Re: BLENDER 2.79

[Marcin Cieslak]

On Sat, 18 Apr 2020, Adam Weinberger wrote:


On Sat, Apr 18, 2020 at 2:31 PM Tomasz CEDRO  wrote:




BGE is gone and done, and in most cases FreeBSD does not keep old
versions of ports around, and that's especially true for massive and
complex projects like Blender.


But I think we keep different port version around on major API changes
for a while. Not to mention things like node that can have up to 4 versions
in the ports tree.

Whatever is the future of BGE, one should simply not lose the functionality
by doing "pkg upgrade". Not sure this even made into UPDATING

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: python 2.7 marked as deprecated and EOL while 2.7.18 RC is available

[Marcin Cieslak]

On Sat, 18 Apr 2020, D'Arcy Cain wrote:


As for mailman, it looks like we only need to upgrade to a current
version.  The site says that 3.x uses Python 3.  Requires it in fact.


Mailman 3 is a completely different system compared to Mailman 2.


CKAN is not in FreeBSD ports but is an example of an application
that is not trivial to migrate due to their use of Pylons web technology:

https://github.com/ckan/ckan/issues/4681

converted into

https://github.com/ckan/ckan/projects/3

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: FreeBSD Port: libv4l-1.18.0 error upgrade

[Marcin Cieslak]

On Sat, 18 Apr 2020, Hans Petter Selasky wrote:


On 2020-04-18 16:12, Marcin Cieslak wrote:

On Fri, 17 Apr 2020, Hans Petter Selasky wrote:


On 2020-04-17 12:52, Alex V. Petrov wrote:

Making all in libdvbv5
gmake[4]: Entering directory
'/usr/ports/multimedia/libv4l/work/v4l-utils-1.18.0/lib/libdvbv5'


You need to deinstall the libv4l package first. Only then it will build 
due to confliciting header file include order!


Can this be improved upon somehow? Like pushing build directory include 
directories

first?



Yes, see:

https://svnweb.freebsd.org/changeset/ports/531931


thanks - it works.

I was able to build libv4l 1.18.0 having 1.6.3_4 installed without issues!

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: FreeBSD Port: libv4l-1.18.0 error upgrade

[Marcin Cieslak]

On Fri, 17 Apr 2020, Hans Petter Selasky wrote:


On 2020-04-17 12:52, Alex V. Petrov wrote:

Making all in libdvbv5
gmake[4]: Entering directory
'/usr/ports/multimedia/libv4l/work/v4l-utils-1.18.0/lib/libdvbv5'


You need to deinstall the libv4l package first. Only then it will build due 
to confliciting header file include order!


Can this be improved upon somehow? Like pushing build directory include 
directories
first?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Query on installing v6eval to FreeBSD 12.1

[Marcin Cieslak]

On Wed, 8 Apr 2020, radhika somaiya wrote:


Hi team,

I am very new to this FreeBSD world. Currently have set up my system having
dual booting of FreeBSD 12.1 release with preinstalled windows 10.

After booting into FreeBSD, I want to set up the Tahi self test tool where
I am facing the complication error while compiling the v6eval package. I
have downloaded v6eval latest package 3.3.5 version.


Hi Radhika,

Where did you download this from? I ad a look at 
https://github.com/IxLabs/tahi-linux/
and v6eval included there has not been updated since 2013 and needs some changes
to work with the newest FreeBSD. For example, it still uses the old  
interface.

I see that there is some version of OpenSSL bundled with the repository, do you 
know
why?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: hdf5 update and problem

[Marcin Cieslak]

On Sun, 29 Mar 2020, Kevin Oberman wrote:


On Sun, Mar 29, 2020 at 1:18 PM ajtiM via freebsd-ports <
freebsd-ports@freebsd.org> wrote:


Hi!

After update of hdf5 there are problems with blender which doesn't
build anymore:
executing: /usr/ports/graphics/blender/work/.build/bin/blender --help
ld-elf.so.1: Shared object "libhdf5.so.103" not found, required by
"libAlembic.so.1.7" Traceback (most recent call last):
  File "blender.1.py", line 56, in 

Also there is a problem with graphics/vigra too,

Thank you.


--
Ernst Lubitsch’s Ninotchka:



The source of libAlembic.so needs to be rebuilt to support the updated hdf5.
I don't use blender, but the generic answer is:
# pkg which libAlembic.so.1.7
The response is the port that is still linked to the hold hdf5 shareable.
Rebuild it and that will fix this one. Also, notify whoever updated hdf5
that the port needs a PORT_REVISION bump so that the package gets updated.


Here I have:

pkg which /usr/local/lib/libAlembic.so.1.7 

/usr/local/lib/libAlembic.so.1.7 was installed by package alembic-1.7.11_2

ldd /usr/local/lib/libAlembic.so.1.7

/usr/local/lib/libAlembic.so.1.7:
libImath-2_4.so.24 => /usr/local/lib/libImath-2_4.so.24 (0x80144a000)
libIlmThread-2_4.so.24 => /usr/local/lib/libIlmThread-2_4.so.24 
(0x80165c000)
libIex-2_4.so.24 => /usr/local/lib/libIex-2_4.so.24 (0x801863000)
libHalf-2_4.so.24 => /usr/local/lib/libHalf-2_4.so.24 (0x801ac9000)
libm.so.5 => /lib/libm.so.5 (0x801d0c000)
libhdf5.so.103 => /usr/local/lib/libhdf5.so.103 (0x80200)
libsz.so.2 => /usr/local/lib/libsz.so.2 (0x802598000)
libz.so.6 => /lib/libz.so.6 (0x8027b)
libdl.so.1 => /usr/lib/libdl.so.1 (0x8029c9000)
libc++.so.1 => /usr/lib/libc++.so.1 (0x802bca000)
libcxxrt.so.1 => /lib/libcxxrt.so.1 (0x802e99000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x8030b8000)
libthr.so.3 => /lib/libthr.so.3 (0x8032cb000)
libc.so.7 => /lib/libc.so.7 (0x800825000)
libIexMath-2_4.so.24 => /usr/local/lib/libIexMath-2_4.so.24 
(0x8034f4000)

pkg which /usr/local/lib/libhdf5.so.103

/usr/local/lib/libhdf5.so.103 was installed by package hdf5-1.10.6

You probably need to (re)install hdf5, but it should be enough to 
reinstall/upgrade alembic


smime.p7s
Description: S/MIME Cryptographic Signature

Re: Pkg repository is broken...

[Marcin Cieslak]

On Sat, 7 Mar 2020, Janky Jay, III wrote:


Still broken for me as well. I just figured I'd look into it when I got a 
chance. Also using FBSD 12.1 RELEASE.


Maybe it depends on the mirror being used?

smime.p7s
Description: S/MIME Cryptographic Signature

Re: good gui bit-torrent client?

[Marcin Cieslak]

On Fri, 28 Feb 2020, Robert Huff wrote:



I used to use azureus/vuze, but it hasn't been maintained is
quite a while.


I'm pretty happy with "transmission", but I prefer to work from
the command line (transmission-daemon, transmission-remote).

Once one gets used to specify a torrent number like this:

 transmission-remote -t 20 -l

(list information about the torrent file number 20)

it is pretty easy to use.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: About protocols in openssl

[Marcin Cieslak]

On Thu, 27 Feb 2020, Willem Jan Withagen wrote:

/home/jenkins/workspace/ceph-master/src/pybind/mgr/.tox/py3/lib/python3.7/site-packages/cryptography/hazmat/bindings/_openssl.abi3.so: 
Undefined symbol "SSLv3_client_method"


This looks to me that you are trying to build ceph in the virtualenv which 
probably
pulls required python packages on its own.

Can you make it to depend and use the existing security/py-cryptography port?

I don't know how exactly the ceph port is supposed to work but it seems to
require virtualenv for its inner workings. You might want to force
virtual env to use FreeBSD-provided libraries with --system-site-packages
but AFAIK it will no longer download anything, i.e. all dependencies
need to packaged.

I have just ran "make test" in my ports tree to test py-cryptography
and got 1 error (TestECDSACertificate.test_load_ecdsa_no_named_curve)
but nothing related to SSLv3_client_method being not there.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Timidity++ needs libarc as run dependency ??

[Marcin Cieslak]

On Thu, 27 Feb 2020, Mateusz Piotrowski wrote:


On 2/26/20 3:20 PM, Hans Petter Selasky wrote:

On 2020-02-26 13:18, Mateusz Piotrowski wrote:

On 2/26/20 10:55 AM, Hans Petter Selasky wrote:
ld-elf.so.1: Shared object "libarc.so.1" not found, required by 
"timidity"


pkg install libarc


It looks like libarc is already included in LIB_DEPENDS. It is not 
included in the timidity++ package. It is only included with the 
timidity++-${PKGNAMESUFFIX} packages (like timidity++-emacs).


I've never used timidity myself, so I am not sure why this is this way.

Is there a reason why you installed timidity++ instead of one of the 
timidity++-${PKGNAMESUFFIX} packages?




No specific reason. I just wanted to test some MIDI files.

And did "pkg instal xxx". Is timidity++ not a valid port/package?


It is a valid port. I am not sure why libarc is not a run-time dependency 
here. Maybe just an oversight.


Maybe there is some bug with

.if !defined(PKGNAMESUFFIX)

during package building?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Wed, 26 Feb 2020, Miroslav Lachman wrote:


Marcin Cieslak wrote on 2020/02/26 13:50:

On Wed, 26 Feb 2020, Miroslav Lachman wrote:

And what is not working now with the line fontFamilyList->append... 
removed()?



Should I report it upstream or is it FreeBSD specific problem?


I think it is a bit too early to tell. I think it is not really 
FreeBSD-specific

but there might be some quirks.

I am rebuilding qt5-webkit now with debug symbols using your phantomfs
port (thanks!) and I'll poke around a bit with debugger.


Just the small update.
phantomjs with modified version of qt5-webkit does not render any text / 
fonts on the web page screenshots taken by

page.render('/tmp/page.png');

So... it builds, it does not crashes but rendered pages are "empty" 
(pictures, background colors etc are rendered OK, but no fonts)


This is expected. This line is responsible for finding fonts
used to render the webpage. Michael just wanted to confirm
this line is indeed causing the problem; we now have to figure
out what exactly is going wrong here.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Wed, 26 Feb 2020, Miroslav Lachman wrote:

And what is not working now with the line fontFamilyList->append... 
removed()?



Should I report it upstream or is it FreeBSD specific problem?


I think it is a bit too early to tell. I think it is not really FreeBSD-specific
but there might be some quirks.

I am rebuilding qt5-webkit now with debug symbols using your phantomfs
port (thanks!) and I'll poke around a bit with debugger.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Mon, 24 Feb 2020, Miroslav Lachman wrote:


Marcin Cieslak wrote on 2020/02/24 18:51:

On Mon, 24 Feb 2020, Marcin Cieslak wrote:


On Mon, 24 Feb 2020, Miroslav Lachman wrote:

   frame #13: 0x0008039aa3ed 
libQt5WebKit.so.5`WebCore::CSSParser::parseSheet(this=0x7fffa490, 
sheet=0x0008155f5e40, string=0x7fffb888, 
textPosition=0x7fffb650, ruleSourceDataResult=0x, 
logErrors=false) at CSSParser.cpp:423:5


Looks like it has some trouble parsing font declarations in the HTML/CSS 
code you have been testing

on. This can be a webkit bug...

Would be good to have a minimal HTML/CSS testcase to reproduce this and 
this should

probably be reported upstream.


There is this report:

https://github.com/qtwebkit/qtwebkit/issues/933

It says it crashes if you have no fonts installed (no TTF).

Could this "fix" it?


Thank you for the tip, I will look at it.

The following fonts related packages are installed:

(root@testjail) ~/# pkg info | grep -i font
dejavu-2.37_1  Bitstream Vera Fonts clone with a wider range 
of characters

encodings-1.0.5,1  X.Org Encoding fonts
font-bh-ttf-1.0.3_4X.Org Bigelow & Holmes TTF font
font-misc-ethiopic-1.0.3_4 X.Org miscellaneous Ethiopic font
font-misc-meltho-1.0.3_4   X.Org miscellaneous Meltho font
fontconfig-2.12.6,1XML-based font configuration API for X Windows
freetype2-2.10.1   Free and portable TrueType font rendering 
engine

libfontenc-1.1.4   The fontenc Library
mkfontscale-1.2.1  Creates an index of scalable font files for X
woff2-1.0.2_4  Library and converter tools for the WOFF 2.0 
web font format

xorg-fonts-truetype-7.7_1  X.Org TrueType fonts



This is the code of the tested example - fetching from https://phantomjs.org/


Can you check if "fc-list" can see the fonts?

Can you publish your work in progress phantomjs port so I can try this here?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Mon, 24 Feb 2020, Marcin Cieslak wrote:


On Mon, 24 Feb 2020, Miroslav Lachman wrote:

   frame #13: 0x0008039aa3ed 
libQt5WebKit.so.5`WebCore::CSSParser::parseSheet(this=0x7fffa490, 
sheet=0x0008155f5e40, string=0x7fffb888, 
textPosition=0x7fffb650, 
ruleSourceDataResult=0x, logErrors=false) at 
CSSParser.cpp:423:5


Looks like it has some trouble parsing font declarations in the 
HTML/CSS code you have been testing

on. This can be a webkit bug...

Would be good to have a minimal HTML/CSS testcase to reproduce 
this and this should

probably be reported upstream.


There is this report:

https://github.com/qtwebkit/qtwebkit/issues/933

It says it crashes if you have no fonts installed (no TTF).

Could this "fix" it?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Mon, 24 Feb 2020, Miroslav Lachman wrote:


"phantomjs" can run simple "Hello world" example
https://github.com/ariya/phantomjs/blob/master/examples/hello.js

But when I try to run some real work (fetching web page) it segfaulted.


Looks like Qt5Webkit has a problem reading CSS for the web page you have
been testing it on. Can you share the code of the page?


The output is long so I put it on line: https://hastebin.com/ibefonagob.bash


* thread #1, name = 'phantomjs', stop reason = signal SIGSEGV: invalid address 
(fault address: 0x10)

frame #9: 0x000803992791 
libQt5WebKit.so.5`WebCore::CSSValuePool::createFontFamilyValue(this=0x000806ee1d18,
 familyName=0x00081551f338, fromSystemFontID=Yes) at CSSValuePool.cpp:129:63
frame #10: 0x0008039cf92c 
libQt5WebKit.so.5`WebCore::CSSParser::parseSystemFont(this=0x7fffa490, 
important=false) at CSSParser.cpp:6907:54
frame #11: 0x0008039b6beb 
libQt5WebKit.so.5`WebCore::CSSParser::parseValue(this=0x7fffa490, 
propId=CSSPropertyFont, important=false) at CSSParser.cpp:3058:13
frame #12: 0x0008043d74bf 
libQt5WebKit.so.5`cssyyparse(parser=0x7fffa490) at CSSGrammar.y:1547:39
frame #13: 0x0008039aa3ed 
libQt5WebKit.so.5`WebCore::CSSParser::parseSheet(this=0x7fffa490, 
sheet=0x0008155f5e40, string=0x7fffb888, 
textPosition=0x7fffb650, ruleSourceDataResult=0x, 
logErrors=false) at CSSParser.cpp:423:5

Looks like it has some trouble parsing font declarations in the HTML/CSS code 
you have been testing
on. This can be a webkit bug...

Would be good to have a minimal HTML/CSS testcase to reproduce this and this 
should
probably be reported upstream.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot build qt5-webkit with debug

[Marcin Cieslak]

On Mon, 24 Feb 2020, Michael Osipov wrote:


(root@testjail) ~/# gdb /usr/local/bin/phantomjs
GNU gdb 6.1.1 [FreeBSD]
Copyright 2004 Free Software Foundation, Inc.


