mupdf/patches: patch-document_c 
>   patch-image_c 
>   patch-page_c 
>   patch-search_c 
>   patch-utils_c 
>   textproc/zathura/plugins/poppler: Makefile distinfo 
>   textproc/zathura/plugins/ps: Makefile distinfo 
> Added files:
>   textproc/zathura/plugins/mupdf/patches: patch-config_mk 
>   patch-plugin_h 
> Removed files:
>   textproc/zathura/plugins/mupdf/patches: patch-index_c 
>   patch-links_c 
>   patch-plugin_c 
>   patch-render_c 
> 
> Log message:
> use ports compiler for c11, remove devel/gettext MODULES, remove unnecessary 
> PERMIT_*

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/10/16 07:51:21

Modified files:
textproc/zathura/plugins: Makefile.inc 
textproc/zathura/plugins/cb: Makefile distinfo 
textproc/zathura/plugins/djvu: Makefile distinfo 
textproc/zathura/plugins/mupdf: Makefile distinfo 
textproc/zathura/plugins/mupdf/patches: patch-document_c 
patch-image_c 
patch-page_c 
patch-search_c 
patch-utils_c 
textproc/zathura/plugins/poppler: Makefile distinfo 
textproc/zathura/plugins/ps: Makefile distinfo 
Added files:
textproc/zathura/plugins/mupdf/patches: patch-config_mk 
patch-plugin_h 
Removed files:
textproc/zathura/plugins/mupdf/patches: patch-index_c 
patch-links_c 
patch-plugin_c 
patch-render_c 

Log message:
use ports compiler for c11, remove devel/gettext MODULES, remove unnecessary 
PERMIT_*

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/10/16 07:50:35

Modified files:
textproc/zathura/core: Makefile distinfo 
textproc/zathura/core/pkg: PLIST 

Log message:
update to zathura-0.3.6

Re: NEW: emulators/libretro-pcsx_rearmed

[Frederic Cambus]

On Wed, Sep 28, 2016 at 12:18:58AM -0600, Anthony J. Bentley wrote:
> > libretro-pcsx_rearmed provides a Sony PlayStation emulation core to the
> > RetroArch emulator frontend, based on PCSX ReARMed.
> > 
> > It is yet another PCSX fork based on the PCSX-Reloaded project, which itself
> > contains code from PCSX, PCSX-df and PCSX-Revolution.
> > 
> > Can be tested with the retroarch package.
> 
> Anyone willing to ok this port?

I have been playing a few Bomberman games over the last few weeks and
the emulator works nicely.

Some nitpicking, though: I would change underscore to dash in DISTNAME
and DESCR. The various README files name the emulator as "PCSX-ReArmed",
and it's a bit akward to have underscores in ports/packages names.

Also, it would be nice to remove the failing calls to Git in Makefile,
possibly doing something like this:

--- Makefile.orig   Sun Oct 16 10:57:23 2016
+++ MakefileSun Oct 16 10:58:01 2016
@@ -240,9 +240,7 @@ libpcsxcore/gte_nf.o: libpcsxcore/gte.c
$(CC) -c -o $@ $^ $(CFLAGS) -DFLAGLESS
 
 frontend/revision.h: FORCE
-   @(git describe || echo) | sed -e 's/.*/#define REV "\0"/' > $@_
-   @diff -q $@_ $@ > /dev/null 2>&1 || cp $@_ $@
-   @rm $@_
+   @echo '#define REV "0"' > $@
 
 %.o: %.S
$(CC_AS) $(CFLAGS) -c $^ -o $@

Re: NEW: emulators/libretro-genesis-plus-dx

[Frederic Cambus]

On Wed, Sep 28, 2016 at 12:18:40AM -0600, Anthony J. Bentley wrote:

> > Genesis Plus GX is an open-source Sega 8/16 bit emulator focused on accuracy
> > and portability.
> 
> Anyone willing to ok this port?

Builds and works fine for me, tried it with a few games and everything
worked as expected.

On a side note, RetroArch is great, it's really nice not to have an 
unified interface for all the emulators, and not to have to relearn key
mapping for each emulator, clearly this is the way forward.

