Re: [gentoo-user] Userspace CPU frequency scaling not available on one machine

[Ramon Fischer]

You can disable the intel p state driver by adding an additional kernel 
parameter through your bootloader.

For GRUB2 you can do the following:

/etc/default/grub:
GRUB_CMDLINE_LINUX="intel_pstate=disable"

terminal:
sudo grub-mkconfig -o /boot/grub/grub.cfg
sudo reboot

After that, the other governors should be available, too.

-Ramon

On 11/01/2017 04:12 AM, Walter Dnes wrote:

  I've cobbled together a script that runs under eselect to list/set
COU speeds/governors on my machines.  I set max speed when I'm using
the machine, and min speed when not.  It has run fine so far, except
recently on an Intel Silvermont machine (64-bit Atom family) running
Gentoo kernel 4.12.12.  Userspace CPU control is not available, even
though I've asked for it under "make menuconfig".  On other machines
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
shows...

conservative userspace ondemand performance

...but on this one machine it only shows...

performance powersave

...so I know it's the machine, not my script, that's the problem.  Note
that I set conservative, performance, powersave, and userspace in the
"make menuconfig" process, and made userspace the default governor.  One
difference I noticed on the Silvermont is CONFIG_X86_INTEL_PSTATE=y
which is the opposite of the machines that things work properly on.  So
I went into "make menuconfig" to try to switch it off, but I get...

-*-   Intel P state control

  Yes, it is *FORCED* on on that one machine.  Now what?  I'm attaching
/proc/config.gz from the Silvermont in case it helps.  Feel free to
comment on other stuff in config.gz as well.

Re: [gentoo-user] Compilation error with StructureSynth

[David Haller]

Hello,

On Wed, 01 Nov 2017, tu...@posteo.de wrote:
>I am trying to compile this wonderful beast:
>* media-gfx/structure-synth
> Available versions:  (~)1.5.0
> Homepage:http://structuresynth.sourceforge.net/
> Description: A program to generate 3D structures by specifying a 
> design grammar
>
>
>and it gives me this:
>
>x86_64-pc-linux-gnu-g++ -c -march=native -O2 -pipe -Wall -W -D_REENTRANT 
>-DQT_NO_DEBUG -DQT_SCRIPT_LIB -DQT_XML_LIB -DQT_OPENGL_LIB -DQT_GUI_LIB 
>-DQT_CORE_LIB -DQT_SHARED -I/usr/share/qt4/mkspecs/linux-g++ -I. 
>-I/usr/include/qt4/QtCore -I/usr/include/qt4/QtGui -I/usr/include/qt4/QtOpenGL 
>-I/usr/include/qt4/QtXml -I/usr/include/qt4/QtScript -I/usr/include/qt4 -I. 
>-IStructureSynth/GUI -ISyntopiaCore/GLEngine -ISyntopiaCore/Math 
>-ISyntopiaCore/Misc -IStructureSynth/Parser -ISyntopiaCore/Exceptions 
>-IStructureSynth/Model/Rendering -IStructureSynth/Model 
>-IThirdPartyCode/MersenneTwister -ISyntopiaCore/Logging 
>-IStructureSynth/JavaScriptSupport -ISyntopiaCore/GLEngine/Raytracer -I. -o 
>JavaScriptParser.o StructureSynth/JavaScriptSupport/JavaScriptParser.cpp
>In file included from StructureSynth/JavaScriptSupport/Debug.cpp:16:0:
>StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Sphere.h:25:4: 
>error: 'GLUquadric' does not name a type
>GLUquadric* myQuad;
>^

Copy the media-gfx/structure-synth subdir from the portage-tree to
your local overlay and add the attached files as suggested by the
filenames (create the files/ subdir in advance and put the patch there).