Can you try gdb from ports?

/usr/local/bin/gdb831 works much better with a modern C++ applications.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Debugging build with gmake, running npm toolchain

[Marcin Cieslak]

On Wed, 12 Feb 2020, Stefan Bethke wrote:


The second issue is that npm install is invoked, which will most likely prevent 
the port to be built as a package. Is there a way to handle the download of the 
package-lock.json dependencies in the fetch stage?


I have some hack in my private node-sass port: 
https://github.com/saper/ports-exp/tree/master/textproc/node-sass

jrm@ has solved this in the late net-im/mastodon port by simply putting all 
JavaScript together
in one tarball: 
https://svnweb.freebsd.org/ports/head/net-im/mastodon/?pathrev=472547

https://ftfl.ca/blog/2017-05-23-mastodon-freebsd.html

The port got removed in 
https://svnweb.freebsd.org/ports?view=revision=474751

saper

smime.p7s
Description: S/MIME Cryptographic Signature

Re: lang/mono: testers wanted for 6.8.0.96 update!

[Marcin Cieslak]

On Tue, 21 Jan 2020, Phillip R. Jaenke wrote:


Just go here: https://reviews.freebsd.org/D2330


https://reviews.freebsd.org/D23300

Thanks for picking it up! I have upstreamed some FreeBSD
changes to mono in the past and I am happy this gets
picked up again!

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Retiring GNU objdump 2.17.50

[Marcin Cieslak]

On Thu, 9 Jan 2020, Ed Maste wrote:


world or kernel). It is required to build a limited number of ports,
and is used by some developers.


My use case is troubleshooting tool - I can always ask somebody
on the remote computer to check their binaries with a tool that
is in the base system. ("Can you objdump -T this library?",
"Can you objdump -R something | grep symbolname?").

I think the first use case can be solved with "nm -D", I don't see
the replacement for the latter. Since I also support Mac, Linux users it
nice to have one tool that works the same across all the platforms;
but give I need to take "otool" for older MacOS into account I don't
mind to have something else under a different name, with different
options, as long as it is well known and preferably always available
for troubleshooting (not an additionally installed something).
Haven't worked with llvm-objdump yet.

I like the disassembler feature but for myself I can install
binutils.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: getting Frotz 2.50 committed

[Marcin Cieslak]

On Sun, 5 Jan 2020, David Griffith wrote:



I submitted an update for games/frotz at 
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=242821.  Now what?


For the archives: the patch has been committeed by tcberner@

Thanks!

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Frotz upgrade to 2.50

[Marcin Cieslak]

On Fri, 20 Dec 2019, David Griffith wrote:

With this patch applied, you get both frotz without audio support and dumb 
frotz.  For audio, some more libraries are required, but that should not pull 
in anything graphical.  For the graphics-capable variant of Frotz, SDL2 is 
needed among other things.


I propose to add two additional ports entries for Frotz.  One is 
"frotz-audio" which is simply curses Frotz with the audio subsystem compiled 
in and dumb Frotz along for the ride.  This would be mutually-exclusive from 
the regular silent "frotz" entry for obvious reasons.  The other would be 
"frotz-sdl" for the SDL interface.


I don't know frotz so I have no idea what the differences are but we could
decide to have more OPTIONS or FLAVOURS (pretty new facility in FreeBSD ports)
if really needed. I wonder why DUMB is optional really - it does not take
much to build I guess so it could always be shipped (or never if one feels so).

One can always decided to ship the tools with "all batteries included"
- I don't know what the users really want. Sadly this port has no maintainer
who would probably knew more.


For now, here's the patch for silent Frotz:

===cut here===
diff -ruP frotz-orig/Makefile frotz/Makefile


can you add -N ? otherwise we only get:


Only in frotz-orig/files: patch-src_curses_ux__audio__oss.c


Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Frotz upgrade to 2.50

[Marcin Cieslak]

On Thu, 19 Dec 2019, David Griffith wrote:



My reply is at the bottom.  Please put your reply there too.
On Wed, 18 Dec 2019, Marcin Cieslak wrote:

On Mon, 16 Dec 2019, David Griffith wrote:



I released version 2.50 of Frotz last month.  Would someone please update 
ports to install this version?


I have never used that software but I gave it a try and:

* moved it to Gitlab
* applied fix for a https://gitlab.com/DavidGriffith/frotz/issues/180
* I think we do not need to generate soundcard.h manually anymore?

Now I am getting:

===>  Building for frotz-2.50
gmake[2]: Entering directory 
'/usr/home/saper/sw/FreeBSD/ports/games/frotz/work/frotz-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64'

** Generating src/common/defs.h
** Generating src/common/git_hash.h
** ERROR UTF-8 support only works with ncursesw!
exit 2
gmake[2]: *** [Makefile:391: src/curses/ux_defines.h] Error 2

but then second "make" works just fine...


The OSS interface is deprecated for now.  There's no need to go hunting 
around for soundcard.h anymore in favor of libao.  I've resisted removing 
ux_audio_oss.c entirely because I'd like to someday reenable that for older 
machines and massage it into something for older Sun workstations.


Thank you. Is there any quick way to test if sound is working?

What about that ncursesw thing? I think it is NetBSD-specific?


How is this patch supposed to be applied?


Just like as you did. Can you check if your ports tree is up to date?

The Makefile should have the following line:

# $FreeBSD: head/games/frotz/Makefile 442400 2017-06-02 15:43:42Z sunpoet $

If 442400 is a lower number, the ports tree needs updating, the way
to do it depends on the way you got it (I have used the portsnap method):

https://www.freebsd.org/doc/handbook/ports-using.html

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Frotz upgrade to 2.50

[Marcin Cieslak]

On Mon, 16 Dec 2019, David Griffith wrote:



I released version 2.50 of Frotz last month.  Would someone please update 
ports to install this version?


I have never used that software but I gave it a try and:

* moved it to Gitlab
* applied fix for a https://gitlab.com/DavidGriffith/frotz/issues/180
* I think we do not need to generate soundcard.h manually anymore?

Now I am getting:

===>  Building for frotz-2.50
gmake[2]: Entering directory 
'/usr/home/saper/sw/FreeBSD/ports/games/frotz/work/frotz-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64'
** Generating src/common/defs.h
** Generating src/common/git_hash.h
** ERROR UTF-8 support only works with ncursesw!
exit 2
gmake[2]: *** [Makefile:391: src/curses/ux_defines.h] Error 2

but then second "make" works just fine...

Marcin

my patch to the port below if you want to try:

diff -ruN /usr/ports/games/frotz/distinfo ./distinfo
--- /usr/ports/games/frotz/distinfo 2015-05-26 20:50:09.0 +0200
+++ ./distinfo  2019-12-18 23:51:11.422706000 +0100
@@ -1,2 +1,3 @@
-SHA256 (DavidGriffith-frotz-2.44_GH0.tar.gz) = 
dbb5eb3bc95275dcb984c4bdbaea58bc1f1b085b20092ce6e86d9f0bf3ba858f
-SIZE (DavidGriffith-frotz-2.44_GH0.tar.gz) = 276815
+TIMESTAMP = 1576709471
+SHA256 
(DavidGriffith-frotz-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64_GL0.tar.gz) = 
0fdbf055edd31e2f07b52f720a74a098d81b766a00a4103955edb2171d13b0d6
+SIZE (DavidGriffith-frotz-9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64_GL0.tar.gz) 
= 346458
diff -ruN /usr/ports/games/frotz/files/patch-Makefile ./files/patch-Makefile
--- /usr/ports/games/frotz/files/patch-Makefile 1970-01-01 01:00:00.0 
+0100
+++ ./files/patch-Makefile  2019-12-18 23:55:18.081693000 +0100
@@ -0,0 +1,11 @@
+--- Makefile.orig  2019-12-18 22:55:03 UTC
 Makefile
+@@ -297,7 +297,7 @@ ifneq ($(and $(wildcard $(GIT_DIR)),$(sh
+   $(NAME)src/doc/*.6 $(NAME)src/doc/frotz.conf* \
+   $(NAME)src/doc/Xresources  > /dev/null
+ else
+-@echo "Not in a git repository or git command not found.  Cannot make a 
tarball."
++  @echo "Not in a git repository or git command not found.  Cannot make a 
tarball."
+ endif
+ 
+ all: $(FROTZ_BIN) $(DFROTZ_BIN) $(SFROTZ_BIN)

diff -ruN /usr/ports/games/frotz/files/patch-src_curses_ux__audio__oss.c 
./files/patch-src_curses_ux__audio__oss.c
--- /usr/ports/games/frotz/files/patch-src_curses_ux__audio__oss.c  
2014-12-05 18:18:12.0 +0100
+++ ./files/patch-src_curses_ux__audio__oss.c   1970-01-01 01:00:00.0 
+0100
@@ -1,20 +0,0 @@
 src/curses/ux_audio_oss.c.orig 2014-03-06 06:08:15 UTC
-+++ src/curses/ux_audio_oss.c
-@@ -80,7 +80,7 @@ static void sigterm_handler(int signal) 
-   _exit(0);

- }
- 
--static void sigint_handler(int signal) {

-+static void oss_sigint_handler(int signal) {
-   num_repeats = 1;
- }
- 
-@@ -141,7 +141,7 @@ static void play_sound(int volume, int r

-   sigaddset(_mask, SIGTERM);
-   sa.sa_flags = 0;
-   sigaction(SIGTERM, , NULL);
--  sa.sa_handler = sigint_handler;
-+  sa.sa_handler = oss_sigint_handler;
-   sigaction(SIGINT, , NULL);
- 
-   for (num_repeats = repeats; num_repeats > 0;

diff -ruN /usr/ports/games/frotz/Makefile ./Makefile
--- /usr/ports/games/frotz/Makefile 2017-06-02 17:43:42.0 +0200
+++ ./Makefile  2019-12-19 00:03:18.599172000 +0100
@@ -2,7 +2,7 @@
 # $FreeBSD: head/games/frotz/Makefile 442400 2017-06-02 15:43:42Z sunpoet $

 PORTNAME=  frotz
-PORTVERSION=   2.44
+PORTVERSION=   2.50
 CATEGORIES=games

 MAINTAINER=po...@freebsd.org
@@ -11,8 +11,9 @@
 LICENSE=   GPLv2+
 LICENSE_FILE=  ${WRKSRC}/COPYING

-USE_GITHUB=yes
-GH_ACCOUNT=DavidGriffith
+USE_GITLAB=yes
+GL_ACCOUNT=DavidGriffith
+GL_COMMIT= 9867a1f14da1e9c0707492d2ac74d1e8ffdd3a64

 USES=  gmake ncurses

@@ -25,8 +26,7 @@
 PORTDOCS=  AUTHORS BUGS ChangeLog HOW_TO_PLAY README
 PORTEXAMPLES=  frotz.conf-big frotz.conf-small

-OPTIONS_DEFINE=DOCS EXAMPLES DUMB OSS
-OPTIONS_DEFAULT=   OSS
+OPTIONS_DEFINE=DOCS EXAMPLES DUMB
 DUMB_DESC= Install dfrotz (dumb frotz)

 OSS_MAKE_ENV=  SOUND_DEFS='-DOSS_SOUND -DSOUND_DEV=\"/dev/dsp\"'
@@ -34,9 +34,6 @@
man/man6/dfrotz.6.gz
 DUMB_PORTDOCS= DUMB

-pre-build-OSS-on:
-   cd ${WRKSRC} && ${DO_MAKE_BUILD} soundcard.h
-
 do-install:
${INSTALL_PROGRAM} ${WRKSRC}/${PORTNAME} ${STAGEDIR}${PREFIX}/bin/
${INSTALL_MAN} ${WRKSRC}/doc/${PORTNAME}.6 
${STAGEDIR}${MAN6PREFIX}/man/man6/

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Web Service Discovery host daemon

[Marcin Cieslak]

On Mon, 14 Oct 2019, Carmel NY wrote:


I am considering installing "wsdd" .
I have not been able to find a port for it. I just want to confirm that
there isn't a port available for this program before I go about
installing it from source.


No, I don's see it:

$ pkg search wsdd

(nothing)

$ pkg search discovery
jakarta-commons-discovery-0.5  Java library for locating resources (including 
classes)
p5-HTML-RSSAutodiscovery-1.21_1 Methods for retrieving RSS-ish information from 
an HTML document
rubygem-aws-sdk-applicationdiscoveryservice-1.21.0 Official AWS Ruby gem for 
AWS Application Discovery Service
rubygem-aws-sdk-servicediscovery-1.18.0 Official AWS Ruby gem for AWS Cloud Map 
(ServiceDiscovery)

$ (cd /usr/ports && make search name="wsdd")

(nothing)


It looks like the author already provides startup file for FreeBSD
so I think it would be trivial to prepare and submit the port:

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/

Looks like these are only few files to install.

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/licenses.html

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/makefile-manpages.html

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/install.html

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/rc-scripts.html

https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/using-python.html

saper

smime.p7s
Description: S/MIME Cryptographic Signature

Re: inkscape core dumps at start

[Marcin Cieslak]

On Mon, 15 Oct 2018, Marcin Cieslak wrote:

> On Sun, 14 Oct 2018, Marcin Cieslak wrote:
> 
> > On Wed, 10 Oct 2018, Erich Dollansky wrote:
> > 
> > > Hi,
> > > 
> > > On Tue, 9 Oct 2018 14:57:55 +0200
> > > Walter Schwarzenfeld  wrote:
> > > 
> > > > seems the same as in this PR
> > > > 
> > > > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=232073
> > > > 
> > > yes, it is. So, we are now already two having the same problem.
> > 
> > Getting the same on 11.1-R without updating inkscape:
> > 
> > Version: 0.92.2
> > Installed on   : Thu Mar  1 02:02:11 2018 CET
> > 
> > I have updated a bunch of libraries a moment ago and
> > one of those is probably a problem.
> 
> https://github.com/caskformula/homebrew-caskformula/issues/51
> 
> https://bugzilla.gnome.org/show_bug.cgi?id=795338
> 
> Haven't tried rebuilding libgtkmm with __STDC_ISO_10646__ defined yet.

The patch below (taken from Bugzilla) can be put into 
/usr/ports/devel/glibmm/files

After rebuilding glibmm inkscape starts again.

I wonder why we run into this issue, should

USE_CXXSTD= gnu++11

be added to the port's makefile?

Marcin

--- glib/glibmm/ustring.cc.orig 2017-08-30 14:13:30 UTC
+++ glib/glibmm/ustring.cc
@@ -1364,7 +1364,7 @@ ustring::FormatStream::to_string() const
 #ifdef GLIBMM_HAVE_WIDE_STREAM
   const std::wstring str = stream_.str();
 
-#if defined(__STDC_ISO_10646__) && SIZEOF_WCHAR_T == 4
+#if (defined(__STDC_ISO_10646__) || defined(_LIBCPP_VERSION)) && 
SIZEOF_WCHAR_T == 4
   // Avoid going through iconv if wchar_t always contains UCS-4.
   glong n_bytes = 0;
   const auto buf = make_unique_ptr_gfree(g_ucs4_to_utf8(
@@ -1452,7 +1452,7 @@ operator>>(std::wistream& is, ustring& u
   std::wstring wstr;
   is >> wstr;
 
-#if defined(__STDC_ISO_10646__) && SIZEOF_WCHAR_T == 4
+#if (defined(__STDC_ISO_10646__) || defined(_LIBCPP_VERSION)) && 
SIZEOF_WCHAR_T == 4
   // Avoid going through iconv if wchar_t always contains UCS-4.
   glong n_bytes = 0;
   const auto buf = make_unique_ptr_gfree(g_ucs4_to_utf8(
@@ -1483,7 +1483,7 @@ operator<<(std::wostream& os, const ustr
 {
   GError* error = nullptr;
 
-#if defined(__STDC_ISO_10646__) && SIZEOF_WCHAR_T == 4
+#if (defined(__STDC_ISO_10646__) || defined(_LIBCPP_VERSION)) && 
SIZEOF_WCHAR_T == 4
   // Avoid going through iconv if wchar_t always contains UCS-4.
   const auto buf = make_unique_ptr_gfree(
 g_utf8_to_ucs4(utf8_string.raw().data(), utf8_string.raw().size(), 
nullptr, nullptr, ));

smime.p7s
Description: S/MIME Cryptographic Signature

Re: libdl.so missing after "pkg upgrade" - firefox, python27...