OK fcambus@

Re: openvpn-auth-ldap problem

[Stuart Henderson]

On 2016/10/11 19:00, Leclerc, Sebastien wrote:
> When starting OpenVPN with module openvpn-auth-ldap activated, I get the 
> following:
> 
> openvpn:/usr/local/lib/openvpn-auth-ldap.so: /usr/lib/libobjc.so.6.0 : 
> WARNING: symbol(__objc_class_name_Protocol) size mismatch, relink your program
> openvpn:/usr/local/lib/openvpn-auth-ldap.so: /usr/lib/libobjc.so.6.0 : 
> WARNING: symbol(__objc_class_name_Object) size mismatch, relink your program
> 
> Openvpn and openvpn-auth-ldap are installed as packages.
> 
> This is from a snapshot (amd64) :
> OpenBSD 6.0-current (GENERIC.MP) #2518: Sun Oct  2 21:41:07 MDT 2016
> openvpn-2.3.11
> openvpn-auth-ldap-2.0.3.p2
> 
> Same result from 5.9 with patch (amd64) :
> OpenBSD 5.9 (GENERIC.MP) #8: Thu Jul 14 20:12:22 CEST 2016
> openvpn-2.3.10
> openvpn-auth-ldap-2.0.3p1
> 
> 
> Sebastien Leclerc
> 

Two problems.

One: base and ports gcc use the same library version for libobjc.
I'm sure this is wrong and hid the problem (otherwise pkg_create
checks would have caught it).

Index: 4.9/Makefile
===
RCS file: /cvs/ports/lang/gcc/4.9/Makefile,v
retrieving revision 1.41
diff -u -p -r1.41 Makefile
--- 4.9/Makefile18 Sep 2016 15:05:15 -  1.41
+++ 4.9/Makefile16 Oct 2016 10:57:43 -
@@ -4,7 +4,7 @@ ONLY_FOR_ARCHS = amd64 arm hppa i386 mip
 DPB_PROPERTIES = parallel
 
 V = 4.9.4
-REVISION = 2
+REVISION = 3
 FULL_VERSION = $V
 FULL_PKGVERSION = $V
 
@@ -44,7 +44,7 @@ SHARED_LIBS = estdc++ 17.0 \
gcj 4.0 \
gcj-tools   4.0 \
gij 4.0 \
-   objc6.0 \
+   objc7.0 \
ssp 4.0 \
lto_plugin  3.0 \
go  3.0 \


Two, even with this fixed, the linker is preferring the version
in /usr/local/lib/ (possibly because it's a higher-numbered version?)
I'm not sure how to fix this. So the diff above switches us from
building a broken package if gcc-libs is installed, to failing
to build if gcc-libs is installed. So it doesn't really gain us
much in practical terms for openvpn-auth-ldap.

I also tried to workaround by building with ports gcc, ports
infrastructure isn't really setup for objc builds - it's easy enough to
add that, but in the case of openvpn-auth-ldap the configure script is
failing when doing checks on the objc preprocessor.

It would be nice if someone with knowledge about the objc toolchain
or an interest in fixing openvpn-auth-ldap could take a look.

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Stuart Henderson]

On 2016/10/16 12:09, Simon Mages wrote:
> Ok, thanks for the feedback.
> 
> But i really don't get the llvm dependency.
> I used the llvm in base on current to build
> this port. Why shouldi suddenly use this
> one?

LLVM is not enabled in base. Anyway, setting MODULES=lang/clang will
add the dep automatically.

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Simon Mages]

Ok, thanks for the feedback.

But i really don't get the llvm dependency.
I used the llvm in base on current to build
this port. Why shouldi suddenly use this
one?

BR

Simon


Am Sonntag, 16. Oktober 2016 schrieb Stuart Henderson :

