Hi,
When using shotwell for full-frame (~40 MB) photos, recent updates
caused import of said files to barf and not import with mysterious
"Backing store not supported" errors in the log. Seems this error comes
from gdk-pixbuf2, which passes hard constraints to libjpeg regarding
memory
file
-DISTNAME = sqlite2mdoc-0.1.4
+DISTNAME = sqlite2mdoc-0.1.10
MAINTAINER = Kristaps Dzonsons
@@ -21,6 +21,6 @@ CONFIGURE_ARGS = PREFIX="${PREFIX}"
ALL_TARGET = sqlite2mdoc
-NO_TEST = Yes
+TEST_TARGET = regress
.include
Index
Now that we have pandoc on amd64, would it be OK to add a preformatted
manual for shellcheck? The diff adds a simple hash check to remind about
(though not enforce) updating it if the source file changes.
lowdown can build this decently and it's a bit smaller than pandoc:
% sed 1d
This is an update of misc/subsurface to version 4.9.7. Not much has
changed. I've been using it now for a little while and it works fine.
...except that it reliably segfaults on exit and has since 4.8.5.
Printing doesn't seem to work properly in some cases. Otherwise works
fine in planning
This patch updates slant to 0.0.23. The difference being that it uses
the newest openradtool, which uses the sqlbox library instead of ksql.
Index: Makefile
===
RCS file: /cvs/ports/sysutils/slant/Makefile,v
retrieving revision 1.5
AliceVision is a photogrammetric computer vision framework. The port
depends upon the last half-dozen or so submitted ports (coin, osi, clp,
flann, etc.). It takes a long time to build.
This is pretty heavily (but mechanically) patched because they unwisely
used variable names that are also
This updates eigen3 to the latest version. Beyond the trivial, it adds
BLAS as a dep and changes the fftw dep to fftw3. It looks like eigen2
support has been removed (it was deprecated in 3.2.1).
Needed for alicevision (upcoming).
? patches/patch-scripts_buildtests_in
?
This port is a programming library of geometric algorithms. I've
limited it to amd64 right now because it has a soup of cmake platform
checks that I'm not sure how to generalise. Tested on amd64 by looking
at mesh files with vorpaview.[1]
This requires the updated glfw sent before.
Generates
Trivial update of graphics/glfw to 3.3. Tested on amd64. (Am I
supposed to do anything with the SHARED_LIBS version?)
Needed for alicevision (upcoming).
Index: Makefile
===
RCS file: /cvs/ports/graphics/glfw/Makefile,v
retrieving
A COIN-OR (Computational Optimization Infrastructure for Operations
Research) project focussing on combinatorial optimisation for graphs and
networks. Configured to use glpk as the solver. Note that this doesn't
use the OSI solver because it'll try to use the CBC mixed-integer
solver, which
This is the COIN-OR (Computational Optimization Infrastructure for
Operations Research) linear programming solver, also as an OSI module.
Depends upon coinutils and osi, just sent out.
All tests run on amd64 and docs are installed (just doxygen rubbish but
still).
This is a requirement for
This is the COIN-OR (Computational Optimization Infrastructure for
Operations Research) open solver interface. It depends upon coinutils,
just sent out.
All tests run on amd64 and docs are generated by doxygen.
It doesn't bundle with any solvers of its own---this is just the
abstract classes.
These are utility classes for the many COIN-OR (Computational
Optimization Infrastructure for Operations Research) projects that will
be forthcoming.
All tests run (on amd64 at least) and documentation is installed. This
asks for fortran, but as noted in the port, I don't think it actually
uses
flann is a C++ solver for approximate nearest neighbours.
The tests don't work as they require downloading things that don't exist
any more and the docs are a PDF, so I left it off.
This is a requirement for alicevision (upcoming).
flann.tar.gz
Description: application/gzip
ceres-solver is a lightweight C++ template library for linear algebra.
It can use either openmp (which I don't think is supported?), tbb, or
c++11 threads. I go with c++11 because it seems best supported across
all architectures, though tbb works as tested. Only tested on amd64.
`make test`
>>> This updates subsurface to the current 4.9.3. Joint effort with Tracey
>>> Emery, thank you!