HTH,
-dnh

-- 
WANTED: Schroedingers Cat, dead or alive.# Copyright 1999-2013 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2

EAPI=4

inherit eutils qt4-r2

MY_P="StructureSynth-Source-v${PV}"
DESCRIPTION="A program to generate 3D structures by specifying a design grammar"
HOMEPAGE="http://structuresynth.sourceforge.net/;
SRC_URI="mirror://sourceforge/structuresynth/${MY_P}.zip"

LICENSE="|| ( LGPL-2.1 GPL-3 )"
SLOT="0"
KEYWORDS="~amd64 ~x86"
IUSE=""

RDEPEND="virtual/opengl
dev-qt/qtcore:4
dev-qt/qtgui:4
dev-qt/qtopengl:4
dev-qt/qtscript:4"
DEPEND="${RDEPEND}
app-arch/unzip"

S="${WORKDIR}/Structure Synth Source Code"

src_prepare() {
epatch "${FILESDIR}/structure-synth-1.5.0-gl.patch"

qmake -project -o ${PN}.pro -after "CONFIG+=opengl" \
-after "QT+=xml opengl script" \
QMAKE_LIBS_OPENGL="-lGLU -lGL" || die "qmake failed"
}

src_install() {
dobin ${PN}
dodoc roadmap.txt changelog.txt bugs.txt
domenu ${PN}.desktop
newicon images/structuresynth.png ${PN}.png
}
diff -x '*~' -purN a/SyntopiaCore/GLEngine/Raytracer/Sampler.h b/SyntopiaCore/GLEngine/Raytracer/Sampler.h
--- a/SyntopiaCore/GLEngine/Raytracer/Sampler.h	2010-11-04 20:02:46.0 +0100
+++ b/SyntopiaCore/GLEngine/Raytracer/Sampler.h	2017-11-01 04:42:46.320970684 +0100
@@ -1,5 +1,6 @@
 #pragma once
 
+#include 
 #include "SyntopiaCore/Math/Random.h"
 #include "SyntopiaCore/Math/Vector3.h"
 
diff -x '*~' -purN a/SyntopiaCore/GLEngine/Raytracer/VoxelStepper.cpp b/SyntopiaCore/GLEngine/Raytracer/VoxelStepper.cpp
--- a/SyntopiaCore/GLEngine/Raytracer/VoxelStepper.cpp	2010-09-08 21:25:30.0 +0200
+++ b/SyntopiaCore/GLEngine/Raytracer/VoxelStepper.cpp	2017-11-01 04:46:51.052969311 +0100
@@ -122,7 +122,7 @@ namespace SyntopiaCore {
 		currentT = p;
 
 		// We do not intersect grid.
-		if (!found) return false;
+		if (!found) return 0;
 }
 
 stepX = (dir.x() > 0) ? 1 : -1;
diff -x '*~' -purN a/SyntopiaCore/GLEngine/Sphere.h b/SyntopiaCore/GLEngine/Sphere.h
--- a/SyntopiaCore/GLEngine/Sphere.h	2010-08-11 15:12:22.0 +0200
+++ b/SyntopiaCore/GLEngine/Sphere.h	2017-11-01 04:42:55.864970632 +0100
@@ -1,5 +1,6 @@
 #pragma once
 
+#include 
 #include "SyntopiaCore/Math/Vector3.h"
 #include "Object3D.h"
 

[gentoo-user] Userspace CPU frequency scaling not available on one machine

[Walter Dnes]

  I've cobbled together a script that runs under eselect to list/set
COU speeds/governors on my machines.  I set max speed when I'm using
the machine, and min speed when not.  It has run fine so far, except
recently on an Intel Silvermont machine (64-bit Atom family) running
Gentoo kernel 4.12.12.  Userspace CPU control is not available, even
though I've asked for it under "make menuconfig".  On other machines
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
shows...

conservative userspace ondemand performance

...but on this one machine it only shows...

performance powersave

...so I know it's the machine, not my script, that's the problem.  Note
that I set conservative, performance, powersave, and userspace in the
"make menuconfig" process, and made userspace the default governor.  One
difference I noticed on the Silvermont is CONFIG_X86_INTEL_PSTATE=y
which is the opposite of the machines that things work properly on.  So
I went into "make menuconfig" to try to switch it off, but I get...

-*-   Intel P state control

  Yes, it is *FORCED* on on that one machine.  Now what?  I'm attaching
/proc/config.gz from the Silvermont in case it helps.  Feel free to
comment on other stuff in config.gz as well.