> On 2016/10/16 07:58, Simon Mages wrote:
> > > first, nice port.  I would like to test my own code with it.  But you
> > > missed the clang dependency.
> >
> > Well, i thought that a new port will not be taged for a past release. If
> > this is wrong i can
> > add the dependency.
>
> I don't understand what you're saying here.
>
> > > bsd,port.mk(5):
> > > REVISION
> > > Revision number of the current package.  Defaults to empty
> (very
> > > first package),...
> > >
> > > Why do you not fixed the missing RCS tags of the portscheck you made?
> >
> > What should i fix there? I thought cvs will add
> > them on commit.
>
> You should have the "$OpenBSD$" part (with the preceding # in Makefile),
> cvs will fill in the committer account, revision, timestamp, etc. Some
> of the missing tags were on patches/*; these should be generated with
> "make update-patches" which adds the tags for you.
>
> Other comments -
>
> ..
>
> # this port will apply patches to minisat
> pre-build:
> cd ${WRKSRC} && ${MAKE_PROGRAM} minisat2-download
>
> You can't do that, package build machines don't have network access.
> (Also note that it is unsafe! There is no checking of hash or signature
> in their Makefile). It will need to be downloaded by setting DISTFILES
> instead, like this
>
> DISTFILES=  ${DISTNAME}${EXTRACT_SUFX} \
> minisat2_2.2.1.orig.tar.gz:0
>
> MASTER_SITES0=  http://ftp.debian.org/debian/pool/main/m/minisat2/
>
> and 'make makesum' to regenerate distinfo. It will then be untarred
> during the extract stage but might not be in the place you need; replace
> your pre-build line with any necessary mv and patch commands. Better to
> put those in post-patch rather than pre-build. I think this may also
> allow you to remove the lwp dependency.
>
> ..
>
> skip this:
>
> +  ifeq ($(origin CC),default)
> +CC = clang
> +  endif
> +  ifeq ($(origin CXX),default)
> +CXX= clang++
> +  endif
>
> and use this in Makefile instead:
>
> MODULES=lang/clang
> MODCLANG_LANGS= c++
> MODCLANG_ARCHS= *
>
> ..
>
> Put the GH_* bits up where you have DISTNAME (like in Makefile.template).
>
> ..
>
> # 4-clause BSD license
>
> we don't differentiate between versions of the BSD license in these
> markers, better to just put "BSD" - if users care about specific
> details they should be checking for themselves. (We do have more
> specifics for GPL because the different versions are incompatible
> with different other licenses).
>
>

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Stuart Henderson]

On 2016/10/16 07:58, Simon Mages wrote:
> > first, nice port.  I would like to test my own code with it.  But you
> > missed the clang dependency.
> 
> Well, i thought that a new port will not be taged for a past release. If
> this is wrong i can
> add the dependency.

I don't understand what you're saying here.

> > bsd,port.mk(5):
> > REVISION
> > Revision number of the current package.  Defaults to empty (very
> > first package),...
> >
> > Why do you not fixed the missing RCS tags of the portscheck you made?
> 
> What should i fix there? I thought cvs will add
> them on commit.

You should have the "$OpenBSD$" part (with the preceding # in Makefile),
cvs will fill in the committer account, revision, timestamp, etc. Some
of the missing tags were on patches/*; these should be generated with
"make update-patches" which adds the tags for you.

Other comments -

..

# this port will apply patches to minisat
pre-build:
cd ${WRKSRC} && ${MAKE_PROGRAM} minisat2-download

You can't do that, package build machines don't have network access.
(Also note that it is unsafe! There is no checking of hash or signature
in their Makefile). It will need to be downloaded by setting DISTFILES
instead, like this

DISTFILES=  ${DISTNAME}${EXTRACT_SUFX} \
minisat2_2.2.1.orig.tar.gz:0

MASTER_SITES0=  http://ftp.debian.org/debian/pool/main/m/minisat2/

and 'make makesum' to regenerate distinfo. It will then be untarred
during the extract stage but might not be in the place you need; replace
your pre-build line with any necessary mv and patch commands. Better to
put those in post-patch rather than pre-build. I think this may also
allow you to remove the lwp dependency.

..

skip this:

+  ifeq ($(origin CC),default)
+CC = clang
+  endif
+  ifeq ($(origin CXX),default)
+CXX= clang++
+  endif

and use this in Makefile instead:

MODULES=lang/clang
MODCLANG_LANGS= c++
MODCLANG_ARCHS= *

..

Put the GH_* bits up where you have DISTNAME (like in Makefile.template).

..

# 4-clause BSD license

we don't differentiate between versions of the BSD license in these
markers, better to just put "BSD" - if users care about specific
details they should be checking for themselves. (We do have more
specifics for GPL because the different versions are incompatible
with different other licenses).

