Kyle R W Milz k...@getaddrinfo.net wrote:
With RC4 being replaced by ChaCha in libc's arc4random, is it good
practice to stop using it?
I locally disabled the RC4 ciphersuites in Firefox for a bit, but
you pretty quickly run into sites that will not offer any other
ciphers.
--
Christian
math/grpn looks abandoned by its upstream. Jens Getreu has put it
on GitHub and incorporated the various patches that are floating
around.
Do we want to update to this?
(I'm a bit hesitant. How many projects have been forked on GitHub
and then promptly been abandoned again?)
Anyway, here's a
I think everybody who proposes a bitcoin port should consider whether
something that seriously deals with people's money doesn't warrant
special auditing and whether they are prepared to invest that work.
I'm very uncomfortable with shipping our usual it seems to build
and run for me packages and
This switches misc/ytree from the bizarrely obsolete re_comp(3) to
POSIX regcomp(3) and thus removes the need for linking with libcompat.
ok?
Index: Makefile
===
RCS file: /cvs/ports/misc/ytree/Makefile,v
retrieving revision 1.15
Theo wants to remove libcompat.a and I agree.
For those who don't know, libcompat is a small static library with
compatibility stubs for a handful of long obsolete, forgotten APIs from
the 1980s:
Modern replacement:
* gtty(3), stty(3) - tcgetattr(3)
Kirill Bychkov ki...@linklevel.net wrote:
All this story started from the problem with dead MASTER_SITES. After some
searching I've found that file is hosted on SF [0], but with another name -
logcheck with the same size and checksum.
I've decided to change it's name from logsentry to
Mysteriously, PyPy appears to link with libcompat during the build.
From a brief look, I couldn't even figure out where this is coming
from. Anybody know what this is about? It's hard to imagine that
PyPy would actually need libcompat.
[translation:info] [PyPy 2.1.0]
[platform:msg] Set platform
Christian Weisgerber na...@mips.inka.de wrote:
Modern replacement:
* gtty(3), stty(3) - tcgetattr(3) etc.
* re_comp(3), re_exec(3)- regcomp(3) etc.
* ftime(3) - gettimeofday(2)
* cuserid(3
Giovanni Bechis giova...@bigio.snb.it wrote:
pre-configure:
${SUBST_CMD} ${WRKSRC}/Makefile
+ @perl -pi -e 's/@\$$\(/\$$\(/g' ${WRKSRC}/Makefile \
+ ${WRKSRC}/bk/Makefile ${WRKSRC}/iniparser-2.17/Makefile
Hmm. You still have chunks to the same effect in
Christian Weisgerber na...@mips.inka.de wrote:
Mysteriously, PyPy appears to link with libcompat during the build.
From a brief look, I couldn't even figure out where this is coming
from. Anybody know what this is about? It's hard to imagine that
PyPy would actually need libcompat
David Coppa dco...@gmail.com wrote:
The diff below fixes some ui glitches in the qt interface, to make
it more like the gtk one.
Comments?
What does this actually fix? Sounds like a suggestion for upstream,
rather than the sort of things we fix in a port.
--
Christian naddy Weisgerber
Christian Weisgerber na...@mips.inka.de wrote:
sthen@ has (been) volunteered to find all ports that still require
libcompat and will provide a list in the next few days.
Combining various efforts, this is the list:
astro/xphoonftime
cad/spice
Christian Weisgerber na...@mips.inka.de wrote:
games/xbl
java/tanukiwrapper
libcompat is not used. Port fixed.
databases/pgpool
security/nessus/core
security/nessus/libnasl
libcompat is picked up by configure, but not used.
No change required.
In infrastructure/db/config.site, we
Replace cuserid() with getlogin() and get rid of -lcompat.
ok?
Index: Makefile
===
RCS file: /cvs/ports/games/late/Makefile,v
retrieving revision 1.14
diff -u -p -r1.14 Makefile
--- Makefile11 Mar 2013 11:07:39 - 1.14
Jérémie Courrèges-Anglas:
I have diffs for metamail, elm, and gopher.
metamail is a candidate for the most terrible code in the ports
tree. It can't be saved. Considering that e-mail is a primary
attack surface, I think we should kill it. Unfortunately, there
are even ports that depend on
Jérémie Courrèges-Anglas:
Use gopher_re_{comp,exec} #defines to deal with the various #ifdefs.