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications


config.gz
Description: Binary data

[gentoo-user] Compilation error with StructureSynth

[tuxic]

Hi,

I am trying to compile this wonderful beast:
* media-gfx/structure-synth
 Available versions:  (~)1.5.0
 Homepage:http://structuresynth.sourceforge.net/
 Description: A program to generate 3D structures by specifying a 
design grammar


and it gives me this:

x86_64-pc-linux-gnu-g++ -c -march=native -O2 -pipe -Wall -W -D_REENTRANT 
-DQT_NO_DEBUG -DQT_SCRIPT_LIB -DQT_XML_LIB -DQT_OPENGL_LIB -DQT_GUI_LIB 
-DQT_CORE_LIB -DQT_SHARED -I/usr/share/qt4/mkspecs/linux-g++ -I. 
-I/usr/include/qt4/QtCore -I/usr/include/qt4/QtGui -I/usr/include/qt4/QtOpenGL 
-I/usr/include/qt4/QtXml -I/usr/include/qt4/QtScript -I/usr/include/qt4 -I. 
-IStructureSynth/GUI -ISyntopiaCore/GLEngine -ISyntopiaCore/Math 
-ISyntopiaCore/Misc -IStructureSynth/Parser -ISyntopiaCore/Exceptions 
-IStructureSynth/Model/Rendering -IStructureSynth/Model 
-IThirdPartyCode/MersenneTwister -ISyntopiaCore/Logging 
-IStructureSynth/JavaScriptSupport -ISyntopiaCore/GLEngine/Raytracer -I. -o 
JavaScriptParser.o StructureSynth/JavaScriptSupport/JavaScriptParser.cpp
In file included from StructureSynth/JavaScriptSupport/Debug.cpp:16:0:
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Sphere.h:25:4: 
error: ‘GLUquadric’ does not name a type
GLUquadric* myQuad;
^
make: *** [Makefile:580: Debug.o] Error 1
make: *** Waiting for unfinished jobs
In file included from 
StructureSynth/JavaScriptSupport/JavaScriptBuilder.cpp:15:0:
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Sphere.h:25:4: 
error: ‘GLUquadric’ does not name a type
GLUquadric* myQuad;
^
In file included from 
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Raytracer/RenderThread.h:7:0,
 from 
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Raytracer/RayTracer.h:10,
 from StructureSynth/JavaScriptSupport/JavaScriptBuilder.cpp:16:
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Raytracer/Sampler.h:
 In constructor ‘SyntopiaCore::GLEngine::GaussianFilter::GaussianFilter(double, 
double)’:
StructureSynth/JavaScriptSupport/../../SyntopiaCore/GLEngine/Raytracer/Sampler.h:27:4:
 warning: ‘SyntopiaCore::GLEngine::GaussianFilter::extent’ is initialized with 
itself [-Winit-self]
GaussianFilter(double ext, double alpha) : extent(extent), alpha(alpha) {
^
In file included from 
StructureSynth/GUI/../../SyntopiaCore/GLEngine/Raytracer/RenderThread.h:7:0,
 from 
StructureSynth/GUI/../../SyntopiaCore/GLEngine/Raytracer/RayTracer.h:10,
 from StructureSynth/GUI/MainWindow.cpp:27:
StructureSynth/GUI/../../SyntopiaCore/GLEngine/Raytracer/Sampler.h: In 
constructor ‘SyntopiaCore::GLEngine::GaussianFilter::GaussianFilter(double, 
double)’:
StructureSynth/GUI/../../SyntopiaCore/GLEngine/Raytracer/Sampler.h:27:4: 
warning: ‘SyntopiaCore::GLEngine::GaussianFilter::extent’ is initialized with 
itself [-Winit-self]
GaussianFilter(double ext, double alpha) : extent(extent), alpha(alpha) {
^
make: *** [Makefile:620: JavaScriptBuilder.o] Error 1
In file included from 
StructureSynth/JavaScriptSupport/JavaScriptParser.cpp:12:0:
./SyntopiaCore/GLEngine/Sphere.h:25:4: error: ‘GLUquadric’ does not name a type
GLUquadric* myQuad;
^
make: *** [Makefile:651: JavaScriptParser.o] Error 1
StructureSynth/GUI/MainWindow.cpp: At global scope:
StructureSynth/GUI/MainWindow.cpp:1583:8: warning: ‘int 
StructureSynth::GUI::{anonymous}::match(QString, QString, double*, double*, 
double*, double*)’ defined but not used [-Wunused-function]
int match(QString s, QString m, double* a, double* b = 0, double* c = 0, 
double* d = 0) {
^
 * ERROR: media-gfx/structure-synth-1.5.0::gentoo failed (compile phase):
 *   emake failed
 * 
 * If you need support, post the output of `emerge --info 
'=media-gfx/structure-synth-1.5.0::gentoo'`,
 * the complete build log and the output of `emerge -pqv 
'=media-gfx/structure-synth-1.5.0::gentoo'`.
 * The complete build log is located at 
'/var/tmp/portage/media-gfx/structure-synth-1.5.0/temp/build.log'.
 * The ebuild environment file is located at 
'/var/tmp/portage/media-gfx/structure-synth-1.5.0/temp/environment'.
 * Working directory: 
'/var/tmp/portage/media-gfx/structure-synth-1.5.0/work/Structure Synth Source 
Code'
 * S: '/var/tmp/portage/media-gfx/structure-synth-1.5.0/work/Structure Synth 
Source Code'