CVS: cvs.openbsd.org: ports

[Stuart Henderson]

CVSROOT:/cvs
Module name:ports
Changes by: st...@cvs.openbsd.org   2016/10/16 03:16:52

Modified files:
textproc/mupdf : Makefile distinfo 
textproc/mupdf/patches: patch-Makefile patch-Makerules 
patch-Makethird patch-docs_man_mupdf_1 
patch-platform_gl_gl-main_c 
patch-platform_x11_pdfapp_c 
patch-source_fitz_list-device_c 
patch-source_pdf_pdf-xref_c 
patch-source_tools_mudraw_c 
patch-source_tools_pdfclean_c 
patch-source_tools_pdfposter_c 
patch-source_tools_pdfshow_c 
textproc/mupdf/pkg: PLIST 
Added files:
textproc/mupdf/patches: patch-source_tools_murun_c 
patch-source_tools_mutool_c 

Log message:
update to mupdf-1.9a

Re: [PATCH] for mupdf 1.8

[Stuart Henderson]

Thanks - I've just commited an update to 1.9 which I think should include this 
fix.

On 2016/10/15 23:39, Holger Mikolon wrote:
> On some PDF documents I created with pdflatex mupdf 1.8p4 segfaults
> in fz_process_path() (source/fitz/path.c)
> 
> The following upstream fix solves the issue for me. 
> http://git.ghostscript.com/?p=mupdf.git;a=commit;h=47078209046ed26ff7828a73d5d43afe08a6941d
> 
> Below is a proposed patch file "patch-source_fitz_path_c"
> based on the above upstream diff.
> 
> Holger
> ;-se
> 
> 
> $ cat /usr/ports/textproc/mupdf/patches/patch-source_fitz_path_c
> 
> Bug 696630: Avoid needless (and potentially wrong) dereference.
> 
> --- source/fitz/path.c.orig   Sat Nov 29 00:01:34 2014
> +++ source/fitz/path.cSat Nov 29 00:01:39 2014
> @@ -580,8 +580,8 @@
>  {
>   int i, k, cmd_len;
>   float x, y, sx, sy;
> - uint8_t *cmds = path->cmds;
> - float *coords = path->coords;
> + uint8_t *cmds;
> + float *coords;
>  
>   switch (path->packed)
>   {
> 

Re: [NEW] cbmc - Bounded Model Checker for C and C++

[Simon Mages]

Hi,

Am Samstag, 15. Oktober 2016 schrieb Jan Klemkow :

> Hi Simon,
>
> first, nice port.  I would like to test my own code with it.  But you
> missed the clang dependency.

Well, i thought that a new port will not be taged for a past release. If
this is wrong i can
add the dependency.


> Your port is new and don't need a
> REVISION.

Ok, will fix that.


> bsd,port.mk(5):
> REVISION
> Revision number of the current package.  Defaults to empty (very
> first package),...
>
> Why do you not fixed the missing RCS tags of the portscheck you made?

What should i fix there? I thought cvs will add
them on commit.


> Little nitpick: as far as I know its more common to use spaces between
> variables and their equals signs like in
> /usr/ports/infrastructure/templates/Makefile.template.

Yea, i can fix that.