The port uses
AC_CHECK_LIB(compat, re_comp, REGEXLIBS=$REGEXLIBS -lcompat)
so it will link in libcompat if present, but it shouldn't matter when we
get rid of it. POSIX regexps are checked for
Jérémie Courrèges-Anglas:
Here's the metamail diff.
I provide a knob to use tc[gs]etattr if desired. I didn't use
cfmakeraw() since it seems that this one hasn't made its way into POSIX.
While here I replaced the gets() calls with fgets but if that's not
desired I can remove those parts.
Landry Breuil lan...@rhaalovely.net wrote:
For the .1x - AfterStep uses the format though if that is unwanted I can
modify the upstream source and use .1 as extension.
I have no idea how our man handles this -
It handles it just fine. See x11/afterstep and a number of other
ports that
What actually uses textproc/sgmlformat?
This package contains SGML formatting tools that were once a part
of the base FreeBSD distribution. This port should be regarded
as a temporary solution to formatting SGML files until the
wrinkles in jade, a DSSSL processor, get ironed out.
Jérémie Courrèges-Anglas:
Third iteration, with input from Antoine (completely unrelated and bogus
diff for patch-src_config_h)
I saw that chunk and thought it was correct!
lynx also references /etc/mailcap, so it would make sense if this
file was considered part of the base system. I don't
Jérémie Courrèges-Anglas:
If no one chimes in, ok to delete it. This being said, here's a
compile-tested (and may^Wprobably incomplete) attempt to convert it to
regex(3).
+@@ -392,7 +392,7 @@ static void build_ts(char *gi, char* cp)
+ tok = Split(cp, i, S_STRDUP);
+ T.var_RE_name
Jérémie Courrèges-Anglas:
Another config.site definition we'll need to change (yes, I'm keeping
notes): ac_cv_func_re_comp
Drop ac_cv_lib_compat_main too and databases/pgpool should be happy.
Yes. My current list:
ac_cv_lib_compat_main
ac_cv_func_re_comp
ac_cv_header_sgtty_h
--
Jérémie Courrèges-Anglas:
Diff for security/cfs. Please close your eyes before running make.
I can't run make because cfs only builds on obsolete platforms. :-
+--- getpass.c.orig Wed Dec 3 22:51:34 1997
getpass.cSat Nov 30 02:21:29 2013
+@@ -45,7 +45,7 @@
+ #include cfs.h
Jérémie Courrèges-Anglas j...@wxcvbn.org wrote:
If no one chimes in, ok to delete it. This being said, here's a
compile-tested (and may^Wprobably incomplete) attempt to convert it to
regex(3).
FWIW, here's my stab at the regexp to regex conversion.
--- instant/tables.c.orig 1996-09-08
Brad Smith b...@comstyle.com wrote:
A bug fix update to Audacious 3.4.2.
OK?
No. This now picks up SDL2 over SDL1 if the former is installed.
--
Christian naddy Weisgerber na...@mips.inka.de
On a tangent, has anybody given thought to shrinking the dependency
list of audacious-plugins like we did with MPlayer? Doesn't the
FFmpeg input plugin cover various audio formats, so we won't need
the corresponding individual plugins? (MP3, Vorbis, AAC, FLAC, ...)
--
Christian naddy Weisgerber
This replaces the one use of ftime() on BSD with gettimeofday() in
cad/spice. gettimeofday() was already available at the time that
code was written in the 1980s. I've followed the existing style
and not used #elif.
This also removes the useless inclusion of sgtty.h, which will
go away, too.
Here's a diff to replace ftime() with gettimeofday() in emulators/fs-uae.
Unfortunately, this is strictly a local diff. Upstream purposely
uses ftime() for portability between Unix and MS-Windows. There's
a variety of ways to handle this--sprinkle some #ifdef's, add a
gettimeofday()
Landry Breuil lan...@rhaalovely.net wrote:
- you could aswell use
${INSTALL_DATA} ${WRKSRC}/{README,COPYING,TODO} ${PREFIX}/share/doc/wmconfig
instead of cd ${WRKSRC} (but that's really cosmetic)
Speaking of really cosmetic, I kinda dislike { } brace expansion
because it is a non-POSIX
These will have to change once libcompat is removed:
* existence of re_comp()
* existence of sgtty.h
* main() with -lcompat, i.e., existence of libcompat
I think I have done an exhaustive search and nothing else in
config.site is affected, but it wouldn't hurt for somebody to
double-check.