[Marcin Cieslak]

On Mon, 15 Oct 2018, Mathieu Arnold wrote:

> > 
> > Why this happens? What am I doing wrong?
> 
> 11.1 is no longer supported, you must have set IGNORE_OSVERSION, which
> is why pkg was happy to break your system at your request.
> You will need to update to 11.2.

Thanks, I know that 11.1 stopped being supported and I try to get out of
the trouble before I upgrade (never upgrade a broken system).

I don't think I have IGNORE_OSVERSION set anywhere or changed ABI or OSVERSION.
(I have a feeling that pkg relaxed OSVERSION checks recently, since they
 were pretty common previously).

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: inkscape core dumps at start

[Marcin Cieslak]

On Sun, 14 Oct 2018, Marcin Cieslak wrote:

> On Wed, 10 Oct 2018, Erich Dollansky wrote:
> 
> > Hi,
> > 
> > On Tue, 9 Oct 2018 14:57:55 +0200
> > Walter Schwarzenfeld  wrote:
> > 
> > > seems the same as in this PR
> > > 
> > > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=232073
> > > 
> > yes, it is. So, we are now already two having the same problem.
> 
> Getting the same on 11.1-R without updating inkscape:
> 
> Version: 0.92.2
> Installed on   : Thu Mar  1 02:02:11 2018 CET
> 
> I have updated a bunch of libraries a moment ago and
> one of those is probably a problem.

gdb backtrace looks like this:

gdb --args /usr/local/bin/inkscape --g-fatal-warnings
GNU gdb 6.1.1 [FreeBSD]
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-marcel-freebsd"...
(gdb) run
Starting program: /usr/local/bin/inkscape --g-fatal-warnings
warning: Lowest section in /usr/local/lib/libicudata.so.62 is .hash at 
0120

Program received signal SIGABRT, Aborted.
0x00080cc5a84a in thr_kill () from /lib/libc.so.7
(gdb) bt
#0  0x00080cc5a84a in thr_kill () from /lib/libc.so.7
#1  0x00080cc5a814 in raise () from /lib/libc.so.7
#2  0x00080cc5a789 in abort () from /lib/libc.so.7
#3  0x0008094a0e38 in Glib::ConvertError::throw_func () from 
/usr/local/lib/libglibmm-2.4.so.1
#4  0x0008094b5679 in Glib::Error::throw_exception () from 
/usr/local/lib/libglibmm-2.4.so.1
#5  0x0008094c2f95 in Glib::operator<< () from 
/usr/local/lib/libglibmm-2.4.so.1
#6  0x00080224a9b9 in Glib::ustring::FormatStream::stream 
(this=0x7fffbae8, value=0x802b4a3df "Font size") at ustring.h:1072
#7  0x000802249a95 in Glib::ustring::format (a1=@0x7fffcc40, a2=@0x802ac52cb, a3=@0x7fffcc38, 
a4=@0x802b8ddc1) at ustring.h:1173
#8  0x00080223af7c in sp_text_toolbox_prep (desktop=0x816769c00, 
mainActions=0x817f974d0, holder=0x81925dbb0) at 
src/widgets/text-toolbar.cpp:1792
#9  0x0008022f5e42 in setup_aux_toolbox (toolbox=0x8174fd410, 
desktop=0x816769c00) at src/widgets/toolbox.cpp:1425
#10 0x0008022f51f6 in Inkscape::UI::ToolboxFactory::setToolboxDesktop 
(toolbox=0x8174fd410, desktop=0x816769c00) at src/widgets/toolbox.cpp:1203
#11 0x000801c18730 in Inkscape::UI::UXManagerImpl::connectToDesktop 
(this=0x81810dfd0, toolboxes=@0x7fffdb78, desktop=0x816769c00) at 
src/ui/uxmanager.cpp:261
#12 0x000802259e83 in SPDesktopWidget::createInstance 
(namedview=0x817443000) at src/widgets/desktop-widget.cpp:1799
#13 0x0008022592b5 in sp_desktop_widget_new (namedview=0x817443000) at 
src/widgets/desktop-widget.cpp:1748
#14 0x0008023f065c in sp_file_new (templ=@0x7fffe188) at 
src/file.cpp:143
#15 0x0008023f133a in sp_file_new_default () at src/file.cpp:214
#16 0x0040c35f in sp_main_gui (argc=1, argv=0x7fffe828) at 
src/main.cpp:1160
#17 0x0040b51a in main (argc=2, argv=0x7fffe828) at src/main.cpp:814

(gdb) frame 6
#6  0x00080224a9b9 in Glib::ustring::FormatStream::stream 
(this=0x7fffbae8, value=0x802b4a3df "Font size") at ustring.h:1072
l1072 stream_ << ustring(value);

Looks like a Unicode handling problem in libgtkmm with libc++, looks like this 
has been reported
also for the MacOS:

https://github.com/caskformula/homebrew-caskformula/issues/51

https://bugzilla.gnome.org/show_bug.cgi?id=795338

Haven't tried rebuilding libgtkmm with __STDC_ISO_10646__ defined yet.

Marcin


smime.p7s
Description: S/MIME Cryptographic Signature

Re: libdl.so missing after "pkg upgrade" - firefox, python27...

[Marcin Cieslak]

On Mon, 15 Oct 2018, Jan Beich wrote:

> Marcin Cieslak  writes:
> 
> > Hello,
> >
> > during preparations to upgrade 11.1-RELEASE p15 system I got firefox-esr
> > upgraded to 
> >
> > Oct 14 22:18:30 radziecki pkg: pkg upgraded: 1.10.5_1 -> 1.10.5_3 
> >
> > Oct 14 22:44:51 radziecki pkg: firefox-esr upgraded: 52.8.0,1 -> 60.2.2_1,1 
> >
> > Oct 14 22:54:37 radziecki pkg: python27 upgraded: 2.7.14_1 -> 2.7.15 
> >
> > However, both python2 and firefox fail due to a missing "libdl.so",
> > which as I understand got introduced in 12.0.
> >
> > Why this happens? What am I doing wrong?
> 
> FreeBSD 11.1 reached EOL on 2018-09-30, so the package cluster switched
> to 11.2 as the oldest supported 11.* release where libdl.so is available.
> 
> https://svnweb.freebsd.org/changeset/base/320872
> 
> pkg upgrade should've warned you about OSVERSION mismatch. Did you ignore?

It didn't. I have only ALIAS statement uncommented in pkg.conf.

Python 2.7 I have installed from ports:

> pkg info -A python27
python27-2.7.15:
FreeBSD_version: 1101001
cpe: cpe:2.3:a:python:python:2.7.15:freebsd11:x64

Firefox installed via pkg:

> pkg info -A firefox-esr
firefox-esr-60.2.2_1,1:
FreeBSD_version: 1102000
cpe: 
cpe:2.3:a:mozilla:firefox_esr:60.2.2:freebsd11:x64:1
no_provide_shlib: yes
repo_type  : binary
repository : FreeBSD

It is frustrating that some small hack breaks stuff like this, something like 
this has not happened
to me for a long time.

I have solved this by installing libdl from HEAD SVN, trying to install firefox 
via ports
was too painful:

svn co https://svn.freebsd.org/base/head/lib/libdl
cd libdl
make
make install

(worked with 11.1 sources from https://svn.freebsd.org/base/releng/11.1 in 
/usr/src).

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

libdl.so missing after "pkg upgrade" - firefox, python27...

[Marcin Cieslak]

Hello,

during preparations to upgrade 11.1-RELEASE p15 system I got firefox-esr
upgraded to 

Oct 14 22:18:30 radziecki pkg: pkg upgraded: 1.10.5_1 -> 1.10.5_3 

Oct 14 22:44:51 radziecki pkg: firefox-esr upgraded: 52.8.0,1 -> 60.2.2_1,1 

Oct 14 22:54:37 radziecki pkg: python27 upgraded: 2.7.14_1 -> 2.7.15 

However, both python2 and firefox fail due to a missing "libdl.so",
which as I understand got introduced in 12.0.

Why this happens? What am I doing wrong?

Marcin Cieślak
___
freebsd-ports@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"

Re: inkscape core dumps at start

[Marcin Cieslak]

On Wed, 10 Oct 2018, Erich Dollansky wrote:

> Hi,
> 
> On Tue, 9 Oct 2018 14:57:55 +0200
> Walter Schwarzenfeld  wrote:
> 
> > seems the same as in this PR
> > 
> > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=232073
> > 
> yes, it is. So, we are now already two having the same problem.

Getting the same on 11.1-R without updating inkscape:

Version: 0.92.2
Installed on   : Thu Mar  1 02:02:11 2018 CET

I have updated a bunch of libraries a moment ago and
one of those is probably a problem.

Marcin Cieślak


smime.p7s
Description: S/MIME Cryptographic Signature

libdl.so missing after "pkg upgrade" - firefox, python27...

[Marcin Cieslak]

Hello,

during preparations to upgrade 11.1-RELEASE p15 system I got firefox-esr
upgraded to 

Oct 14 22:18:30 radziecki pkg: pkg upgraded: 1.10.5_1 -> 1.10.5_3 

Oct 14 22:44:51 radziecki pkg: firefox-esr upgraded: 52.8.0,1 -> 60.2.2_1,1 

Oct 14 22:54:37 radziecki pkg: python27 upgraded: 2.7.14_1 -> 2.7.15 

However, both python2 and firefox fail due to a missing "libdl.so",
which as I understand got introduced in 12.0.

Why this happens? What am I doing wrong?

Marcin Cieślak

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

On Thu, 1 Mar 2018, Yuri wrote:

> On 02/26/18 17:11, Marcin Cieslak wrote:
> > So I don't know what has changed and why but the temporary fix is to
> > use "if" to check if the desired files are not already there, and
> > then proceeding with "post-fetch" only if the files are not found.
> 
> 
> Some time back I was working on nodejs support:
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=204577
> 
> So far, I couldn't make it stable because after a week or two files fetched
> from the npm server change without an apparent reason.
> 
> I will return to it to see if the situation improved.

Hi Yuri,

thanks for the pointer - it is very useful! Also I find a very useful reference
to Olivier Duchateau's work as well.

I share your pain. I am maintaining this local port for 2 years and it is a 
pain.
I must admit the pain comes not only from npm volatility (I somehow managed
to handle that) but mostly from the fact that I still needed to maintain
node 0.10, node 0.12 and io.js support while they were gone.

It is also interesting how the two different approaches - your approach
to help build larger applications ports simply (that could help, for example,
the current net-im/mastodon port by jrm@) or more comprehensive apporach
from Oliver - ruby-style, one port per registry package.

The latter approach is more correct but will be problematic in case of
when package A requires package B~1.0 and package C requires package B~2.0.

So we will need some why to have "local" modules installed under application
specific node_modules 
(/usr/local/lib/node_modules/package_c/node_modules/package_b_2.0).

Good news is that for binary modules (where I am mostly interested) we will
need to simply disallow this - not because we believe we are the best in the 
world,
but because trying dlopen() second binary module with the same name into node
will not work as expected.

I also would like to have possibility to build for multiple node engines 
(node4-,
node6-, node8- and node9-) which is a must for binary packages. I would even
use NODE_MODULES_VERSION API number instead of 4, 6, 8, 9.

So, for binary modules we have to have Oliver's approach, for the rest, it will 
be hard.

I'll try to examine your patches and come up with something reasonable.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase? [solved]

[Marcin Cieslak]

On Sun, 25 Feb 2018, Marcin Cieslak wrote:

> Hello,
> 
> why is poudriere on 11.1-STABLE trying to build by custom port which has the 
> following
> in the Makefile:
> 
> post-fetch:
>   ${MKDIR} ${DISTDIR}/${PORTNAME}
>   ${MKDIR} ${NPM_CACHE}
>   ${CP} ${FILESDIR}/package-lock.json ${DISTDIR}/${PORTNAME}
>   ${CP} ${FILESDIR}/package.json ${DISTDIR}/${PORTNAME}
>   (cd ${DISTDIR}/${PORTNAME} && ${SETENV} NPM_CONFIG_CACHE=${NPM_CACHE} 
> npm install --ignore-scripts)
> 
> ( complete port: 
> https://github.com/saper/ports-exp/tree/master/textproc/node-sass )
> 
> The results are those:
> 
> ===>   Returning to build of node-sass-4.7.2
> ===
> ===
> ===>  License MIT accepted by the user
> ===> Fetching all distfiles required by node-sass-4.7.2 for building
> /bin/mkdir -p /portdistfiles/node-sass
> /bin/mkdir -p /portdistfiles/npm
> /bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
> /portdistfiles/node-sass
> (cd /portdistfiles/node-sass && /usr/bin/env 
> NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
> npm WARN saveError ENOENT: no such file or directory, open 
> '/portdistfiles/node-sass/package.json'
> npm WARN enoent ENOENT: no such file or directory, open 
> '/portdistfiles/node-sass/package.json'
> npm WARN node-sass No description
> npm WARN node-sass No repository field.
> npm WARN node-sass No README data
> npm WARN node-sass No license field.
> 
> added 4 packages in 1.185s
> ===
> ===
> ===>  License MIT accepted by the user
> ===> Fetching all distfiles required by node-sass-4.7.2 for building
> /bin/mkdir -p /portdistfiles/node-sass
> /bin/mkdir -p /portdistfiles/npm
> /bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
> /portdistfiles/node-sass
> (cd /portdistfiles/node-sass && /usr/bin/env 
> NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
> npm WARN saveError ENOENT: no such file or directory, open 
> '/portdistfiles/node-sass/package.json'
> npm WARN enoent ENOENT: no such file or directory, open 
> '/portdistfiles/node-sass/package.json'
> npm WARN node-sass No description
> npm WARN node-sass No repository field.
> npm WARN node-sass No README data
> npm WARN node-sass No license field.
> 
> up to date in 0.514s
> => SHA256 Checksum OK for sass-node-sass-v4.7.2_GH0.tar.gz.
> ===
> ===
> ===
> ===
> ===>  License MIT accepted by the user
> ===> Fetching all distfiles required by node-sass-4.7.2 for building
> /bin/mkdir -p /portdistfiles/node-sass
> /bin/mkdir -p /portdistfiles/npm
> /bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
> /portdistfiles/node-sass
> cp: /portdistfiles/node-sass/package-lock.json: Permission denied
> *** Error code 1
> 
> Stop.
> make: stopped in /usr/ports/textproc/node-sass
> 

For the archives:

I have solved it (and some other related problem that occured later) by
copying even more files to $WRKDIR like crazy:

DIST_NPM_CACHE= ${DISTDIR}/npm
BUILD_NPM_CACHE=${WRKDIR}/npm

.include 
.if ${COMPILER_TYPE} == clang
MAKE_ENV+=  GYP_DEFINES=clang=1
.endif

fetch-specials:
${MKDIR} ${DISTDIR}/${PORTNAME}
${MKDIR} ${DIST_NPM_CACHE}
if [ ! -f ${DISTDIR}/${PORTNAME}/package-lock.json ]; then ${CP} 
${FILESDIR}/package-lock.json ${DISTDIR}/${PORTNAME}; \
${CP} ${FILESDIR}/package.json ${DISTDIR}/${PORTNAME}; \
(cd ${DISTDIR}/${PORTNAME} && ${SETENV} 
NPM_CONFIG_CACHE=${DIST_NPM_CACHE} npm install --ignore-scripts) \
fi

post-extract:
(cd ${DISTDIR}/${PORTNAME} && find node_modules | ${PAX} -rw ${WRKSRC})
${MKDIR} ${BUILD_NPM_CACHE}
(cd ${DIST_NPM_CACHE} && find . | ${PAX} -rw ${BUILD_NPM_CACHE})

do-build:
(cd ${WRKSRC} && ${SETENV} ${CONFIGURE_ENV} ${MAKE_ENV} 
NPM_CONFIG_CACHE=${BUILD_NPM_CACHE} npm install --nodedir=${LOCALBASE})

Those things I simpler with yarn. Maybe we could consider using yarn as a basis 
for some way to bring binary node packages into ports tree.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

On Tue, 27 Feb 2018, Marcin Cieslak wrote:

> On Sun, 25 Feb 2018, Yuri wrote:
> 
> > On 02/25/18 05:37, Marcin Cieslak wrote:
> > > Yes, this is my private port that I am using to produce FreeBSD binaries
> > > for node-sass. Getting binary npm modules into our ports tree is another
> > > conversation.
> > > 
> > > The problem here is that a whole thing worked for me before for months
> > > so I am aware of all those limitations for particular build phases
> > > (it took me long to figure out that).
> > 
> > 
> > npm is an extremely volatile technology. Some package might work now, and 
> > then
> > break in a week due to a dependency package update.
> > 
> > It continuously automatically updates files that are downloaded as
> > dependencies.
> > 
> > NodeJS is largely incompatible with the FreeBSD ports system because of this
> > volatility.
> > 
> > NodeJS is also a very insecure technology. It brings files directly from
> > github without any vetting. So if somebody will update some github package
> > with malware, it is extremely likely that next day this malware will end up 
> > on
> > your production servers. There is nobody in between, you have to always 
> > trust
> > hundreds of parties.
> 
> I think I have some idea how we can tame this somewhat without allowing for
> a wild fetch.
> 
> It seems that I need to learn more about the code that checks the completness
> of the distfiles, since "make checksum" insists on redoing things all again:
> 
> # rm -rf distinfo 
> # make makesum
> 

(...)

So I don't know what has changed and why but the temporary fix is to
use "if" to check if the desired files are not already there, and
then proceeding with "post-fetch" only if the files are not found.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

On Sun, 25 Feb 2018, Yuri wrote:

> On 02/25/18 05:37, Marcin Cieslak wrote:
> > Yes, this is my private port that I am using to produce FreeBSD binaries
> > for node-sass. Getting binary npm modules into our ports tree is another
> > conversation.
> > 
> > The problem here is that a whole thing worked for me before for months
> > so I am aware of all those limitations for particular build phases
> > (it took me long to figure out that).
> 
> 
> npm is an extremely volatile technology. Some package might work now, and then
> break in a week due to a dependency package update.
> 
> It continuously automatically updates files that are downloaded as
> dependencies.
> 
> NodeJS is largely incompatible with the FreeBSD ports system because of this
> volatility.
> 
> NodeJS is also a very insecure technology. It brings files directly from
> github without any vetting. So if somebody will update some github package
> with malware, it is extremely likely that next day this malware will end up on
> your production servers. There is nobody in between, you have to always trust
> hundreds of parties.

I think I have some idea how we can tame this somewhat without allowing for
a wild fetch.

It seems that I need to learn more about the code that checks the completness
of the distfiles, since "make checksum" insists on redoing things all again:

# rm -rf distinfo 
# make makesum
# cat distinfo
TIMESTAMP = 1519691985
SHA256 (sass-node-sass-v4.7.2_GH0.tar.gz) = 
21cdea5c6bf73825eaec06e78a0bcc54ed75c0953e05c72fe4b4316d756b9e35
SIZE (sass-node-sass-v4.7.2_GH0.tar.gz) = 398635
# env TERM=dumb make checksum
===>  License MIT accepted by the user
===>   node-sass-4.7.2 depends on file: /usr/local/sbin/pkg - found
===>   node-sass-4.7.2 depends on package: npm>=0 - found
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /usr/ports/distfiles/node-sass
/bin/mkdir -p /usr/ports/distfiles/npm
cp -f /home/saper/sw/FreeBSD/ports/textproc/node-sass/files/package-lock.json 
/usr/ports/distfiles/node-sass
cp -f /home/saper/sw/FreeBSD/ports/textproc/node-sass/files/package.json 
/usr/ports/distfiles/node-sass
(cd /usr/ports/distfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/usr/ports/distfiles/npm npm install --ignore-scripts)
npm WARN lifecycle node-sass@4.7.2~install: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/install.js /usr/ports/distfiles/node-sass
npm WARN lifecycle node-sass@4.7.2~postinstall: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/build.js /usr/ports/distfiles/node-sass
npm WARN prepublish-on-install As of npm@5, `prepublish` scripts are deprecated.
npm WARN prepublish-on-install Use `prepare` for build steps and 
`prepublishOnly` for upload-only.
npm WARN prepublish-on-install See the deprecation note in `npm help scripts` 
for more information.
npm WARN lifecycle node-sass@4.7.2~prepublish: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 not-in-install && node scripts/prepublish.js || in-install 
/usr/ports/distfiles/node-sass
up to date in 1.952s
=> SHA256 Checksum OK for sass-node-sass-v4.7.2_GH0.tar.gz.
# env TERM=dumb make checksum
===>  License MIT accepted by the user
===>   node-sass-4.7.2 depends on file: /usr/local/sbin/pkg - found
===>   node-sass-4.7.2 depends on package: npm>=0 - found
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /usr/ports/distfiles/node-sass
/bin/mkdir -p /usr/ports/distfiles/npm
cp -f /home/saper/sw/FreeBSD/ports/textproc/node-sass/files/package-lock.json 
/usr/ports/distfiles/node-sass
cp -f /home/saper/sw/FreeBSD/ports/textproc/node-sass/files/package.json 
/usr/ports/distfiles/node-sass
(cd /usr/ports/distfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/usr/ports/distfiles/npm npm install --ignore-scripts)
npm WARN lifecycle node-sass@4.7.2~install: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/install.js /usr/ports/distfiles/node-sass
npm WARN lifecycle node-sass@4.7.2~postinstall: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/build.js /usr/ports/distfiles/node-sass
npm WARN prepublish-on-install As of npm@5, `prepublish` scripts are deprecated.
npm WARN prepublish-on-install Use `prepare` for build steps and 
`prepublishOnly` for upload-only.
npm WARN prepublish-on-install See the deprecation note in `npm help scripts` 
for more information.
npm WARN lifecycle node-sass@4.7.2~prepublish: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 not-in-install && node scripts/prepublish.js || in-install 
/usr/ports/distfiles/node-sass
up to date in 1.921s
=> SHA256 Checksum OK for sass-node-sass-v4.7.2_GH0.tar.gz.

So this is not poudriere's fault.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

On Sun, 25 Feb 2018, Adam Weinberger wrote:

> > > (full log:
> > > https://gist.github.com/saper/0be3f9b4e9eb819d3ec00f5273705517#file-node-sass-4-7-2-log
> > > )
> > > 
> > > Why are /bin/mkdir -p and /bin/cp commands invoked again in the "extract"
> > > phase?
> > > 
> > > FreeBSD 11.1-STABLE #10 r321629 running poudriere-3.2.4 (I have updated
> > > poudriere today)
> 
> First of all, just so you know, npm module ports won't be added to the ports
> tree. I just wanted to be clear in case you were planning on submitting it.
> There were a couple threads about this a while back.
> 
> With that out of the way, you're hitting errors for two reasons here.

Yes, this is my private port that I am using to produce FreeBSD binaries
for node-sass. Getting binary npm modules into our ports tree is another 
conversation.

The problem here is that a whole thing worked for me before for months
so I am aware of all those limitations for particular build phases
(it took me long to figure out that).

> The first problem is that npm needs network access, and poudriere prevents
> ports from dialing out. You can work around this with the
> ALLOW_NETWORKING_PACKAGES variable in poudriere.conf.
> 
> The other problem you're hitting is that npm saves a cache, and for some
> reason you're trying to run it from /usr/ports/distfiles/node-sass. Ports
> themselves should never, ever have the ability to create files in
> /usr/ports/distfiles. If you run 'npm install' from ${WRKSRC}, then you can
> do:
>   cd ${WRKSRC} && ${SETENV} HOME=/tmp npm install

Aren't ports allowed to write in /usr/ports/distfiles during the fetch phase?
That would be new to me...

This whole thing was working for me before (with older poudriere/bsd.port.mk 
maybe).
I try to force npm to fetch everything (including dependent modules)
in the fetch phase:

 42 post-fetch:
 43 ${MKDIR} ${DISTDIR}/${PORTNAME}
 44 ${MKDIR} ${NPM_CACHE}
 45 ${CP} ${FILESDIR}/package-lock.json ${DISTDIR}/${PORTNAME}
 46 ${CP} ${FILESDIR}/package.json ${DISTDIR}/${PORTNAME}
 47 (cd ${DISTDIR}/${PORTNAME} && ${SETENV} 
NPM_CONFIG_CACHE=${NPM_CACHE} npm install --ignore-scripts)

I copy everything I need from distfiles to ${WRKSRC} in the extract phase:

 49 post-extract:
 50 (cd ${DISTDIR}/${PORTNAME} && ${FIND} node_modules | ${PAX} -rw 
${WRKSRC})
 51 

I build everything without touching the network:

 52 do-build:
 53 (cd ${WRKSRC} && ${SETENV} ${CONFIGURE_ENV} ${MAKE_ENV} 
NPM_CONFIG_CACHE=${NPM_CACHE} npm install --nodedir=${LOCALBASE})

The problem in particular is that "post-fetch" target is invoked ALSO during 
the "extract" target,
I don't know why:

===
   
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
 
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm

It did that already in the fetch phase:

===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm

and it tries that again in the checksum phase:

===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm

Maybe there is an additional check for a successful fetch that I have missed.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

On Sun, 25 Feb 2018, Marcin Cieslak wrote:

> Hello,
> 
> why is poudriere on 11.1-STABLE trying to build by custom port which has the 
> following
> in the Makefile:
> 
> post-fetch:
>   ${MKDIR} ${DISTDIR}/${PORTNAME}
>   ${MKDIR} ${NPM_CACHE}
>   ${CP} ${FILESDIR}/package-lock.json ${DISTDIR}/${PORTNAME}
>   ${CP} ${FILESDIR}/package.json ${DISTDIR}/${PORTNAME}
>   (cd ${DISTDIR}/${PORTNAME} && ${SETENV} NPM_CONFIG_CACHE=${NPM_CACHE} 
> npm install --ignore-scripts)
> 
> ( complete port: 
> https://github.com/saper/ports-exp/tree/master/textproc/node-sass )
> 
> The results are those:
> 
( this was wrong, the full log is also corrected )

===>   Returning to build of node-sass-4.7.2
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
/bin/cp /usr/ports/textproc/node-sass/files/package.json 
/portdistfiles/node-sass
(cd /portdistfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
npm WARN lifecycle node-sass@4.7.2~install: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/install.js /portdistfiles/node-sass
npm WARN lifecycle node-sass@4.7.2~postinstall: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/build.js /portdistfiles/node-sass
npm WARN prepublish-on-install As of npm@5, `prepublish` scripts are deprecated.
npm WARN prepublish-on-install Use `prepare` for build steps and 
`prepublishOnly` for upload-only.
npm WARN prepublish-on-install See the deprecation note in `npm help scripts` 
for more information.
npm WARN lifecycle node-sass@4.7.2~prepublish: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 not-in-install && node scripts/prepublish.js || in-install 
/portdistfiles/node-sass
added 356 packages in 10.684s
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
/bin/cp /usr/ports/textproc/node-sass/files/package.json 
/portdistfiles/node-sass
(cd /portdistfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
npm WARN lifecycle node-sass@4.7.2~install: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/install.js /portdistfiles/node-sass
npm WARN lifecycle node-sass@4.7.2~postinstall: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 node scripts/build.js /portdistfiles/node-sass
npm WARN prepublish-on-install As of npm@5, `prepublish` scripts are deprecated.
npm WARN prepublish-on-install Use `prepare` for build steps and 
`prepublishOnly` for upload-only.
npm WARN prepublish-on-install See the deprecation note in `npm help scripts` 
for more information.
npm WARN lifecycle node-sass@4.7.2~prepublish: cannot run in wd %s %s (wd=%s) 
node-sass@4.7.2 not-in-install && node scripts/prepublish.js || in-install 
/portdistfiles/node-sass
npm WARN ajv-keywords@1.5.1 requires a peer of ajv@>=4.10.0 but none is 
installed. You must install peer dependencies yourself.

up to date in 2.13s
=> SHA256 Checksum OK for sass-node-sass-v4.7.2_GH0.tar.gz.
===
===
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
cp: /portdistfiles/node-sass/package-lock.json: Permission denied
*** Error code 1

Stop.
make: stopped in /usr/ports/textproc/node-sass
=>> Cleaning up wrkdir
===>  Cleaning for node-sass-4.7.2
build of textproc/node-sass | node-sass-4.7.2 ended at Sun Feb 25 02:33:55 UTC 
2018


> (full log: 
> https://gist.github.com/saper/0be3f9b4e9eb819d3ec00f5273705517#file-node-sass-4-7-2-log
>  )
> 
> Why are /bin/mkdir -p and /bin/cp commands invoked again in the "extract" 
> phase?
> 
> FreeBSD 11.1-STABLE #10 r321629 running poudriere-3.2.4 (I have updated 
> poudriere today)
> 
> What's wrong? That part of port code didn't change for some time.
> 
> Marcin


smime.p7s
Description: S/MIME Cryptographic Signature

poudriere: "Permission denied" in the extract phase?

[Marcin Cieslak]

Hello,

why is poudriere on 11.1-STABLE trying to build by custom port which has the 
following
in the Makefile:

post-fetch:
${MKDIR} ${DISTDIR}/${PORTNAME}
${MKDIR} ${NPM_CACHE}
${CP} ${FILESDIR}/package-lock.json ${DISTDIR}/${PORTNAME}
${CP} ${FILESDIR}/package.json ${DISTDIR}/${PORTNAME}
(cd ${DISTDIR}/${PORTNAME} && ${SETENV} NPM_CONFIG_CACHE=${NPM_CACHE} 
npm install --ignore-scripts)

( complete port: 
https://github.com/saper/ports-exp/tree/master/textproc/node-sass )

The results are those:

===>   Returning to build of node-sass-4.7.2
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
(cd /portdistfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
npm WARN saveError ENOENT: no such file or directory, open 
'/portdistfiles/node-sass/package.json'
npm WARN enoent ENOENT: no such file or directory, open 
'/portdistfiles/node-sass/package.json'
npm WARN node-sass No description
npm WARN node-sass No repository field.
npm WARN node-sass No README data
npm WARN node-sass No license field.

added 4 packages in 1.185s
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
(cd /portdistfiles/node-sass && /usr/bin/env 
NPM_CONFIG_CACHE=/portdistfiles/npm npm install --ignore-scripts)
npm WARN saveError ENOENT: no such file or directory, open 
'/portdistfiles/node-sass/package.json'
npm WARN enoent ENOENT: no such file or directory, open 
'/portdistfiles/node-sass/package.json'
npm WARN node-sass No description
npm WARN node-sass No repository field.
npm WARN node-sass No README data
npm WARN node-sass No license field.

up to date in 0.514s
=> SHA256 Checksum OK for sass-node-sass-v4.7.2_GH0.tar.gz.
===
===
===
===
===>  License MIT accepted by the user
===> Fetching all distfiles required by node-sass-4.7.2 for building
/bin/mkdir -p /portdistfiles/node-sass
/bin/mkdir -p /portdistfiles/npm
/bin/cp /usr/ports/textproc/node-sass/files/package-lock.json 
/portdistfiles/node-sass
cp: /portdistfiles/node-sass/package-lock.json: Permission denied
*** Error code 1

Stop.
make: stopped in /usr/ports/textproc/node-sass

(full log: 
https://gist.github.com/saper/0be3f9b4e9eb819d3ec00f5273705517#file-node-sass-4-7-2-log
 )

Why are /bin/mkdir -p and /bin/cp commands invoked again in the "extract" phase?