>>>
>>> The port-lib-depends-check asks for a x11/gtk+3,-guic RUN_DEPENDS as
>>> well, but I've left that out because this is a Qt application. I'm not
>>> sure if this is correct.
>>
>> You need to
This updates subsurface to the current 4.9.3. Joint effort with Tracey
Emery, thank you!
The port-lib-depends-check asks for a x11/gtk+3,-guic RUN_DEPENDS as
well, but I've left that out because this is a Qt application. I'm not
sure if this is correct.
So this has a really ugly (but otherwise
When running pkg-config for eopenssl11 (openssl-1.1.1d), the values for
eopenssl are printed instead. This fixes it.
Was:
% pkg-config --cflags eopenssl11
-I/usr/local/include/eopenssl
Now is:
% pkg-config --cflags eopenssl11
-I/usr/local/include/eopenssl11
?
Ping!
Forwarded Message
Subject: [UPDATE] misc/libdivecomputer
Date: Tue, 18 Jun 2019 18:04:17 +0200
From: Kristaps Dzonsons
To: ports@openbsd.org
This is just a REVISION update since I've located the source of the
removed AES files[1]. The license is appropriate
Ping!
Forwarded Message
Subject: Re: [UPDATE] misc/subsurface (downgrade)
Date: Tue, 18 Jun 2019 23:33:27 +0200
From: Kristaps Dzonsons
To: ports@openbsd.org
> This is an update of Subsurface to 4.8.6.
>
> I've removed the patches to libdivecomputer commentin
Is anybody a user of kdenlive?
I'm not able to actually add any clips to the timeline, especially if
they happen to "snap" to the beginning of the timeline (0 seconds).
Crashes immediately with no warning. If I am able to get a clip into
the timeline, it crashes when trying to work with them.
> One last nitpick: users and rc scripts are usually named after
> programs, not packages. So please use slant-collectd/slant_collectd.
> With that it's ok with me.
Enclosed is slant-0.0.21, which renames slant.rc to slant_collectd.rc.
It also fixes a bug found by schwarze@, hence the version
>> Enclosed is a port attempt for slant, https://kristaps.bsd.lv/slant.
>> Depends on openradtool, which was recently submitted by jturner.
>>
>> Previous attempts tried to be too smart about stopping the collector and
>> CGI script. This just YOLOs and jams in an upgraded database schema.
>> The
Enclosed is a port attempt for slant, https://kristaps.bsd.lv/slant.
Depends on openradtool, which was recently submitted by jturner.
Previous attempts tried to be too smart about stopping the collector and
CGI script. This just YOLOs and jams in an upgraded database schema.
The schema upgrade
> This is an update of Subsurface to 4.8.6.
>
> I've removed the patches to libdivecomputer commenting out AES as their
> files are "unlicensed", with a patch sitting upstream to put the actual
> license text in the files themselves.
>
> https://github.com/libdivecomputer/libdivecomputer/pulls
>
This is just a REVISION update since I've located the source of the
removed AES files[1]. The license is appropriate (the "unlicense").
I've notified upstream of this. The release version of libdivecomputer
has not changed since the last update.
Best,
Kristaps
[1]
This is an update of Subsurface to 4.8.6.
I've removed the patches to libdivecomputer commenting out AES as their
files are "unlicensed", with a patch sitting upstream to put the actual
license text in the files themselves.
https://github.com/libdivecomputer/libdivecomputer/pulls
One thing that
> There are two common methods used by programs that depend on database
> schema. One is probably best described as "YOLO", i.e. don't bother
> checking anything, not ideal but it is very common. The other is to have
> a schema version stored in the database and refuse to run if it's not as
>
> It's really not expected that a package runs something which changes
> packaged files or their permissions.
>
> Automatically running something which stops and starts daemons is
> also not expected.
If we go by the letter of the install process, where we're supposed to
"stop/disable all
> BUILD_DEPENDS=STEM->=0.6.9:www/kwebapp www/kcgi databases/ksql
> kcgi and ksql are RDEPS of kwebapp already, so this seems redundant.
>
> update-plist moves slant-cgi at the end of PLIST.
>
> port-lib-depends-check reports "curses tls z" as missing.
Enclosed fixes these.
However,
Hi, this is a re-post of slant, a remote system monitor for UNIX
machines (right now just OpenBSD, with more on the way).