Does
Kirill Bychkov:
Shouldn't OpenBSD be mentioned here in curl/curl.h, since we have
sys/select.h:
/* HP-UX systems version 9, 10 and 11 lack sys/select.h and so does oldish
libc5-based Linux systems. Only include it on systems that are known to
require it! */
#if defined(_AIX) ||
Kirill Bychkov:
Shouldn't OpenBSD be mentioned here in curl/curl.h, since we have
sys/select.h:
/* HP-UX systems version 9, 10 and 11 lack sys/select.h and so does oldish
libc5-based Linux systems. Only include it on systems that are known to
require it! */
#if defined(_AIX) ||
Libcompat has been removed from base. I should point out that the
old files do not magically disappear on upgraded machines, you need
to delete them manually:
# rm /usr/lib/libcompat* /usr/include/{re_comp,regexp}.h
# rm /usr/share/man/man3/{re_comp,re_exec,regexp}.3
sgtty.h and sys/timeb.h
Does anybody still use cad/chipmunk or cad/spice?
Both are incredibly crufty ports whose main code base dates back
to the 1980s, was barely carried over into the 1990s, and has been
on minimum life support since. These are the sort of ports we
always run into when we clean up something, and we
This welcomes textproc/ispell in the 1990s and switches it from
sgtty ioctls to POSIX termios.
ok?
Index: Makefile
===
RCS file: /cvs/ports/textproc/ispell/Makefile,v
retrieving revision 1.59
diff -u -p -r1.59 Makefile
--- Makefile
Christian Weisgerber na...@mips.inka.de wrote:
sgtty.h and sys/timeb.h have been temporarily left in place, but
these will go next. Below is a list of ports that need fixing.
Some more are likely to show up once these are done. I could use
some assistance there.
astro/sunclock
Brian Callahan bcal...@devio.us wrote:
Here's a fix for ee after the Libcompat fallout. What this patch does is
use our curses instead of ee's included subset of curses.
FreeBSD, which has ee in base, also builds it against the system
curses.
+@@ -180,8 +185,8 @@ then
+ TARGET=curses
Although it calls ftime() only on WIN32, games/csmash uses struct timeb
internally to pass around time values. I've copied our definition of
timeb into the central header file. Acceptable?
Other cleanup:
* Don't patch away the SDL header subdirectory and then add it
back to the include path.
Ian Darwin i...@darwinsys.com wrote:
--- Makefile 10 Jul 2013 14:40:29 - 1.1.1.1
+++ Makefile 7 Dec 2013 19:46:30 -
@@ -7,6 +7,7 @@ COMMENT= Tanuki Java Service Wrapper
SVN_REV= 1780
DISTNAME=wrapper-code-${SVN_REV}-trunk
PKGNAME=
This switches the editors/beav port from old 4.3BSD tty ioctls to
termios. While there, I also did a general overhaul of the port:
* Use the latest (2004) Debian version as upstream.
* Improved description, from Debian.
* Miscellaneous clean-up.
I don't know if Kevin still wants to be listed as
Ian Darwin i...@darwinsys.com wrote:
Found a plugin that plays all the m4a's I had lying around.
Port tarball at http://darwinsys.com/tmp/audio_xmms-mp4-port.tgz
How does this compare with faad-xmms? ;-)
--
Christian naddy Weisgerber na...@mips.inka.de
Ian Darwin i...@darwinsys.com wrote:
Port tarball at http://darwinsys.com/tmp/audio_xmms-mp4-port.tgz
How does this compare with faad-xmms? ;-)
Meh
(does an rm -r and crawls back under rock)
To be fair, I didn't remember that we had faad-xmms right away
either.