I will mail the contents of the mentioned filed if wanted (I dont want
to pollute the mailinglist with dust of bytes in beforehand... :)

Is there a way around this?

Cheers
Meino

Re: [gentoo-user] insane backends

[Philip Webb]

171030 Neil Bothwick wrote:
> On Sun, 29 Oct 2017 22:39:31 -0400, Philip Webb wrote:
>> 'sane-backends-1.0.27' just stabilised, so I updated to it.
>> Previously I had selected  4  items from the list in SANE_BACKENDS,
>> which is shown by 'emerge', ie 'epson epson2 plustek plustek_pp'.
>> My scanner is an 'Epson V550 Photo'.
>> 
>> Wanting to test which of the  4  items above was needed, I tried each.
>> Inadvertently, I tried emerging 'sane-backends' with none included
>> & discovered that my scanner still works regardless !
> ISTR that with use_expanded variables like this,
> if you don't set the variable then all options are enabled,
> so you've just built an un-Gentoolike version that works everywhere.

That doesn't matter much beyond compile time, which is not excessive :
Xsane will interrogate the scanner & when it's told it's an Epson V550
it will load the appropriate driver & go from there.
>From the point of view of system maintenance,
it's much simpler to let it do that than try to pick out the driver myself.

> You could check this by using qlist
> to see what has been installed with and without SANE_BACKENDS set.

'qlist sane-backends' doesn't show anything like that.

Thanks for the explanation.

-- 
,,
SUPPORT ___//___,   Philip Webb
ELECTRIC   /] [] [] [] [] []|   Cities Centre, University of Toronto
TRANSIT`-O--O---'   purslowatchassdotutorontodotca

Re: [gentoo-user] [OT] Extracting printer settings from a gcode file?

[tuxic]

On 11/01 12:15, Adam Carter wrote:
> >
> > Would be nice, if it could be possible to extract them from the gcode
> > example
> > files in a way, that made it possible to feed them back into the slicer
> > software manually (not expecting to get a config file ready to be read
> > directly
> > with that certain slicer software I want to use...)
> >
> 
> Suggest you load the gcode file, then in slicer go to each of the three
> settings tabs and save as a new preset. Then when you go to slice your
> model, you can select those profiles. IIRC that is, i haven't touched this
> for a long time.

Hi Adam,

OH! :)
...and I thought, that the gcode is a write-only format, which cannot be 
"reversed"
into something a slicer would understand (only the printer)...

You see: It's time, that my printer finally arrived (it got stuck somehwere 
between
the UK and here) and get out of this "theoreticallu it may be. that"-thinking...