It depends upon kwebapp 0.6.9, which was submitted by qbit a few days ago.
https://kristaps.bsd.lv/slant
Since I last posted it, it has gained a lot of client smarts
Enclosed is a patch to the cvsweb port to add a new flavour, "chroot".
This configures the script (and its default configuration) to run in the
/var/www chroot.
By default, the flavour is "unsafe", which uses the existing way of
referring to /var/www/whatever.
Best,
Kristaps
Index: Makefile
A few months ago, frei0r-plugins was posted to the list. By teaching
mlt to use these plugins, I can get kdenlive to do colour correction
properly. Enclosed is the patch to mlt so that it can use the proper
plugin values. (kdenlive needs no patch: it detects it via mlt.)
This requires the
>> I am setting this up atm. I will try to put together a readme we can include
>> for nginx / httpd configurations.
>
> Before/while you are at it:
>
> Kristaps: Slant doesnt't appear to support specifying port number in the
> URL in .slantrc, like:
>
> $ tail -n 3 ~/.slantrc
> servers
>
This is needed for slant, just posted. Version notes since the last update:
2018-10-09: version 0.6.3
Take unique statements into account when warning whether a search
statement might return multiple results. Also add some examples for
using unique.
Fix a crash in kwebapp-xliff(1).
Slant is a *simple* remote multi-system monitor currently only for
OpenBSD machines, though easily portable. Beyond the collector and a
CGI script, it also has a nice, customisable ncurses interface for
top-like monitoring of your hosts. Yay.
More information at https://kristaps.bsd.lv/slant.
>> This updates divecmd to 0.1.3.
>
> Given that this is very new software under very active development
> and that no other ports depend on it, i committed right away,
> to avoid trouble with the upcoming lock.
>
>> Attached is both the port and a patch of the contents.
>
> Only the diff was
This updates divecmd to 0.1.3. Attached is both the port and a patch of
the contents. Beyond the hash, this update includes ssrf2dcmd in its
install. This can be used for converting from Subsurface native XML
files. More update notes:
- Clarify that pressure sensor may refer to nonexistent
The attached port updates Subsurface from 4.8.1 to 4.8.2, released just
this month. The update consists of removing some patches that have been
merged into the mainline.
I've also attached a diff from the current one.
Upstream version notes:
This depends upon libdivecomputer, which hasn't been committed but is in
this list somewhere. It started as fork of libdivecomputer's dctool.
Mostly used for free diving.
Description:
divecmd is a suite of open source command-line UNIX tools to extract
and manipulate dives from dive computers.
> -WANTLIB += c
> +WANTLIB += c expat
Should note "# uses pledge()" above this line, no? All binaries use
"stdio rpath" dropping down to just "stdio".
(Grep for HAVE_PLEDGE in all main() functions for context.)
> Thanks for the feedback and pointers. Bringing in Kristaps into the
> discussion.
I don't have any extra plans for libdiff---nor is it likely to change as
a library. At least in terms of lowdown. I just figured a library for
easy diffing would be useful. So by all means push the library
> New tarball attached with the following changes:
>
> - Set USE_WXNEEDED, we use Qt5Webkit.
> - Regen WANTLIB
> - Use x11/qt5 module and drop CMAKE_PREFIX_PATH ...
> - Set MODGNU_AUTOCONF_DEPENDS and MODGNU_AUTOMAKE_DEPENDS by hand
> because we don't use CONFIGURE_STYLE=gnu witch set this
> Imported dives from DiveRite Nitek Q directly into Subsurface
> without issue. Will test an IR link to a Uwatec later. That may or
> may not work as I've never played with USB IR in OpenBSD.
It should work with it, though I've never used IR. From what I see,
libdivecomputer (both on its own
> Not sure that's entirely true - i got a very noticeable delay during
> first startup with your version that i didn't get with mine, and i
> assume that's related to some kind of network activity, though
> admittedly i did not investigate with tcpdump and/or ktrace.
Ingo,
Hm, that's
> Regarding libdivecomputer, i'm looking for OKs and for confirmation
> from kristaps@ for the attached version:
Works perfectly. Tested on amd64, was able to interface with the
library and pull down dives. Thank you so much, Ingo!
Now time to get these patches upstreamed...