Ian has suggested to
Failures in the latest amd64 bulk build, very briefly:
math/octave undefined reference to `__guard'
textproc/redland-bindings,-python swig: -php4 is no longer supported
mail/zarafa/zarafa 'SWIG_From_long' was not declared in this scope
www/py-wtforms
Stuart Henderson st...@openbsd.org wrote:
Not sure about cause yet, some may be PIE-related too:
These are not:
lang/gcc/4.8: gcj: fatal error: can't specify '-D' without '--main'
Sporadic build failure; also happens on amd64.
mail/zarafa/zarafa: 'SWIG_From_long' was not declared in this
Stuart Henderson st...@openbsd.org wrote:
lang/nhc98: segfaults in build
NOT_FOR_ARCHS= ${LP64_ARCHS} powerpc
BROKEN-hppa=Segfault during build since the PIE switch
Do we need this for bootstrapping or reference purposes or can we
just remove it?
lang/petite-chez: undefined reference to
Failures in the latest amd64 bulk build:
textproc/redland-bindings swig: -php4 is no longer supported
mail/zarafa/zarafa 'SWIG_From_long' was not declared in this scope
x11/py-wxPython ld: cannot find -lwx_gtk2_media
Logs on request.
--
Christian naddy Weisgerber
Juan Francisco Cantero Hurtado i...@juanfra.info wrote:
lang/petite-chez: undefined reference to `__guard'
ONLY_FOR_ARCHS =i386
Hmm, FreeBSD has a port of 8.4 that at least adds amd64.
The license of Petite Chez is weird and the port is very outdated. We
have better
This is the fix for the null pointer dereference (CVE-2013-6954)
for OPENBSD_5_4, backported from png 1.6.8.
I don't have a 5.4-stable system at hand to actually test this
there.
OK?
OPENBSD_5_3 has 1.5.10, which is not affected according to the libpng
homepage.
Index: Makefile
Gleydson Soares:
here is an update for ggrep-2.16.
works fine for me @amd64
Okay by me with these further changes:
(gperf isn't used by the build, as far as I can tell.
Remove a no longer relevant comment.)
--- Makefile.orig Fri Jan 3 21:50:49 2014
+++ MakefileFri Jan 3 22:08:10
Matthieu has a FreeType update (2.5.2) ready that will include a
reorganization of the header files:
This large patch changes the header file directory layout from
`INCLUDE/freetype/...' to `INCLUDE/...', effectively removing one
level. Since the file `ft2build.h' is also located in
Landry Breuil:
games/renpy
/usr/local/include/GL/glew.h:1188:24: error: GL/glu.h: No such file or
directory
This fails because -I${X11BASE}/include is now missing.
graphics/py-Imaging
graphics/py-Pillow
Fixed, but now
graphics/py-matplotlib
fails, again because -I${X11BASE}/include
David Coppa:
Attached you have the right diff, they're three commits from upstream:
http://cmake.org/gitweb?p=cmake.git;a=commit;h=7435ae7c97590a0713d86b63add761d5e1607ec1
http://cmake.org/gitweb?p=cmake.git;a=commit;h=bc49d820a276243324003bc34d92a69194938adf
David Coppa:
These still fail:
graphics/darktable
graphics/openscenegraph
lang/io
x11/ogre
These are probably using a local FindFreetype.cmake module file.
Indeed they are.
So, we have two ways here: we can remove them at pre-configure stage
so the port will use the
Christian Weisgerber na...@mips.inka.de wrote:
These are probably using a local FindFreetype.cmake module file.
Indeed they are.
Add another one:
geo/mapserver
graphics/darktable
graphics/openscenegraph
lang/io
x11/ogre
--
Christian naddy Weisgerber na
Alexis de BRUYN alexis.mailingl...@de-bruyn.fr wrote:
Short story: the latest package snapshost (i386) is signed with
55pkg.pub, but the @signer in +CONTENTS is 54pkg.
Same problem here too :
OpenBSD alex.test 5.5 GENERIC.MP#8 amd64
I fixed the amd64 packages last night (CET). The
Pascal Stumpf pascal.stu...@cubes.de wrote:
Update to 20131230 snapshot, with UTF-8 support.
Doesn't this need iconv?
--
Christian naddy Weisgerber na...@mips.inka.de
Pascal Stumpf:
Or, as aja@ prefers, just build with iconv per default. It does add a
dependency to an otherwise dependency-less port, for a feature that's
not needed by everybody, so I'm asking if anybody prefers the FLAVOR
approach.