FreeBSD 11.1-STABLE #10 r321629 running poudriere-3.2.4 (I have updated 
poudriere today)

What's wrong? That part of port code didn't change for some time.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: monitoring ruby app memory usage with passenger-recycler (was: Cannot require library for passenger in ruby)

[Marcin Cieslak]

On Fri, 23 Feb 2018, Karli Sjöberg wrote:

> On Thu, 2018-02-22 at 18:28 +0000, Marcin Cieslak wrote:
> > > than
> > > nothing. It now runs and does what you´d expect :)
> > > 
> > > I have attached it to this email.
> > Thanks! Now we should try to integrate it into the passenger itself
> > :)
> 
> No it´s just 'foreman_maintain' that needs fixing. I´m actually d
> oing the git dance right now getting it accepted upstream :)
> 
> https://github.com/theforeman/foreman_maintain/pull/146

Wish you luck :)

I have pushed https://github.com/theforeman/foreman_maintain/pull/143 a bit 
earlier.

(I still think that most of your code should go into passenger, but some 
refactoring
might be needed there first).

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: monitoring ruby app memory usage with passenger-recycler (was: Cannot require library for passenger in ruby)

[Marcin Cieslak]

On Thu, 22 Feb 2018, Karli Sjöberg wrote:

> > How do I interpret the output of 'procstat -v ' though? Can I
> > sum
> > up all of the RES or PRES numbers to get the total virtual memory
> > that
> > this process is consuming?
> 
> I hacked at it a bit and came up with an updated patch that includes
> the changes you made, plus my own.
> 
> Since "private dirty RSS" isn´t available, I´ve used "maximum RSS" from
> 'procstat -r ', which is far from perfect, but it´s better than
> nothing. It now runs and does what you´d expect :)
> 
> I have attached it to this email.

Thanks! Now we should try to integrate it into the passenger itself :)

I think you might be looking for "private resident pages" (PRES),
but I am not sure. 

For your particular problem I would log procstat -v output in its entirety
at regular intervals and see where the memory grows.

Happy you got there somehow :)

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: monitoring ruby app memory usage with passenger-recycler (was: Cannot require library for passenger in ruby)

[Marcin Cieslak]

On Mon, 19 Feb 2018, Karli Sjöberg wrote:

> > What is the tool you are trying to deploy? foreman_maintain?

Sorry, I was too quick. This foreman_main is used to download/update foreman
from RedHat Satellite servers, which is not what you want.

However, the passenger-recycler script is a standalone script which
could possibly used with this little fix:

https://github.com/theforeman/foreman_maintain/pull/143

To apply this, please add files/patch-bin_passenger-recycler in the port I've 
sent before:

--- bin/passenger-recycler.orig 2018-02-20 21:09:18 UTC
+++ bin/passenger-recycler
@@ -41,10 +41,10 @@ def process_status?(pid)
 end
 
 require 'phusion_passenger'
+PhusionPassenger.locate_directories
 require 'phusion_passenger/platform_info'
 require 'phusion_passenger/platform_info/ruby'
 require 'phusion_passenger/admin_tools/memory_stats'
-PhusionPassenger.locate_directories
 stats = PhusionPassenger::AdminTools::MemoryStats.new
 unless stats.platform_provides_private_dirty_rss_information?
   puts 'Please run as root or platform unsupported'


But the real problem is that Passenger does think it supports
getting process private dirty RSS memory information only on Linux:

https://github.com/phusion/passenger/blob/219ad24159ae4033a342e6ad753cfee05d98bae0/src/ruby_supportlib/phusion_passenger/admin_tools/memory_stats.rb#L121

In file src/ruby_supportlib/phusion_passenger/admin_tools/memory_stats.rb:

120   def platform_provides_private_dirty_rss_information?
121 return os_name_simple == "linux"
122   end

This support probably would need to be written. So, unfortunately, 
passenger-recycler
as written will exit with:

Please run as root or platform unsupported

But passenger-recycler does something very simple - if the amount of "private 
resident RSS
memory" is higher than specified amount in the configuration file, it kills the 
process.

You might get a much better monitoring on FreeBSD by monitoring the output of

procstat -v 

which will give you details about the memory usage of the process and see how 
it is growing
over time. Maybe you can pinpoint the shared library that causes this, if you 
are lucky.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot require library for passenger in ruby

[Marcin Cieslak]

On Mon, 19 Feb 2018, Karli Sjöberg wrote:

> > What is the tool you are trying to deploy? foreman_maintain?

Can you try the port below? Unfortunately I do not see how foreman_maintain may 
fix your passenger
installation issues, but here it is:

# This is a shell archive.  Save it in a file, remove anything before
# this line, and then unpack it by entering "sh file".  Note, it may
# create directories; files and directories will be owned by you and
# have default permissions.
#
# This archive contains:
#
#   rubygem-foreman_maintain
#   rubygem-foreman_maintain/Makefile
#   rubygem-foreman_maintain/distinfo
#   rubygem-foreman_maintain/pkg-descr
#
echo c - rubygem-foreman_maintain
mkdir -p rubygem-foreman_maintain > /dev/null 2>&1
echo x - rubygem-foreman_maintain/Makefile
sed 's/^X//' >rubygem-foreman_maintain/Makefile << 
'15090155e0ad279a3a88902efd75aea3'
X# $FreeBSD: head/devel/rubygem-foreman/Makefile 448125 2017-08-17 15:05:14Z 
sunpoet $
X
XPORTNAME=  foreman_maintain
XPORTVERSION=   0.1.3
XCATEGORIES=devel rubygems
XMASTER_SITES=  RG
X
XMAINTAINER=r...@freebsd.org
XCOMMENT=   Facility to keep Foreman/Satellite services running
X
XRUN_DEPENDS=   rubygem-clamp>0:devel/rubygem-clamp \
X   rubygem-highline>0:devel/rubygem-highline \
X   rubocop:devel/rubygem-rubocop \
X   pry:devel/rubygem-pry
X
XNO_ARCH=   yes
XUSES=  gem shebangfix
XUSE_RUBY=  yes
X
XPLIST_FILES=   bin/foreman-maintain bin/passenger-recycler
X
XSHEBANG_FILES= ${PLIST_FILES}
X
X.include 
15090155e0ad279a3a88902efd75aea3
echo x - rubygem-foreman_maintain/distinfo
sed 's/^X//' >rubygem-foreman_maintain/distinfo << 
'b2bf4dcb1a0a898485a2b273846fd43e'
XTIMESTAMP = 1519156673
XSHA256 (rubygem/foreman_maintain-0.1.3.gem) = 
35f10c7df685072de50f515bfe5c694791d01f5d2eb2ef02c32b693f7e68f73b
XSIZE (rubygem/foreman_maintain-0.1.3.gem) = 58880
b2bf4dcb1a0a898485a2b273846fd43e
echo x - rubygem-foreman_maintain/pkg-descr
sed 's/^X//' >rubygem-foreman_maintain/pkg-descr << 
'9944d73c5ee733bda3fa3a0bea6fbf70'
XA maintenance tool for Foreman/Satellite Ruby applications
X
XWWW: https://github.com/theforeman/foreman_maintain
9944d73c5ee733bda3fa3a0bea6fbf70
exit


smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot require library for passenger in ruby

[Marcin Cieslak]

On Mon, 19 Feb 2018, Karli Sjöberg wrote:

> 
> 
> Den 18 feb. 2018 23:49 skrev Marcin Cieslak <sa...@saper.info>:
>   On Sat, 17 Feb 2018, Karli Sjöberg via freebsd-ports wrote:
>   CONFIG
>   > 
> /usr/local/lib/ruby/gems/2.4/gems/foreman_maintain-0.1.3/bin/passenger-recycler:8:
>  warning: previous definition of CONFIG
>was here
>   > 
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:135:in 
> `require': cannot load such file --
>   /phusion_passenger/utils/tmpio (LoadError)
>   > from 
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:135:in 
> `rescue in require'
>   > from 
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:39:in 
> `require'
>   > from 
> /usr/local/lib/ruby/gems/2.4/gems/passenger-5.2.0/src/ruby_supportlib/phusion_passenger.rb:240:in
>   `require_passenger_lib'
>   > from 
> /usr/local/lib/ruby/gems/2.4/gems/passenger-5.2.0/src/ruby_supportlib/phusion_passenger/platform_info.rb:26:in
>   `'
> 
>   Where did you get passenger 5.2.0?
> 
>   Can you try to install passenger 5.1.12 from ports and see if it works 
> better?
> 
> 
> Right, yeah, no, I started with 5.1.12 from ports and when it didn't work, I 
> installed a newer version with 'gem'. I can deinstall it
> again if you want to see the backtrace?

Would be good, please. I wonder how many passenger dependencies got installed 
via rubygems - would be good to
have them restored, too. What is the tool you are trying to deploy? 
foreman_maintain?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Cannot require library for passenger in ruby

[Marcin Cieslak]

On Sat, 17 Feb 2018, Karli Sjöberg via freebsd-ports wrote:

> Hey all!
> 
> My hope for this email is that someone will give me a smack in the
> right direction, like 'You´re holding it wrong!'  and that´ll be that,
> but you never know...
> 
> So the background story of it all is that I have a Foreman server that
> is eating all of my RAM, basically no matter how much I give it. I´ve
> written to the devs and they say it´s wrong, it shouldn´t be doing
> that, it´s not a feature :) So they gave me a ruby script that fires
> from cron every 5 mins that checks if there are any passenger processes
> that are going bananas and kills them. It´s not a cure, just a
> workaround, but it´s going to give some breathing room for diagnosing
> the root cause in the mean time. Problem is, I can´t get it running!
> 
> /usr/local/lib/ruby/gems/2.4/gems/foreman_maintain-0.1.3/bin/passenger-recycler:10:
>  warning: already initialized constant CONFIG
> /usr/local/lib/ruby/gems/2.4/gems/foreman_maintain-0.1.3/bin/passenger-recycler:8:
>  warning: previous definition of CONFIG was here
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:135:in 
> `require': cannot load such file -- /phusion_passenger/utils/tmpio (LoadError)
> from 
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:135:in 
> `rescue in require'
> from 
> /usr/local/lib/ruby/site_ruby/2.4/rubygems/core_ext/kernel_require.rb:39:in 
> `require'
> from 
> /usr/local/lib/ruby/gems/2.4/gems/passenger-5.2.0/src/ruby_supportlib/phusion_passenger.rb:240:in
>  `require_passenger_lib'
> from 
> /usr/local/lib/ruby/gems/2.4/gems/passenger-5.2.0/src/ruby_supportlib/phusion_passenger/platform_info.rb:26:in
>  `'

Where did you get passenger 5.2.0? 

Can you try to install passenger 5.1.12 from ports and see if it works better?

It looks to my that there is some trouble to figure out the location of Ruby 
libraries...

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: DotNet Core on FreeBSD

[Marcin Cieslak]

On Sun, 10 Sep 2017, Russell Haley wrote:

> Hey guys,
> 
> I am talking to Karel and Tomas at Microsoft's DotNet Core Team about
> putting together a "proposal" for community involvement, which will be
> followed by more information on this list. Most of the talk has been
> me blowing air and being pedantic about nothing (I get excited and
> type alot).

I was tinkering around FreeBSD support when dotnet core was first published.
Stack unwinding for exceptions was very hacky at the time, and there were
some unnecessary discussions about how to implement some FreeBSD sepecific
sysctl's. I got a bit frustrated with that (a whole porting effort is about
coercing Unix to offer part of Win32 APIs).

It was somewhat running on FreeBSD natively back then, haven't checked
recently.

I have also decided not to sign their bad contributor agreement
and some of the sysctl code I decided to put in mono instead :)

Marcin
saper on github

smime.p7s
Description: S/MIME Cryptographic Signature

How to make www/npm pick www/node4 and www/node6 instead of www/node?

[Marcin Cieslak]

Hello,

in the past (before r414303[1])
[1] 
https://svnweb.freebsd.org/ports/head/www/npm/Makefile?view=log=414303

npm had a couple of OPTIONS to select which node engine should be used.

Now this is all gone and replaced with

RUN_DEPENDS=node>=0.8.0:www/node


I am trying to build node-sass binaries for FreeBSD (I used to publish them in 
the past)
and that requires running poudriere against multiple engines.

The following attempt to cheat does not help:

 poudriere bulk -j node4_10_3_i386 -p exp www/node4 textproc/node-sass

(textproc/node-sass is my custom port 
https://github.com/saper/ports-exp/tree/master/textproc/node-sass)

since poudriere starts to build www/node4 and www/node in parallel.


Would that help if the npm dependency were changed to

RUN_DEPENDS=node:www/node

so that only existing executable is needed? How could I tell poudriere to pick 
node4 first?

I used to maintain a private copy of the npm port with lots of OPTIONS and this 
is a PITA.



I have separate poudriere jails for all architectures I have decided to support:

$ poudriere jail -ln | grep ^node
node4_10_3_amd64   
node4_10_3_i386
node6_10_3_amd64   
node6_10_3_i386
node8_10_3_amd64   
node8_10_3_i386

In the past those had OPTIONS set to pick a proper engine as a www/npm 
dependency.


How to do it cleanly now?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: poudriere cross building to arm64.aarch64: Don't know how to make native-xtools

[Marcin Cieslak]

On Sat, 10 Jun 2017, Bryan Drewery wrote:

> On 6/10/17 1:14 PM, Marcin Cieslak wrote:
> > When running poudriere 3.1.17_1 on FreeBSD 11.1-PRERELEASE r319534 on amd64:
> > 
> > poudriere jail -c -j ksh-test-aarch64 -m svn -x -v stable/11 -a 
> > arm64.aarch64
> > 
> > I am getting an error that "make native-xtools" does not seem to be a 
> > failed target.
> > 
> > Unfortunately, I cannot see the state of the filesystem when it fails, 
> > because
> > zfs dataset gets removed on failure.
> > 
> > Any ideas where to look?
> > 
> > Marcin
> > 
> 
> It is trying to run 'make native-xtools' in your *host* /usr/src.  You
> need a source checkout matching your current system there and having
> already done a 'make buildworld' for it.
> 
> This is a bug in the native-xtools build that poudriere is working
> around until it is fixed.

Thank you, I am using poudriere in a small Xen VM which didn't even had 
/usr/src or /usr/obj
mounted.