Best,
Kristaps
Hi Ingo,
Enclosed is another iteration with the changes as specifically noted
below. The last port didn't compile with an existing grantlee and it
crashed on start---I think this had to do with the FB changes, but am
not sure.
The enclosed port builds, installs, and runs with all functionality
>> Hi Stuart and Kristaps,
>>
>> Stuart Henderson wrote on Wed, Aug 01, 2018 at 07:57:25PM +0100:
>>
>>> I'd never expect to find this in multimedia... misc seems
>>> perfect?
>>
>> I don't object to misc either and trust Stuart's judgement on this
>> one.
>>
>> But i think libdivecomputer
018
@@ -6,7 +6,7 @@ VER = 4.8.1
PKGNAME = subsurface-${VER}
DISTNAME = Subsurface-${VER}
-CATEGORIES = misc
+CATEGORIES = multimedia
HOMEPAGE = https://subsurface-divelog.org
MAINTAINER = Kristaps Dzonsons
@@ -57,7 +57,9 @@ pre-configure:
env AUTOCONF_VERSION=2.67 AUTOMAKE_VERSION=1.11
>> http://libdivecomputer.org/
>>
>> "Libdivecomputer is a cross-platform and open source library for
>> communication with dive computers from various manufacturers."
>>
>> Tested by linking against https://github.com/kristapsdz/divecmd
>> with a suunto d6i and using the native dctool(1).
>
> I
This is a port for Subsurface:
Description:
Subsurface can plan and track single- and multi-tank dives using
air, Nitrox or TriMix. It allows tracking of dive locations
including GPS coordinates (which can also conveniently be entered
using a map interface), logging of equipment used and names
This is a dependency for Subsurface, coming in my next e-mail. It's
different from libdivecomputer (also done, to be submitted later) in
that it has special magic to work with Subsurface. I'm not in a
position to solve the disparity between both projects.
sthen@ noted that this could be merged
Hi folks,
I'm working on a package for Subsurface, https://subsurface-divelog.org.
It now compiles fine (some pending stuff upstream for it) and I've been
using it for a while on OpenBSD by way of a modified version of its
baked-in build script.
The question is as follows. I think I can get
> This is another simple maintainer update of sblg from 0.4.13 to 0.4.14.
> The changes listed by the author are:
>
> "Allow for data-sblg-navtag and data-sblg-articletag to support an
> embedded ${sblg-get|xxx}. This is useful for having a navigation list
> (in -L or -C mode) deriving from the
etrieving revision 1.2
diff -u -p -r1.2 Makefile
--- Makefile 29 Jun 2018 18:11:44 -0000 1.2
+++ Makefile 30 Jun 2018 19:21:31 -
@@ -2,7 +2,7 @@
COMMENT = extract C reference manpages from SQLite3 header file
-DISTNAME = sqlite2mdoc-0.1.2
+DISTNAME = sqlite2mdoc-0.1.4
MAINTAINER
=
RCS file: /cvs/ports/converters/sqlite2mdoc/Makefile,v
retrieving revision 1.2
diff -u -p -r1.2 Makefile
--- Makefile 29 Jun 2018 18:11:44 - 1.2
+++ Makefile 30 Jun 2018 15:18:58 -
@@ -2,7 +2,7 @@
COMMENT = extract C reference manpages from SQL
>> and now that sqlite2mdoc is in, here's the diff for the integration part
>> into sqlite3 port. Note the less that 10% size increase in pkg size.
>>
>> 1953518 sqlite3-3.24.0.tgz
>> 2187305 sqlite3-3.24.0p0.tgz
>
> The unpacked size isn't too much bigger either. OK sthen
>
> Kristaps: one nit
>> Rather than relying on github's on-the-fly-generated tarballs (which are
>> subject to
>> change) could you upload a tar.gz as a "release asset" please? Then it would
>> use
>> something like this
>>
>> MASTER_SITES =
>>
Attached is a port for sqlite2mdoc.
pkg/DESCR:
This utility accepts an SQLite3 header file sqlite3.h and produces a
set of decently well-formed mdoc(7) files documenting the C API.
These will be roughly equivalent to the C-language Interface
Specification for SQLite3.
HOMEPAGE:
>> The following patch updates kcaldav to 0.1.6, making it compatible
>> with kcgi-0.10.2.