I was going to propose this--no FLAVOR, just build with
games/pioneers picks up libnotify if installed. There's not switch to
disable it, so I just added it as a dependency.
port-lib-depends-check suggested pretty much a complete change of
WANTLIB. Does this look plausible?
(I don't play pioneers or really care, but it broke in my latest bulk
This fixes the build of shells/nsh after the n2k14 network header
changes. While there, I belatedly corrected a number of type
mismatches due to the time_t change.
ok?
Index: Makefile
===
RCS file: /cvs/ports/shells/nsh/Makefile,v
The devel/readline port clashes with the old libreadline in base.
Yes, the library is named ereadline, but the header files still
collide.
devel/R fails with
../../lib/libR.so.1.1: undefined reference to `rl_sort_completion_matches'
when devel/readline is around:
#if RL_READLINE_VERSION =
Pascal Stumpf:
The problem is that MASTER_SITES will let you download any
CGAL-x.y.tar.gz and give you 3.8. If you want anything else, you have
to change the number at the end of the URL. I can prepare an update to
4.3 for after unlock, but for now, I think correcting the version and
The ports tree is unlocked.
--
Christian naddy Weisgerber na...@mips.inka.de
I suggest to delete archivers/bzip.
bzip is used by nobody. It was patent-encumbered and, from today's
POV, superseded immediately by bzip2. The upstream distfile is
long gone.
ok?
--
Christian naddy Weisgerber na...@mips.inka.de
This removes TCP Wrapper support from databases/openldap and
databases/openldap23.
OK?
Index: openldap/Makefile
===
RCS file: /cvs/ports/databases/openldap/Makefile,v
retrieving revision 1.131
diff -u -p -r1.131 Makefile
---
On 2014-03-24, Donovan Watteau tso...@gmail.com wrote:
Attached is a port for flac123. Tested on loongson.
I'll have a look.
--
Christian naddy Weisgerber na...@mips.inka.de
On 2014-03-24, Donovan Watteau tso...@gmail.com wrote:
Attached is a port for flac123. Tested on loongson.
I have some minor nits for the Makefile, but I think the man page
poses a problem. That file doesn't have a license, we can't just
include it.
--
Christian naddy Weisgerber
CVSROOT: /cvs
Module name: ports
Changes by: na...@cvs.openbsd.org 2014/03/27 06:07:59
Modified files:
security/ppgen : Makefile
sysutils/ansible: Makefile
Log message:
Drop USE_GROFF since groff and mandoc produce identical output.
This was the result of a
sthen's quick hack to unbreak gcc 4.8 after the gets() removal doesn't
work for me. I don't see how it could, since _GLIBCXX_HAVE_GETS is
defined *after* the #undef the substitution introduces.
With the diff below instead, gcc 4.8 builds for me on amd64.
Anybody want to test another arch?
OK?
security/samhain has been broken since the removal of sum(1).
The problem is the c_random.sh script, which returns a 16-bit random
number in decimal. Unless I'm missing something obvious, this is
surprisingly difficult to do in a semi-portable fashion at the shell
level.
I've settled on
Stuart Henderson:
I wondered about just doing $((RANDOM+RANDOM)) but don't know the
pros and cons of that approach - it's not exactly arc4random_uniform..
I wanted to stay in the spirit of portability. RANDOM is a ksh
extension and difficult to check for reliably.
--
Christian naddy
Here are patches to move
security/ophcrack
security/chntpw
security/samdump2
from the ancient des_old.h to the des.h API. These three ports use
more or less the same code, so the patches look similar as well.
Basically this just manually expands the defines in openssl/des_old.h,
so it
This switches mail/mixmaster from the ancient des_ API to DES_.
ok?
Index: patches/patch-Src_crypto_c
===
RCS file: patches/patch-Src_crypto_c
diff -N patches/patch-Src_crypto_c
--- /dev/null 1 Jan 1970 00:00:00 -
+++
Only two des_ heads to cut off...
ok?