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

poudriere cross building to arm64.aarch64: Don't know how to make native-xtools

[Marcin Cieslak]

When running poudriere 3.1.17_1 on FreeBSD 11.1-PRERELEASE r319534 on amd64:

poudriere jail -c -j ksh-test-aarch64 -m svn -x -v stable/11 -a arm64.aarch64

I am getting an error that "make native-xtools" does not seem to be a failed 
target.

Unfortunately, I cannot see the state of the filesystem when it fails, because
zfs dataset gets removed on failure.

Any ideas where to look?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Hosting distfiles on HTTPS w/Let's Encrypt - how? [somehow solved]

[Marcin Cieslak]

On Mon, 5 Jun 2017, Matthew D. Fuller wrote:

> On Sun, Jun 04, 2017 at 09:48:02PM + I heard the voice of
> Marcin Cieslak, and lo! it spake thus:
> > 
> > My temporary solution to this problem is to pin the CA certificate
> > in the port itself:
> 
> Err...
> 
> > -FETCH_ENV= HTTP_AUTH=basic:*:I\ accept\ 
> > www.opensource.org/licenses/cpl:.
> > +FETCH_ARGS+=   --ca-cert="${FILESDIR}/dst_root_ca_x3.crt"
> 
> bsd.port.mk already sets
> 
> FETCH_ENV?= SSL_NO_VERIFY_PEER=1 SSL_NO_VERIFY_HOSTNAME=1
> 
> itself (on !makesum).  If you don't need that FETCH_ENV at all, you
> wouldn't need the _ARGS either (and if you do need the _ENV, you'd
> probably want to pull in the default as well to match...)

That was it, thank you very much!

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

poudriere cross building to arm64.aarch64: Don't know how to make native-xtools

[Marcin Cieslak]

When running poudriere 3.1.17_1 on FreeBSD 11.1-PRERELEASE r319534 on amd64:

poudriere jail -c -j ksh-test-aarch64 -m svn -x -v stable/11 -a arm64.aarch64

I am getting an error that "make native-xtools" does not seem to be a failed 
target.

Unfortunately, I cannot see the state of the filesystem when it fails, because
zfs dataset gets removed on failure.

Any ideas where to look?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Hosting distfiles on HTTPS w/Let's Encrypt - how? [somehow solved]

[Marcin Cieslak]

On Thu, 1 Jun 2017, Marcin Cieslak wrote:

> => Attempting to fetch 
> https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz
> Certificate verification failed for /C=US/O=Let's Encrypt/CN=Let's Encrypt 
> Authority X3
> 34374329736:error:14090086:SSL 
> routines:ssl3_get_server_certificate:certificate verify 
> failed:/usr/src/secure/lib/libssl/../../../crypto/openssl/ssl/s3_clnt.c:1264:
> fetch: https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz: 
> Authentication error

My temporary solution to this problem is to pin the CA certificate in the port 
itself:

commit 7eec5787c09565b0b2dfc4b2cee176c8509474b2
Author: Marcin Cieślak <sa...@saper.info>
Date:   Sun Jun 4 21:31:22 2017 +

Hardwire CA certificate to facilitate HTTPS downloads

Ports do not have a public key infrastructure to facilitate
ports that need to be fetched using https only.

So we hardcode a root certificate used by Let's Encrypt
for now.

diff --git a/shells/ksh93/Makefile b/shells/ksh93/Makefile
index 10f826c..c1ddef2 100644
--- a/shells/ksh93/Makefile
+++ b/shells/ksh93/Makefile
@@ -24,7 +24,7 @@ LICENSE=  EPL
 
 OPTIONS_DEFINE=EXAMPLES STATIC
 
-FETCH_ENV= HTTP_AUTH=basic:*:I\ accept\ www.opensource.org/licenses/cpl:.
+FETCH_ARGS+=   --ca-cert="${FILESDIR}/dst_root_ca_x3.crt"
 LDFLAGS+=  -lm
 MAKE_ENV=  CCFLAGS="${CFLAGS}"
 NO_WRKSUBDIR=  yes
diff --git a/shells/ksh93/files/dst_root_ca_x3.crt 
b/shells/ksh93/files/dst_root_ca_x3.crt
new file mode 100644
index 000..e2bd36f
--- /dev/null
+++ b/shells/ksh93/files/dst_root_ca_x3.crt
@@ -0,0 +1,22 @@
+subject=/O=Digital Signature Trust Co./CN=DST Root CA X3
+issuer=/O=Digital Signature Trust Co./CN=DST Root CA X3
+-BEGIN CERTIFICATE-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+-END CERTIFICATE-

https://github.com/saper/ports-exp/commit/7eec5787c09565b0b2dfc4b2cee176c8509474b2



smime.p7s
Description: S/MIME Cryptographic Signature

Re: Hosting distfiles on HTTPS w/Let's Encrypt - how?

[Marcin Cieslak]

On Thu, 1 Jun 2017, Adam Weinberger wrote:

> I've tried fetching a distfile from my own server (which uses a Let's Encrypt 
> cert) and it fetches fine in a poudriere jail. I'm suspecting that there's 
> something unusual in your web server's SSL configuration, or in how you're 
> generating your LE cert. Do you have any interesting arguments that you're 
> giving dehydrated or your web server?

The only unusual thing in my certificate is that CN belongs to another domain 
and the domain in question
is listed in the subjectAltName along with a primary.

On a system with certificate bundle installed the following works fine:

fetch https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz

My port (https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211164) has barely 
any dependencies, and there is
no certificate bundle in the jail. Adam - can you check if something installs 
NSS CA roots as a dependency in your jail?

I think I understand what happens - bare FreeBSD installation has no CA 
bundles, therefore fetch cannot really
do https. Most ports work either because one of the dependencies installs ca 
root nss or they have a plain HTTP
fallback (from distcache if need be). My distfiles are brand new and the 
distcache does not know them, not there is
any HTTP fallback.

The question is: do we silently require at least one unencrypted HTTP or FTP 
distfile source?
If not, what should be done to bootstrap certificates for fetch - include somme 
roots in base,
turn off certificate validation, other options?

Marcin

smime.p7s
Description: S/MIME Cryptographic Signature

Re: Hosting distfiles on HTTPS w/Let's Encrypt - how?

[Marcin Cieslak]

On Thu, 1 Jun 2017, Freddie Cash wrote:

> In your web server configuration, are you using the Let's Encrypt cert.pem
> or fullchain.pem?

fullchain.pem

> If you use the former, then any client that doesn't have the DST Root CA
> pre-installed will error out. The latest versions of browsers will work, as
> they include the DST Root CA.

My fullchain.pem as delivered by dehydrated does not include the DST Root CA.

> If you use the latter, then it will just work, as the server will send all
> the intermediate certificate info needed to reach the root.

To test this theory, I have added DST Root CA to my customized fullchain.pem
which now contains:

Certificate chain
 0 s:/CN=marcincieslak.com
   i:/C=US/O=Let's Encrypt/CN=Let's Encrypt Authority X3

 1 s:/C=US/O=Let's Encrypt/CN=Let's Encrypt Authority X3
   i:/O=Digital Signature Trust Co./CN=DST Root CA X3

 2 s:/O=Digital Signature Trust Co./CN=DST Root CA X3
   i:/O=Digital Signature Trust Co./CN=DST Root CA X3

so now we have "DST Root CA X3" extra.

And the result is:

=> INIT.2014-12-24.tgz doesn't seem to exist in /portdistfiles/ksh93.
=> Attempting to fetch 
https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz
Certificate verification failed for /O=Digital Signature Trust Co./CN=DST Root 
CA X3
34374329736:error:14090086:SSL routines:ssl3_get_server_certificate:certificate 
verify 
failed:/usr/src/secure/lib/libssl/../../../crypto/openssl/ssl/s3_clnt.c:1264:
fetch: https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz: 
Authentication error
=> Attempting to fetch 
http://distcache.FreeBSD.org/ports-distfiles/ksh93/INIT.2014-12-24.tgz
fetch: http://distcache.FreeBSD.org/ports-distfiles/ksh93/INIT.2014-12-24.tgz: 
Not Found

so it cannot validate "DST Root CA X3" now, because it does not have the 
pre-installed CA bundle.


Marcin Cieślak

smime.p7s
Description: S/MIME Cryptographic Signature

Hosting distfiles on HTTPS w/Let's Encrypt - how?

[Marcin Cieslak]

Hello,

I have posted my port's local distfiles to a machine
that is serving them with SSL behind the Let's Encrypt
certificate (https://distfile.net). This is SSL-only.

However, poudriere fails on certificate check when trying
to fetch it:

===
===>  License EPL accepted by the user
===
===
===>   ksh93-20160716 depends on file: /usr/local/sbin/pkg - not found
===>   Installing existing package /packages/All/pkg-1.10.1.txz
[ksh-test-amd64-exp-job-01] Installing pkg-1.10.1...
[ksh-test-amd64-exp-job-01] Extracting pkg-1.10.1: .. done
===>   ksh93-20160716 depends on file: /usr/local/sbin/pkg - found
===>   Returning to build of ksh93-20160716
===
===
===
===
===>  License EPL accepted by the user
=> INIT.2014-12-24.tgz doesn't seem to exist in /portdistfiles/ksh93.
=> Attempting to fetch 
https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz
Certificate verification failed for /C=US/O=Let's Encrypt/CN=Let's Encrypt 
Authority X3
34374329736:error:14090086:SSL routines:ssl3_get_server_certificate:certificate 
verify 
failed:/usr/src/secure/lib/libssl/../../../crypto/openssl/ssl/s3_clnt.c:1264:
fetch: https://distfile.net/local-ports-distfiles/INIT.2014-12-24.tgz: 
Authentication error
=> Attempting to fetch 
http://distcache.FreeBSD.org/ports-distfiles/ksh93/INIT.2014-12-24.tgz
fetch: http://distcache.FreeBSD.org/ports-distfiles/ksh93/INIT.2014-12-24.tgz: 
Not Found
=> Couldn't fetch it - please try to retrieve this
=> port manually into /portdistfiles/ksh93 and try again.
*** Error code 1

What is the best solution here?

so I really have to add security/ca_root_nss (... and perl)
as a fetch dependency? Any other solution?

A quick look at bsd.sites.mk shows that we have some https-only distfile
sources.

Marcin Cieślak


smime.p7s
Description: S/MIME Cryptographic Signature

Interest in lang/v8-devel upgrade?

[Marcin Cieslak]

I just got v8 git master working
on FreeBSD 11 (with clang 3.7
and https://crrev.com/75e43a6681e01de9422a0211f502723d23b51247)

Is there an interest in having a 4.x series
of V8 available in ports? We could follow
www/iojs for example which is currently at
V8 4.4.63.26

~Marcin
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Re: lang/mono: Update to 4.0.3.19

[Marcin Cieslak]

On Sun, 9 Aug 2015, Kurt Jaeger wrote:

 Hi!
 
  Ok, got that. As Romain is the maintainer of the port, he needs
  to approve PR 201883
 
 Technically, after 14 days oe non-action on a PR, the maintainer is
 seen as timed-out for that PR and anyone can commit. This is almost
 the case with that PR.

We have mono@ mailing list, which is the maintainer. (I'd suggest
moving this thread there).

 There are other issues to consider:
 
 I used
 
 grep -Rl '{PORTSDIR}/lang/mono' *  /tmp/mono

We have USES=mono

The mailing list is full of automated complaints about
failing dependencies. There are also human complaints
that monodevelop is old and/or doesn't work.

Not sure what is the official policy but you can
be maintainer *right now* on behalf of mono@

I also would love some help in getting

https://github.com/mono/mono/pull/1870

committed to mono and getting

https://github.com/mono/mono/pull/1859
https://github.com/mono/mono/pull/1860
https://github.com/mono/mono/pull/1861
https://github.com/mono/mono/pull/1863
https://github.com/mono/mono/pull/1864
https://github.com/mono/mono/pull/1865
https://github.com/mono/mono/pull/1866
https://github.com/mono/mono/pull/1867

backported to 4.0.3

as well as help with how to proceed with

https://github.com/mono/mono/pull/1882

futher - maybe by getting mono working with
system iconv and not the GNU one - see
http://lists.freebsd.org/pipermail/freebsd-ports/2015-July/099675.html

~Marcin


___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Proper way to get base iconv.h with autoconf?

[Marcin Cieslak]

[X-posted, replies set to po...@freebsd.org]

Hello,

I have managed to upstream few things needed
to make mono work on FreeBSD without local patches.

Now I am stuck with the iconv detection code.

Mono (or it's dumbled-down glib called eglib) need
libiconv for two things - charset conversion:


https://github.com/mono/mono/blob/master/eglib/src/giconv.c

and getting default charset information

https://github.com/mono/mono/blob/master/eglib/src/gunicode.c#L223

The latter can be solved by using nl_langinfo (CODESET);
and avoiding the need to link libcharset.so


Mono uses standard iconv.m4 autoconf test, version 11
https://github.com/mono/mono/blob/master/eglib/m4/iconv.m4

which is developed in the gettext and gnulib (The GNU portability
library):

http://git.savannah.gnu.org/cgit/gnulib.git/log/m4/iconv.m4

http://git.savannah.gnu.org/cgit/gettext.git/log/gettext-runtime/m4/iconv.m4

Whenever -I/usr/local/include is added to CPPFLAGS for
some reason (some other library, or jest using /usr/local as
default prefix) the configure script fails to use base header
and libc support and needs using GNU iconv as a dependency;
because this simple test fails:


configure:13910: checking for iconv
configure:13934: cc -o conftest -g -O2 -g -D_GNU_SOURCE -I/usr/local/include  
conftest.c  5
/tmp/conftest-ee6d62.o: In function `main':
/home/saper/sw/mono/eglib/conftest.c:42: undefined reference to `libiconv_open'
/home/saper/sw/mono/eglib/conftest.c:43: undefined reference to `libiconv'
/home/saper/sw/mono/eglib/conftest.c:44: undefined reference to `libiconv_close'
cc: error: linker command failed with exit code 1 (use -v to see invocation)
configure:13934: $? = 1
configure: failed program was:
| /* confdefs.h */
| #define PACKAGE_NAME eglib
| #define PACKAGE_TARNAME eglib
| #define PACKAGE_VERSION 0.3
| #define PACKAGE_STRING eglib 0.3
| #define PACKAGE_BUGREPORT 
http://bugzilla.xamarin.com/enter_bug.cgi?classification=Mono;

--snip--

| /* end confdefs.h.  */
|
| #include stdlib.h
| #include iconv.h
|
| int
| main ()
| {
| iconv_t cd = iconv_open(,);
|iconv(cd,NULL,NULL,NULL,NULL);
|iconv_close(cd);
|   ;
|   return 0;
| }
configure:13958: cc -o conftest -g -O2 -g -D_GNU_SOURCE -I/usr/local/include  
conftest.c  /usr/local/lib/libiconv.so -Wl,-rpath -Wl,/usr/local/lib 5
configure:13958: $? = 0
configure:13968: result: yes

The iconv.h test fails because /usr/local/include/iconv.h
redefines iconv_open (and friends) to libiconv_open 

Any ideas how to fix/improve the iconv.m4 detection code
- to use the FreeBSD's base iconv.h and not require a library?

~Marcin



___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Re: www/iojs 2.0.0 is there

[Marcin Cieslak]

On Wed, 6 May 2015, Daniel Lin wrote:

 Hello, Marcin,
 
 Thanks, but I think iojs updates very often, it would keep many version
 on ports.

You are right. But only 1.0.4, 1.8.1 and now 2.0.0 have distinct binary
module APIs, so extensions compiled for them stay the same.