> bye,
> Jan
>
> log:
>
> ## Entering big-int
> gmake  -C big-int
> gmake[1]: Entering directory
> '/usr/ports/pobj/cbmc-5.5/cbmc-cbmc-5.5/src/big-int'
> clang++ -c -MMD -MP -std=c++11  -Wall -O2  -o bigint-func.o
> bigint-func.cc
> gmake[1]: clang++: Command not found
> gmake[1]: *** [../common:183: bigint-func.o] Error 127
> gmake[1]: Leaving directory
> '/usr/ports/pobj/cbmc-5.5/cbmc-cbmc-5.5/src/big-int'
> gmake: *** [Makefile:50: big-int.dir] Error 2
> *** Error 2 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2671
> '/usr/ports/pobj/cbmc-5.5/.build_done')
> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:1884
> '/usr/ports/packages/amd64/all/cbmc-5.5p0.tgz')
> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2409
> '_internal-package')
> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2389
> 'package')
> *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:1901
> '/var/db/pkg/cbmc-5.5p0/+CONTENTS')
> *** Error 1 in /usr/ports/devel/cbmc
> (/usr/ports/infrastructure/mk/bsd.port.mk:2389 'install')
>
>
> On Fri, Oct 14, 2016 at 08:15:21PM +0200, Simon Mages wrote:
> > Hi,
> >
> > this is my first port, i hope everything is fine.
> >
> > A college of mine was using this tool to analyse some C code. I think
> this
> > tool is pretty interesting and can be put to good use for all sorts of
> programs
> > or parts of it.
> >
> > COMMENT=  Bounded Model Checker for C and C++ programs
> >
> > # cat pkg/DESCR
> > CBMC is a Bounded Model Checker for C and C++ programs. It supports C89,
> C99,
> > most of C11 and most compiler extensions provided by gcc and Visual
> Studio. It
> > also supports SystemC using Scoot. We have recently added experimental
> support
> > for Java Bytecode.
> >
> > CBMC verifies array bounds (buffer overflows), pointer safety,
> ex??cep??tions
> > and user-specified as??ser??tions. Furthermore, it can check C and C++
> for
> > consistency with other languages, such as Verilog. The verification is
> > performed by unwinding the loops in the program and passing the
> re??sul??ting
> > equation to a decision procedure.
> >
> > While CBMC is aimed for embedded software, it also supports dynamic
> memory
> > allocation using malloc and new.
> >
> > CBMC comes with a built-in solver for bit-vector formulas that is based
> on
> > MiniSat. As an alternative, CBMC has featured support for external SMT
> solvers
> > since version 3.3. The solvers we recommend are (in no particular order)
> > Boolector, MathSAT, Yices 2 and Z3. Note that these solvers need to be
> > installed separately and have different licensing conditions.
> >
> > # portcheck
> > Makefile does not have $OpenBSD$ RCS tag at the top
> > patches/patch-Makefiles does not have $OpenBSD$ RCS tag at the top
> > patches/patch-bigint-fix-includes does not have $OpenBSD$ RCS tag at
> the top
> > devel/cbmc
> >
> > Content of the archive:
> > devel/cbmc
> > devel/cbmc/Makefile
> > devel/cbmc/patches
> > devel/cbmc/patches/patch-Makefiles
> > devel/cbmc/patches/patch-bigint-fix-includes
> > devel/cbmc/distinfo
> > devel/cbmc/pkg
> > devel/cbmc/pkg/PLIST
> > devel/cbmc/pkg/DESCR
> >
> > BR
> > Simon
>
>
>

Re: [UPDATE] devel/p5-Coro to 6.511 + patch

[Stuart Henderson]

On 2016/10/15 15:26, Andrew Fresh wrote:
> Coro 6.511 includes some patches to support perl 5.24, but it's 
> apparently not enough, but fortunately someone at Debian has a patch for
> the other problem building.
> 
> This update also disables the CORO_JIT because it violates W^X, yay for
> a perl thing that does that I guess!

Hmm - can we take that as an indication that nobody is actually
using p5-Coro, and combining that with the reluctance of upstream to
follow perl updates, remove the port to save us trouble for future
updates?

> The current 6.49 port also needs to disable the JIT in order to work, so
> not entirely convinced anyone actually uses this port. 

Anyway the above is a separate question. So please zap the REVISION
line then it's OK sthen@.

Re: "pkg_info -S" can't get update signature, "unsigned package"

[Stuart Henderson]

On 2016/10/16 10:27, Andreas Kusalananda Kähäri wrote:
> Hi,
> 
> "pkg_info -S" used to show the "update signature" for a package, i.e.
> the "unique tag showing the package name, and the version number of
> every run time dependency and shared library used to build this
> package".
> 
> Since recently, this gives me an error due to the packages not being
> signed:
> 
> $ pkg_info -S /usr/ports/packages/amd64/all/ccache-3.3.2.tgz
> Error from file:/usr/ports/packages/amd64/all/
> unsigned package
> 
> This is regardless of the whether the package is built recently or a
> long time ago.
> 
> Do I need to sign my own packages to be able to get the update signature
> out of them?!

No, you can use "-D unsigned".

"pkg_info -S" can't get update signature, "unsigned package"

[Andreas Kusalananda Kähäri]

Hi,

"pkg_info -S" used to show the "update signature" for a package, i.e.
the "unique tag showing the package name, and the version number of
every run time dependency and shared library used to build this
package".

Since recently, this gives me an error due to the packages not being
signed:

$ pkg_info -S /usr/ports/packages/amd64/all/ccache-3.3.2.tgz
Error from file:/usr/ports/packages/amd64/all/
unsigned package

This is regardless of the whether the package is built recently or a
long time ago.

Do I need to sign my own packages to be able to get the update signature
out of them?!


Cheers,
Andreas

-- 
Andreas Kusalananda Kähäri
Bioinformatics Developer
NBIS, Uppsala University
http://www.nbis.se/


signature.asc
Description: PGP signature