Index: patches/patch-hydra-oracle-listener_c
===
RCS file: patches/patch-hydra-oracle-listener_c
diff -N patches/patch-hydra-oracle-listener_c
--- /dev/null 1 Jan 1970 00:00:00 -
+++
This one has ifdefs to build with libcrypto or libdes(?) instead.
ok?
Index: patches/patch-src_ntlm_auth_c
===
RCS file: patches/patch-src_ntlm_auth_c
diff -N patches/patch-src_ntlm_auth_c
--- /dev/null 1 Jan 1970 00:00:00 -
Another des_ - DES_ conversion.
Index: patches/patch-libsrc_Wi_xmlenc_algos_c
===
RCS file: patches/patch-libsrc_Wi_xmlenc_algos_c
diff -N patches/patch-libsrc_Wi_xmlenc_algos_c
--- /dev/null 1 Jan 1970 00:00:00 -
+++
Another des_ to DES_ conversion. I'm bound to have fat-fingered
something in one of these. ;-)
Index: patches/patch-src_dissectors_ec_ssh_c
===
RCS file: patches/patch-src_dissectors_ec_ssh_c
diff -N
net/netatalk uses a prefix-less API even older than des_ and found at
the end of openssl/des_old.h.
configure still checks for -ldes, but this looks like a vestige given
that the automake glue refers to OpenSSL, so I decided that it wasn't
worth trying to maintain compatibility with the 1990s.
On 2014-04-19, Stuart Henderson st...@openbsd.org wrote:
On 2014/04/19 13:02, Markus Lude wrote:
during last days netpbm fails to build here on sparc64.
Confirmed.
I suspect this probably needs -fPIC instead of -fpic
Yes, I'll fix this.
--
Christian naddy Weisgerber
I don't know how far ahead or behind sthen@'s and espie@'s builds
are, but here's some of the current ports breakage I'm seeing:
llvm update:
www/mozilla-firefox
www/seamonkey
RAND_egd(_bytes) removal:
graphics/dcmtk
lang/pypy
lang/rubinius
net/papyon
--
Christian naddy Weisgerber
On 2014-04-09, Landry Breuil lan...@rhaalovely.net wrote:
If you already tried beta3, please test the beta5 in the git repo now,
as it should fix the spellchecker feature (there are still some issues
but being worked on). Report success or failure too ;)
Beta8 worked fine for me, too, but
Here's the list of ports that failed to build during the amd64
package build started on 2014-04-22:
databases/hs-HDBC-postgresqlMissing C libraries: ssl, com_err
databases/hs-postgresql-libpq Missing C libraries: ssl, asn1, krb5, ...
geo/postgis json/json.h: No such file
Jérémie Courrèges-Anglas:
RAND_egd(_bytes) removal:
graphics/dcmtk
This fixes the build for me.
ok?
-snip-
Hmm, this still has RAND_load_file... ... OMG, there are seed file
tentacles everywhere.
I guess this is okay. :-/
--
Christian naddy Weisgerber
I'm creating a port for x2goclient (http://www.x2go.org/) but I don't
want to build the browser plugin and the documentation, only the heavy
client. So instead of the regular make, I have to launch make
build_client.
I tried to add MAKE_FLAGS= build_client or change to CONFIGURE_STYLE=
make
Now that we have AI_ADDRCONFIG, these patches can just be removed:
audio/gogglesmm/patches/patch-src_ap_http_cpp
benchmarks/netperf/patches/patch-src_netlib_c
benchmarks/netperf/patches/patch-src_nettest_bsd_c
graphics/gimp/stable/patches/patch-plug-ins_script-fu_script-fu-server_c
Here's the list of ports that failed to build during the amd64
package build started on 2014-04-28:
devel/arm-elf/gcc
configure: error: cannot compute suffix of object files: cannot compile
telephony/asterisk,imap
checking for mandatory modules: IMAP_TK PORTAUDIO FREETDS... fail
Build failures from the amd64 bulk build started on 2014-05-13:
devel/arm-elf/gcc
gmake[4]: *** [addsf3.o] Segmentation fault (core dumped)
security/py-M2Crypto
FAIL: test_load_save (tests.test_rand.RandTestCase)
x11/kde4/l10n/pt
file INSTALL cannot find
Did we somewhere along the way lose the ability to pass through
AC3/DTS audio over S/PDIF? It makes no difference whether sndiod(1)
is running, and it's the same for mplayer-20140106p3 and mplayer-20140412.