I currently build binaries of node-sass (http://github.com/sass/node-sass)
for node 0.10 (API 13), node 0.12 (API 14), iojs 1.0.4 (API 42), io.js 1.8.1
(API 43) and io.js 2.0 (API 44). I put my custom ports on GitHub
at http://github.com/saper/ports-exp

Here's the patch for io.js 2.0.0, we need a little fix for 
http://code.google.com/p/v8/issues/detail?id=4079


diff -ruN /usr/ports/www/iojs/Makefile www/iojs44/Makefile
--- /usr/ports/www/iojs/Makefile2015-04-27 14:58:07.0 +
+++ www/iojs44/Makefile 2015-05-06 09:09:32.0 +
@@ -2,7 +2,7 @@
 # $FreeBSD: head/www/iojs/Makefile 384845 2015-04-27 14:58:07Z sunpoet $
 
 PORTNAME=  iojs
-PORTVERSION=   1.8.1
+PORTVERSION=   2.0.0
 CATEGORIES=www
 
 MAINTAINER=lin...@gmail.com
@@ -40,7 +40,10 @@
${WRKSRC}/deps/v8/tools/gyp/v8.gyp
@${REINPLACE_CMD} -e 's|/usr/bin/env python|${PYTHON_CMD}|' \
${WRKSRC}/configure
+   @${REINPLACE_CMD} -e s|python -c|${PYTHON_CMD} -c| \
+   ${WRKSRC}/deps/v8/build/toolchain.gypi
@${REINPLACE_CMD} -e s|'python',|'${PYTHON_CMD}',| \
+   ${WRKSRC}/tools/icu/icu-generic.gyp \
${WRKSRC}/deps/v8/build/shim_headers.gypi \
${WRKSRC}/deps/v8/tools/gyp/v8.gyp \
${WRKSRC}/deps/v8/src/d8.gyp
diff -ruN /usr/ports/www/iojs/distinfo www/iojs44/distinfo
--- /usr/ports/www/iojs/distinfo2015-04-27 14:58:07.0 +
+++ www/iojs44/distinfo 2015-05-05 12:07:25.0 +
@@ -1,2 +1,2 @@
-SHA256 (iojs-io.js-1.8.1-v1.8.1_GH0.tar.gz) = 
90e7e4d801e433a405b61e0a62c9dd516804dd884895602ee22bcb590eca5f13
-SIZE (iojs-io.js-1.8.1-v1.8.1_GH0.tar.gz) = 31172196
+SHA256 (iojs-io.js-2.0.0-v2.0.0_GH0.tar.gz) = 
23ee904d5ea6e7f29ae8c146618e4a61dae274e4237e897075f30a28a243804d
+SIZE (iojs-io.js-2.0.0-v2.0.0_GH0.tar.gz) = 31378750
diff -ruN 
/usr/ports/www/iojs/files/patch-deps_v8_src_compiler_graph_visualizer.h 
www/iojs44/files/patch-deps_v8_src_compiler_graph_visualizer.h
--- /usr/ports/www/iojs/files/patch-deps_v8_src_compiler_graph_visualizer.h 
1970-01-01 00:00:00.0 +
+++ www/iojs44/files/patch-deps_v8_src_compiler_graph_visualizer.h  
2015-05-06 07:50:31.0 +
@@ -0,0 +1,10 @@
+--- deps/v8/src/compiler/graph-visualizer.h.orig   2015-05-06 
07:49:33.0 +
 deps/v8/src/compiler/graph-visualizer.h2015-05-06 07:49:44.0 
+
+@@ -5,6 +5,7 @@
+ #ifndef V8_COMPILER_GRAPH_VISUALIZER_H_
+ #define V8_COMPILER_GRAPH_VISUALIZER_H_
+ 
++#include stdio.h
+ #include iosfwd
+ 
+ namespace v8 {


___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

www/iojs 2.0.0 is there

[Marcin Cieslak]

Unfortunately, similarly to iojs 1.0.4 - 1.1.0
NODE_MODULE_VERSION has been bumped (from 43 to 44)
to indicate potential v8 API breakage.

Maybe we should keep node 1.8.1 around for those who
depend on those binary modules.

I also keep node 1.0.4 in my private repository for
that reason https://github.com/saper/ports-exp/tree/master/www/iojs42

~Marcin
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Re: [CORRECTION] Change to the header in ports Makefiles, take two

[Marcin Cieslak]

 Thomas Abthorpe portmgr-secret...@freebsd.org wrote:
 Sometimes, as a result of a repocopy, or changed maintainership, the
 Created By and MAINTAINER is no longer in synchronisation. To avoid
 confusion, the first line can be removed, optionally leaving us with a one
 line header.

 # $FreeBSD$

 Removing the line of attribution is to be done only at the consent/request
 of the original contributor.

Why bother? I always thought that this leaving the original maintainer
intact is a nice way to say thank you to the person who brought this
piece to FreeBSD.

//Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Re: Zarafa port

[Marcin Cieslak]

 rfl...@acsalaska.net rfl...@acsalaska.net wrote:
 --=_20120205130645_53400
 Content-Type: text/plain; charset=iso-8859-1
 Content-Transfer-Encoding: quoted-printable

 Hi,

 I've started working on a port of Zarafa, following
 http://wiki.freebsd.org/WantedPorts. I'm not sure if I can finish it,
 but I will try my bestest :).

Independently from OpenBSD work (didn't even look at their patches...)
I have prepared a set of portability fixes (that could possibly
added upstream) as well as FreeBSD specific ones. 

They can be fetches from my Git repository on git.zarafa.com:

https://git.zarafa.com/~saper/zcp/zcp-for-saper/

 - portability branch contains generic portability improvements
   that can go upstream.

 - FreeBSD contains FreeBSD specific stuff, including removal
   of dependency on xlocale.

 - master is clean copy of original zcp master branch.

This is very rough stuff, but it compiles fine. Not yet tested.
Because of my lack of autoconf-foo you have to gmake platform.freebsd.o
at some point in the common directory, because it does
not get picked up.

Configure invocation I was using:

env CPPFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib ./configure \
--enable-unicode \
--prefix=/usr/local \
--with-vmime-prefix=/usr/local \
--with-php-config=/usr/local/bin/php-config \
--with-mapi-configdir=/usr/local/etc/mapi \
--with-userscript-prefix=/usr/local/etc/zarafa/userscripts \
--with-quotatemplate-prefix=/usr/local/etc/zarafa/quotamail \
--with-indexerscripts-prefix=/usr/local/etc/zarafa/indexerscripts \
--with-python=/usr/local/bin/python2.7 --enable-swig

//Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Attn: maintainers of rxvt-derived ports: small crash fixed

[Marcin Cieslak]

Hello,

There are some rxvt-derived ports in the tree and I have fixed
a small possibility to crash the emulator when pasting some
text from not-fully-ICCCM compliant client.

Here's patch for x11/rxvt:
http://www.freebsd.org/cgi/query-pr.cgi?pr=156542

for x11/rxvt-devel:
http://www.freebsd.org/cgi/query-pr.cgi?pr=156544

for x11/rxvt-unicode:
http://www.freebsd.org/cgi/query-pr.cgi?pr=156509

The patch is trivial and changes ca. 4 lines of code,
I just don't have time to fix all the rxvt derivatives
right there, notify upstream etc. 

I would also welcome feedback from X11 experts
whether this patch is a RightThing(tm) to do.

//Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

Re: [legal] port with restrictive license

[Marcin Cieslak]

Dnia 11.09.2010 Dmitry Marakasov amd...@amdmi3.ru napisał/a:
 software authors kindly agreed to remove revocation term (which worried
 be the most), however other terms are in question. Could someone review
 that license and say whether that's ok in ports, whether distfile
 mirroring is OK and whether we should modify it to make user download
 file for himself (like with java)?

Looks scary, but there no limitations for redistribution, even
limited derivative work (if one constitues patches in the FreeBSD port
as such) is allowed.

Java says you can't redistribute it (i.e. files) so we can't mirror them.

Ports can also contain a fully proprietary, closed-source software, so
I guess you are ok with it. 

//Marcin (speaking only for myself and in no means legal advice)

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org

comms/gammu: preliminary patch to talk SDP

[Marcin Cieslak]

Hello,

Attached please find a preliminary support to make gammu discover the 
Bluetooth RFCOMM channel number using sdp(3).


The patch currently has a hard-coded search for the OBEX File Transfer 
class, as I have some trouble understanding reasons for logic behind

bluetooth_checkservicename() function (in common/device/devfunc.c).

This code correctly parses SDP response and extracts the RFCOMM channel
for the desired protocol. I have basically copied the code from
sdpcontrol(8) to do this.

I am unable to test this further as I cannot figure out the correct way
to talk to my phone (My Nokia refuses to talk on the OBEX File
Transfer RFCOMM channel #10 despite being advertised via SDP), but this
code delivers the expected channel number to the Bluetooth connect(2) call.

Probably it can be made usable for most of the people in a short time
therefore I thought somebody will be able to pick it from there.

Btw, the maintainer's email address (b...@...) bounces.

--Marcin

=== Generating patch
=== Viewing diff with more
diff -ruN --exclude=CVS /usr/ports/comms/gammu/files/patch-CMakeLists.txt /usr/home/saper/FreeBSD/ports/gammu/files/patch-CMakeLists.txt
--- /usr/ports/comms/gammu/files/patch-CMakeLists.txt	2007-12-16 21:59:51.0 +0100
+++ /usr/home/saper/FreeBSD/ports/gammu/files/patch-CMakeLists.txt	2009-01-13 05:00:37.883255000 +0100
@@ -1,6 +1,32 @@
 CMakeLists.txt.orig	2007-08-08 16:30:55.0 +0400
-+++ CMakeLists.txt	2007-08-21 00:18:05.0 +0400
-@@ -601,7 +604,7 @@
+--- CMakeLists.txt.orig	2008-12-19 10:31:54.0 +0100
 CMakeLists.txt	2009-01-13 05:00:34.875317481 +0100
+@@ -204,7 +204,11 @@
+ find_package (FBSDBluetooth)
+ if (FBSD_BLUE_FOUND)
+ set(BLUETOOTH_FOUND ON)
+-set(BLUETOOTH_SEARCH FALSE)
++if (FBSD_SDP_FOUND)
++	set(BLUETOOTH_SEARCH TRUE)
++else (FBSD_SDP_FOUND)
++	set(BLUETOOTH_SEARCH FALSE)
++endif (FBSD_SDP_FOUND)
+ endif (FBSD_BLUE_FOUND)
+ find_package (OSXBluetooth)
+ if (OSX_BLUE_FOUND)
+@@ -282,6 +286,12 @@
+ endif (NOT ${FBSD_BLUE_LIBRARIES} STREQUAL )
+ endif (FBSD_BLUE_FOUND)
+ 
++if (FBSD_SDP_FOUND)
++if (NOT ${FBSD_SDP_LIBRARIES} STREQUAL )
++set (GAMMU_LIBS ${GAMMU_LIBS} -l${FBSD_SDP_LIBRARIES})
++endif (NOT ${FBSD_SDP_LIBRARIES} STREQUAL )
++endif (FBSD_SDP_FOUND)
++
+ if (ICONV_FOUND)
+ if (NOT ${ICONV_LIBRARIES} STREQUAL )
+ set (GAMMU_LIBS ${GAMMU_LIBS} -l${ICONV_LIBRARIES})
+@@ -721,7 +731,7 @@
  set (INSTALL_DOC_DIR share/doc/gammu CACHE STRING Path for documentation installation)
  mark_as_advanced (INSTALL_DOC_DIR)
  
diff -ruN --exclude=CVS /usr/ports/comms/gammu/files/patch-FindFBSDBluetooth.cmake /usr/home/saper/FreeBSD/ports/gammu/files/patch-FindFBSDBluetooth.cmake
--- /usr/ports/comms/gammu/files/patch-FindFBSDBluetooth.cmake	1970-01-01 01:00:00.0 +0100
+++ /usr/home/saper/FreeBSD/ports/gammu/files/patch-FindFBSDBluetooth.cmake	2009-01-13 05:04:56.330885000 +0100
@@ -0,0 +1,47 @@
+--- cmake/FindFBSDBluetooth.cmake.orig	2007-09-03 04:06:48.0 +0200
 cmake/FindFBSDBluetooth.cmake	2009-01-13 05:04:44.992120772 +0100
+@@ -10,7 +10,7 @@
+ 
+ if (NOT DEFINED FBSD_BLUE_FOUND)
+ if (NOT CROSS_MINGW)
+-find_path(FBSD_BLUE_INCLUDE_DIR NAMES bluetooth.h
++find_path(FBSD_BLUE_INCLUDE_DIR NAMES bluetooth.h sdp.h
+PATHS
+/usr/include
+/usr/local/include
+@@ -22,14 +22,33 @@
+/usr/local/lib
+ )
+ 
++find_path(FBSD_SDP_INCLUDE_DIR NAMES sdp.h
++   PATHS
++   /usr/include
++   /usr/local/include
++)
++
++find_library(FBSD_SDP_LIBRARIES NAMES sdp
++   PATHS
++   /usr/lib
++   /usr/local/lib
++)
++
+ if(FBSD_BLUE_INCLUDE_DIR AND FBSD_BLUE_LIBRARIES)
+set(FBSD_BLUE_FOUND TRUE CACHE INTERNAL FreeBSD Bluetooth found)
+message(STATUS Found FreeBSD Bluetooth: ${FBSD_BLUE_INCLUDE_DIR}, ${FBSD_BLUE_LIBRARIES})
++   if(FBSD_SDP_INCLUDE_DIR AND FBSD_SDP_LIBRARIES)
++   set(FBSD_SDP_FOUND TRUE CACHE INTERNAL FreeBSD SDP found)
++   message(STATUS Found FreeBSD SDP: ${FBSD_SDP_INCLUDE_DIR}, ${FBSD_SDP_LIBRARIES})
++   else(FBSD_SDP_INCLUDE_DIR AND FBSD_SDP_LIBRARIES)
++   set(FBSD_SDP_FOUND FALSE CACHE INTERNAL FreeBSD SDP not found)
++   message(STATUS FreeBSD SDP not found.)
++   endif(FBSD_SDP_INCLUDE_DIR AND FBSD_SDP_LIBRARIES)
+ else(FBSD_BLUE_INCLUDE_DIR AND FBSD_BLUE_LIBRARIES)
+-   set(FBSD_BLUE_FOUND FALSE CACHE INTERNAL FreeBSD Bluetooth found)
++   set(FBSD_BLUE_FOUND FALSE CACHE INTERNAL FreeBSD Bluetooth not found)
+message(STATUS FreeBSD Bluetooth not found.)
+ endif(FBSD_BLUE_INCLUDE_DIR AND FBSD_BLUE_LIBRARIES)
+ 
+-mark_as_advanced(FBSD_BLUE_INCLUDE_DIR 

www/commonist BROKEN: No ${JAR} even if USE_JAVA=1.5+ defined?

[Marcin Cieslak]

Hello,

One of my ports breaks badly on the tinderbox:

portname:   www/commonist
broken because: does not extract
build errors:
http://pointyhat.freebsd.org/errorlogs/i386-errorlogs/e.7.20080809191637/commonist-0.3.28.log 
(Sun Aug 10 00:58:14 UTC 2008)
overview: 
http://portsmon.FreeBSD.org/portoverview.py?category=wwwportname=commonist


The error message is:

===   commonist-0.3.28 depends on executable: unzip - found
(cd /work/a/ports/www/commonist/work/commonist-0.3.28  
/usr/local/diablo-jdk1.5.0/bin/jar xf lib/lib-util-src.jar src)

/usr/local/diablo-jdk1.5.0/bin/jar: not found

The port defines

USE_JAVA=1.5+

and uses ${JAR} to get the proper location of the tool.