>>
>> In addition to make kcaldav compatible with kcgi, Krtisaps accepted
>> many of patches from the 0.1.4 port, and updated build to use gmake
>> like kcgi.
>>
>> As a result, you'll find a lot of patch
Hello,
Has anybody been able to actually edit video with multimedia/lives?
I've never been able to get past loading a video, at which point the
application crashes without any debug message except the now-familiar
"lives has crashed" message. When I run the application in gdb, I get a
segv here:
I find it hard to believe there are no OpenBSD divers. libdivecomputer
is very handy:
http://libdivecomputer.org/
"Libdivecomputer is a cross-platform and open source library for
communication with dive computers from various manufacturers."
Tested by linking against
> MAKE_FLAGS and ALL_TARGET lines are not needed. Also, you could zap
> some extra empty lines, but that's purely style nit. :)
Vadim,
If MAKE_FLAGS wasn't specified with PREFIX, it wouldn't install
correctly into the /usr/ports/pobj directory. (The fake-amd64 part, on
my machine, was
This is my first port. It didn't need much coersion to build. I use it
to get markdown PDFs on openbsd without groff, though the groff ones do
look better (via -ms).
https://kristaps.bsd.lv/lowdown
"lowdown is just another Markdown translator. It can output traditional
HTML or a document for
>> sblg is a utility for creating static blogs: it knits together articles and
>> templates, generating static HTML files and Atom feeds. It's built for use
>> with make(1), as blogs depend upon articles, etc. No markdown, no "CMS", no
>> CGI, no PHP. Just a simple tool for pulling data from
>> Good question. I figured that if the user has - explicitely - chosen
>> rdrand, then the choice should be respected, even if it's a bit, ahem,
>> stupid.
>>
>> I can change the diff if that's a problem. In the end, this should be
>> discussed with upstream.
>
> Same proposal, diff below for
Hi,
json-c has some... peculiarities, such as how it seeds the random number
generator for its hashing function.
https://github.com/json-c/json-c/blob/master/random_seed.c
I'm not sure how responsive the project is, hence posting here, but the
existing code causes /dev/urandom to be touched,
The pod2mdoc utility is a converter from POD into mdoc. It's meant to
operate like pod2man.
ok schwarze@
The docbook2mdoc utility is a converter from DocBook into mdoc.
ok schwarze@
Both written by our very own kristaps@. Both are still experimental, but
useful to have in ports for
On 01/16/12 00:06, Anders Langworthy wrote:
Hi. First attempt at a port update, observations requested. Works for
me on i386. The license is actually BSD without the clauses (see
autossh.c); the new version uses autoconf. I wasn't sure how to make it
compile with SEPARATE_BUILD, and the
The non-english manpages are broken, all the non-ascii characters get
stripped when you view them. These should either be fixed or disabled.
Check output in e.g. man -m /usr/local/man/es scrotwm.
You probably want to build them with nroff for now, as was done with
the (commented-out) Russian
Hi,
Does anybody use edict around here? From their website:
NOTE: As of Mon, Aug 23, 2010, edictionary is broken. I'm actively
working to fix it. Please check back in a few days.
Eh, I think a few days has passed... anyway word lookup fails
unilaterally for me.
(Many of the project
On 02/10/2011 10:22, Marc Espie wrote:
On Sun, Oct 02, 2011 at 12:28:09AM +0200, Ingo Schwarze wrote:
- The naming is terrible, one is not even versioned, but i will talk
to Kristaps to fix that before the next mandoc-tools release.
Bad Kristaps. ;-) Obviously not used to work as an
On 03/08/2011 01:24, Amit Kulkarni wrote:
I did a make package inside math/lapack and before package creation,
there is a bunch of errors.
No subject found in /tmp/manpage.jSOMEJUNK//usr/local/man/cat3f/claqr1.0
No subject found in /tmp/manpage.jSOMEJUNK//usr/local/man/cat3f/claqr2.0
No subject
Hello, people. I'm curious if anybody's interested in a port for the
sysjail system[1]. sysjail is a jail implementation for OpenBSD
(support for NetBSD has been deprecated) based upon systrace. I'd have
thought the system phased out with the published deficiencies of
systrace[2], but judging
74 matches
Mail list logo