$ mplayer -afm hwac3 dvd.vob
[...]
On 2014-05-16, Christian Weisgerber na...@mips.inka.de wrote:
My ability to test this will likely go away over the next few days,
so if nobody else cares... *shrug*
... aaand I'm all out of boxes that have azalia(4) and S/PDIF outputs.
(AC3 passthrough has never worked with cmpci(4
On 2014-05-17, Sergey Bronnikov este...@gmail.com wrote:
$ cat audio/opusfile/pkg/DESCR
Opusfile provides application developers with a high-level API for
decoding and seeking in .opus files.
Tested on amd64, -current.
I have an interest in this, I'll take a look.
--
Christian naddy
As landry@ discovered, the bzgrep script contains a bashism. I've
lifted a portable replacement from a newer version of GNU gzip's
zgrep script.
ok?
Index: Makefile
===
RCS file: /cvs/ports/archivers/bzip2/Makefile,v
retrieving
+VER= 2.83
DISTNAME= transmission-${VER}
PKGNAME-main= transmission-${VER}
PKGNAME-gtk= transmission-gtk-${VER}
@@ -17,7 +14,7 @@ HOMEPAGE= http://www.transmissionbt.com/
MAINTAINER=Christian Weisgerber na...@openbsd.org
-# GPLv2
+# GPLv2+
PERMIT_PACKAGE_CDROM= Yes
Here's a port for MediaInfo 0.7.69:
| MediaInfo is a convenient unified display of the most relevant technical
| and tag data for video and audio files.
This is the CLI version only. As far as I understand, the GUI
version just presents the same information in a window. Not worth
the effort.
On 2014-06-04, Stuart Henderson st...@openbsd.org wrote:
Missing NO_TEST.
I've now fixed this in my copy.
I would prefer to fix line-endings in the files which are patched,
Hmpf... I see the point. Fixed.
Personally I'd use DISTNAME/EXTRACT_SUFX rather than setting
DISTFILES, either way
On 2014-06-05, James Turner ja...@calminferno.net wrote:
And to clarify a bit, in 0.18 the mp4ff plugin was removed which
provided m4a support from audio/faad so it appears ffmpeg is the only
option now for m4a support.
That's a pity because ffmpeg is a heavy dependency. That said, if
you
On 2014-06-07, Stuart Henderson st...@cvs.openbsd.org wrote:
CVSROOT: /cvs
Module name: ports
Changes by: st...@cvs.openbsd.org 2014/06/07 14:20:42
Modified files:
devel/arm-elf/gcc: Makefile
Log message:
mark BROKEN-i386; segfaults during build
I think this is
On 2014-06-14, Andrew Fresh and...@afresh1.com wrote:
None of the architectures I checked had any content in the existing .bs
files, so those files could be removed from the PLISTs at any time. I
have it on good authority that Debian already prunes them in their
packaging.
As does FreeBSD
On 2014-06-20, Bob Beck b...@obtuse.com wrote:
Anyone care to have a peek at the cvsync port and and try to fix the
client for commitid?
CVSync doesn't support commitid and adding it is not trivial.
jcs@'s patches haphazardly added commitid support to the RCS parser
and
1.0 # 6.2
+DISTNAME= gmp-6.0.0
+SHARED_LIBS += gmp 10.0 # 12.0
+SHARED_LIBS += gmpxx2.0 # 8.0
CATEGORIES=devel math
-HOMEPAGE= http://gmplib.org/
+HOMEPAGE= https://gmplib.org/
MAINTAINER=Christian Weisgerber na
Maintenance update of graphics/jpeg to 9a.
Principal change: Add support for wide gamut color spaces.
Upstream keeps tweaking the boolean handling, but this still blows
up with most C++ code, so we revert it again to the status quo from
version 8.
I started a bulk build of ports in jpeg's
Why exactly do we run the fake step as root?
(Hint: FreeBSD's corresponding stage infrastructure doesn't.)
Because ports want to install with install -o root -g bin? But
that's only because we tell them to. We pass those flags to configure
etc. We just need to stop doing this.
Because some
701 - 800 of 5525 matches
Mail list logo