What's wrong?



signature.asc
Description: OpenPGP digital signature

Port review request: ports/124771

[Marcin Cieslak]

Hello,

I have prepared a new port, www/commonist
http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/124771

(depends on www/jericho-html, PR ports/124770).

I had to write some tricky scripts to fix rather strange packaging of
both ports.

Since these are one of my first Java ports I would appreciate some
review from this list.

Additional question: what's the preferred way of adding JARs as
dependencies? Should they be added both to BUILD_DEPENDS and RUN_DEPENDS?

Thanks in advance,

--Marcin






signature.asc
Description: OpenPGP digital signature

Re: [SOLVED] OpenOffice-2.3 : Error building under 7.0-BETA2

[Marcin Cieslak]

Norberto Meijome wrote:

 [EMAIL PROTECTED] [Mon Nov  5 08:09:34 2007]
 /usr/ports/editors/openoffice.org-2
 $ sudo make -DWITH_DEBUG=2 -DWITH_SYSTEM_FREETYPE -DWITH_CUPS 
 -DWITHOUT_MOZILLA package

 []

 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/odk/pack/gendocu
 dmake: Executing shell macro: $(FIND) $(INCLUDETOPDIRLIST) -type d -print
 dmake: Executing shell macro: $(FIND) $(INCLUDETOPDIRLIST) -type f ! $(QOB) 
 -name *~ -o -name *build.lst -o -name *deliver.log $(QCB) -print
 touch ../../unxfbsdi.pro/misc/cpp_docu_cleanup.flag
 mkdir -p ../../unxfbsdi.pro/bin/odkcommon/docs/cpp/ref/
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/bin/autodoc
  -html ../../unxfbsdi.pro/bin/odkcommon/docs/cpp/ref -name UDK 3.2.7 C/C++ 
 API Reference -lg c++ -p sal 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t sal -t osl -t rtl -p store 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t store -p registry 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t registry -p cppu 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t cppu -t com -t typelib -t uno -p cppuhelper 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t cppuhelper -p salhelper /usr/ports/editors
  /openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc -t salhelper
  -p bridges 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/solver/680/unxfbsdi.pro/inc
  -t bridges
 Autodoc version 2.2.5
 -

 Parsing the repository UDK 3.2.7 C/C++ API Reference ...
 ..80 files found to parse in project sal.
 assertion failed: i  0 ? i_pTokTypeArray[i]  i_pTokTypeArray[i-1] : true 
 in file: ../inc/semantic/callf.hxx at line: 186
 dmake:  Error code 3, while making 
 '../../unxfbsdi.pro/bin/odkcommon/docs/cpp/ref/index.html'
 ---* RULES.MK *---

 ERROR: Error 65280 occurred while making 
 /usr/ports/editors/openoffice.org-2/work/OOG680_m6/odk/pack/gendocu
 dmake:  Error code 1, while making 'build_instsetoo_native'

For the archives:

This is OpenOffice.org bug #i82101#

http://www.openoffice.org/issues/show_bug.cgi?id=82101

Fixed in the 2.4 release (now in ports).

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: problem about software traffic

[Marcin Cieslak]

[EMAIL PROTECTED] wrote:

 Do you have up-to-date ports, and is that directory clean?
   
 not sure

Something is very broken in your system. Do you have enough disk space?

Did you upgrade the system recently? How did you do that?

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: after X.org upgrade, pcidata module is missing

[Marcin Cieslak]

Scott Bennett wrote:
  Can someone tell me where to find the font that X is looking for?
 And then, how and where do I install the font?  /usr/local/lib/X11/fonts
 contains no files with names containing the string fixed, but I know so
 little about the internals of X that I don't know whether the reference
 to a default font called fixed has anything to do with the file name(s)
 that would contain the font in question.

Do you have anything in those /usr/local/lib/X11/fonts directories?
If not, installing some ports from x11-fonts/xorg-fonts* family would
certainly help.

I suppose that your upgrade process didn't go so smooth and you need to
check which X components got installed.

I would verify how far you are with

x11/xorg-libraries
x11-fonts/xorg-fonts
x11/xorg-protos

Checking x11/xorg-apps and x11/xorg-docs may save you some frustration, too.

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: Preparing an Oracle Database XE port/package -- any tips ?

[Marcin Cieslak]

Adrian Penisoara wrote:
 Hi,

(dropping -isp)

 During election observation mission in 2007 an Oracle XE installation
 I've had (just for test) on my FreeBSD notebook saved my (election) day
 and allowed me to analyze and verify election results - and this is a
 Big Thing(tm).
 
 
 Well, this sort of stories really ought to be told and heard around. Have
 you considered writing a short story or even a news article on this topic (I
 guess there are good places where you could submit it) ?

This is not always easy. You need to observe state secrets of a host
country plus some special requirements in a highly political
environment. I will try my best anyway.

 Right. I will be extracting the RPM specs file to form the package
 installation script.
 I'm not sure whether the debian scripts have something better.

Go for install.sh on the
http://wiki.bsdportal.ru/doc:oraclexe_on_freebsd instead. Worked for me
like a charm. (The kernel patches mentioned in the article are already
incorporated).

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: java/eclipse java/jdk16 IPV6 problems

[Marcin Cieslak]

Dominic Fandrey wrote:
 I'm running eclipse with jdk16. If I build the jdk with IPV6 support the
 Eclipse JUnit-Plugin doesn't work and the Debugger complains about not
 being able to connect to a socket. I forgot to copy the exact message,
 before I rebuilt without IPV6. It works fine without it. I can switch
 back to IPV6 and get the exact message if someone thinks that might be
 useful.

Try

sysctl net.inet6.ip6.v6only=0

for now.

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: Thunderbird, enigmail, and GCC 3.4

[Marcin Cieslak]

Coleman Kane wrote:
 How does seamonkey compare to Thunderbird? Is it really bloated?

I switched to seamonkey from the Mozilla suite. I never really liked
Firefox interface and I prefer tighter integration between browser and
mail on UNIX-like systems.

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: Thunderbird, enigmail, and GCC 3.4

[Marcin Cieslak]

Yuri Pankov wrote:
 AFAICT, it only affects amd64. Maybe USE_GCC can be 3.4 on amd64, and 3.4+ 
 elsewhere?

Don't know if it helps at all, but seamonkey works with enigmail very
nice, using gcc version 4.2.1 20070719 on amd64 (7.0-PRERELEASE,
seamonkey 1.1.8).

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: FreeBSD Port: java/gnu-regexp

[Marcin Cieslak]

[EMAIL PROTECTED] wrote in German:

 Ich brauche information darüber, welche license bedingungen bei der
 verwendung von gnu-regexp berücksichtigt werden müssen. Da gnu-regexp, so
 weit ich verstanden habe, ein freeware ist!
 Danke im voraus!

Laut https://savannah.gnu.org/projects/gnu-regexp/ es ist nach LGPL
lizenziert. Bitte beachten, dass FreeBSD-ports Mailingliste
Englischsprachig ist!

(The question was about gnu-regexp licence).

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Apache Commons ports: unfetchable distfiles

[Marcin Cieslak]

Submitter-Id:  current-users
Originator:Marcin Cieslak
Organization:  
Confidential:  no 
Synopsis:  Apache Commons ports: unfetchable distfiles
Severity:  serious
Priority:  high
Category:  ports
Class: update
Release:   FreeBSD 7.0-PRERELEASE amd64
Environment:
System: FreeBSD radziecki.saper.info 7.0-PRERELEASE FreeBSD 7.0-PRERELEASE #9: 
Sat Jan 26 01:36:13 CET 2008 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/VAIO amd64

Apache Commons project are no longer hosted under MASTER_SITES_APACHE 
jakarta/commons subdirectory, they have been moved one directory up.

Please see http://commons.apache.org/charter.html

databases/jakarta-commons-dbutils
devel/jakarta-commons-chain
devel/jakarta-commons-configuration
devel/jakarta-commons-daemon
devel/jakarta-commons-io
devel/jakarta-commons-jxpath
devel/jakarta-commons-modeler
java/jakarta-commons-beanutils
java/jakarta-commons-cli
java/jakarta-commons-codec
java/jakarta-commons-collections
java/jakarta-commons-dbcp
java/jakarta-commons-discovery
java/jakarta-commons-httpclient
java/jakarta-commons-lang
java/jakarta-commons-logging
java/jakarta-commons-pool
java/jakarta-commons-primitives
lang/jakarta-commons-jelly
math/jakarta-commons-math
net/jakarta-commons-net
textproc/jakarta-commons-digester

Description:

How-To-Repeat:

Fix:

http://akson.sgh.waw.pl/~saper/FreeBSD/ports/commons/commons-distfiles.diff

Index: Mk/bsd.sites.mk
===
RCS file: /usr/home/ncvs/ports/Mk/bsd.sites.mk,v
retrieving revision 1.436
diff -u -r1.436 bsd.sites.mk
--- Mk/bsd.sites.mk 7 Feb 2008 16:13:41 -   1.436
+++ Mk/bsd.sites.mk 24 Feb 2008 04:53:53 -
@@ -83,6 +83,16 @@
ftp://ftp.saix.net/pub/apache/dist/%SUBDIR%/
 .endif
 
+.if !defined(IGNORE_MASTER_SITE_APACHE_COMMONS_BINARIES)
+MASTER_SITE_APACHE_COMMONS_BINARIES+=  \
+   ${MASTER_SITE_APACHE:S,%SUBDIR%,commons//binaries,}
+.endif
+
+.if !defined(IGNORE_MASTER_SITE_APACHE_COMMONS_SOURCE)
+MASTER_SITE_APACHE_COMMONS_SOURCE+=\
+   ${MASTER_SITE_APACHE:S,%SUBDIR%,commons//source,}
+.endif
+
 .if !defined(IGNORE_MASTER_SITE_APACHE_HTTPD)
 MASTER_SITE_APACHE_HTTPD+= \
${MASTER_SITE_APACHE:S,%SUBDIR%,httpd/,}
Index: /usr/ports/databases/jakarta-commons-dbutils/Makefile
===
RCS file: /usr/home/ncvs/ports/databases/jakarta-commons-dbutils/Makefile,v
retrieving revision 1.5
diff -u -r1.5 Makefile
--- /usr/ports/databases/jakarta-commons-dbutils/Makefile   19 May 2007 
20:01:05 -  1.5
+++ /usr/ports/databases/jakarta-commons-dbutils/Makefile   24 Feb 2008 
04:38:52 -
@@ -9,8 +9,8 @@
 PORTVERSION=   1.1
 PORTREVISION=  1
 CATEGORIES=databases java
-MASTER_SITES=  ${MASTER_SITE_APACHE_JAKARTA}
-MASTER_SITE_SUBDIR=commons/dbutils/source
+MASTER_SITES=  ${MASTER_SITE_APACHE_COMMONS_SOURCE}
+MASTER_SITE_SUBDIR=${PORTNAME:S,commons-,,}
 PKGNAMEPREFIX= jakarta-
 DISTNAME=  ${PORTNAME}-${PORTVERSION}-src
 
Index: /usr/ports/devel/jakarta-commons-chain/Makefile
===
RCS file: /usr/home/ncvs/ports/devel/jakarta-commons-chain/Makefile,v
retrieving revision 1.3
diff -u -r1.3 Makefile
--- /usr/ports/devel/jakarta-commons-chain/Makefile 19 May 2007 20:02:54 
-  1.3
+++ /usr/ports/devel/jakarta-commons-chain/Makefile 24 Feb 2008 04:38:52 
-
@@ -9,8 +9,8 @@
 PORTVERSION=   1.1
 PORTREVISION=  1
 CATEGORIES=devel java
-MASTER_SITES=  ${MASTER_SITE_APACHE_JAKARTA}
-MASTER_SITE_SUBDIR=${PORTNAME:S,-,/,}/source
+MASTER_SITES=  ${MASTER_SITE_APACHE_COMMONS_SOURCE}
+MASTER_SITE_SUBDIR=${PORTNAME:S,commons-,,}
 PKGNAMEPREFIX= jakarta-
 DISTNAME=  ${PORTNAME}-${PORTVERSION}-src
 
Index: /usr/ports/devel/jakarta-commons-daemon/Makefile
===
RCS file: /usr/home/ncvs/ports/devel/jakarta-commons-daemon/Makefile,v
retrieving revision 1.2
diff -u -r1.2 Makefile
--- /usr/ports/devel/jakarta-commons-daemon/Makefile29 Jul 2007 01:15:26 
-  1.2
+++ /usr/ports/devel/jakarta-commons-daemon/Makefile24 Feb 2008 04:38:52 
-
@@ -8,8 +8,8 @@
 PORTNAME=  commons-daemon
 PORTVERSION=   1.0.1
 CATEGORIES=devel java
-MASTER_SITES=  ${MASTER_SITE_APACHE_JAKARTA}
-MASTER_SITE_SUBDIR=commons/daemon/source
+MASTER_SITES=  ${MASTER_SITE_APACHE_COMMONS_SOURCE}
+MASTER_SITE_SUBDIR=${PORTNAME:S,commons-,,}
 PKGNAMEPREFIX= jakarta-
 DISTNAME=  daemon-${PORTVERSION}
 
Index: /usr/ports/devel/jakarta-commons-io/Makefile
===
RCS file: /usr/home/ncvs/ports/devel/jakarta-commons-io/Makefile,v
retrieving revision 1.8
diff -u -r1.8 Makefile
--- /usr/ports/devel/jakarta-commons-io/Makefile14 Dec 2007 06:24:47 
-  1.8
+++ /usr/ports/devel/jakarta-commons-io/Makefile24 Feb 2008

Re: net-im/psi update plans?

[Marcin Cieslak]

Shaun Amott wrote:
 On Fri, Feb 22, 2008 at 11:57:06PM -0600, Kelly Hays wrote:
 I was wondering what plans were in the works for net-im/psi version
 0.11 and/or development snapshots of 0.12?

 
 I'll be updating the port to 0.12 shortly. I have some additional
 testing to do. I currently have no plans to do do anything with the
 development snapshots, though.

It would be nice if the FreeBSD port could include 0.10 default icon set
available at http://jabber.limanowa.net/psi-iconsets/stellar2.jisp

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

libgtkmozembed.so for Adobe Reader 8

[Marcin Cieslak]

Hello,

New Adobe Reader 8 requires some HTML rendering capabilities mainly for 
his new advertising splash-screen.


Therefore it needs that libgtkmozembed.so is present and set in the 
preferences (without this you get annoying popup at startup but PDF 
functionality works fine).


libgtkmozembed.so is in every linux-* browser port, I use one from 
www/linux-seamonkey. However I think we should not add any of the 
linux-* ports as a dependency.


How this should be solved? A separate port (but still pulling large 
distfile)? A wrapper? Any other ideas?


--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: testing skype

[Marcin Cieslak]

Chuck Robey wrote:
Anybody who has skype up and working, and has a few free minutes, would 
you care to help me test my brand new skype setup?  Write me privately, ok?


There is a test robot you can call and talk to - it's called echo123.
(Sorry don't have skype here right now).

--Marcin

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]

Re: openoffice.org-2.0 does not build on amd64 with 6.1

[Marcin Cieslak]

Vasil Dimov wrote:
 On Tue, Oct 31, 2006 at 12:06:55AM +, Bartosz Fabianowski wrote:
 It seems that you have run into a know problem with the OpenOffice.org 
 build system - it does not work when file names become too long. In the 
 error message you posted, 

I've run into similar problem with epm and 256 character limit (tried
last time with 2.02). idlc seems something new to me.
-- 
   Marcin Cieslak // [EMAIL PROTECTED] 
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to [EMAIL PROTECTED]