Thanks a lot Adam! Your solution is by far better than hand picking 
gcode-commands
and deciphering printer settings from that... :))

Cheers
Meino

Re: [gentoo-user] [OT] Extracting printer settings from a gcode file?

[Adam Carter]

>
> Would be nice, if it could be possible to extract them from the gcode
> example
> files in a way, that made it possible to feed them back into the slicer
> software manually (not expecting to get a config file ready to be read
> directly
> with that certain slicer software I want to use...)
>

Suggest you load the gcode file, then in slicer go to each of the three
settings tabs and save as a new preset. Then when you go to slice your
model, you can select those profiles. IIRC that is, i haven't touched this
for a long time.

Re: [gentoo-user] Alternatives to knutclient

[Peter Humphrey]

On Tue, 31 Oct 2017 17:33:53 +
Wols Lists  wrote:

> In the UK at least also, we have ring mains. These are rated at 30 Amps,
> from which you can take a 13 Amp feed from any socket. Once you start
> taking power over multiple leads the wiring has more resistance, plugs
> introduce resistance, etc, and the voltage can drop rapidly.

I've never seen that, nor heard of it. It sounds like a major fire risk too,
which the wiring regs would never countenance.

> (It is mandatory, according to code, if you hard-wire a feed off a ring
> you must separate it with a 16 Amp fuse.)

I don't recognise that either. It sounds like a special case to me. And
where do I buy a 16A fuse?

> I've never seen one, but I would have thought a motor/generator pair
> with a hefty flywheel would provide very good surge/spike/whatever
> protection, and provide nice clean power. Plus, it would be unlikely to
> burn out if you had to provide protection from several big consecutive
> shocks, like a lightning strike.

Just think of the noise though, and the expense.

-- 
Regards,
Peter.

Re: [gentoo-user] Alternatives to knutclient

[Wols Lists]

On 31/10/17 00:09, mad.scientist.at.la...@tutanota.com wrote:

> 
> the issue is with plugging one thing, into another, into another and
> then into the wall, most outlet strips are cheap, they don't use proper
> sockets and often have/develop a significant resistance, which creates a
> hazard etc.  

In the UK at least also, we have ring mains. These are rated at 30 Amps,
from which you can take a 13 Amp feed from any socket. Once you start
taking power over multiple leads the wiring has more resistance, plugs
introduce resistance, etc, and the voltage can drop rapidly.

(It is mandatory, according to code, if you hard-wire a feed off a ring
you must separate it with a 16 Amp fuse.)
> 

> 
> There are actually multistage devices that can protect low voltage lines
> (i.e. control lines on a transmitter tower) from a direct lightning
> strike assuming they have a good enough ground connection.  of course
> it's likely to destroy the surge arrestor, which isn't cheap, but it
> protects far more expensive equipment when it self sacrifices.
> 
I've never seen one, but I would have thought a motor/generator pair
with a hefty flywheel would provide very good surge/spike/whatever
protection, and provide nice clean power. Plus, it would be unlikely to
burn out if you had to provide protection from several big consecutive
shocks, like a lightning strike.

Cheers,
Wol

Re: [gentoo-user] Alternatives to knutclient

[Wols Lists]

On 30/10/17 23:32, Rich Freeman wrote:
> Often there are elements of a traditionally public service that aren't
> natural monopolies which can be outsourced for a benefit.  Electrical
> generation is often a case of that, but as I suggested you do need to
> ensure you're paying to have extra capacity online.

Exactly. The infrastructure should be a commonhold - the householders
jointly own the wires in town sort of thing. Then they can buy from any
generator, who rents the wires to deliver.

Cheers,
Wol

Re: [gentoo-user] Alternatives to knutclient

[Rich Freeman]

On Tue, Oct 31, 2017 at 10:05 AM, Wols Lists  wrote:
>  If you can, do a Buffet...
> ...
> ...demand that pension funds invest in companies that pay good dividends, 
> well covered...

Without wanting to derail this thread even more, I couldn't help but
notice the irony here...

-- 
Rich

Re: [gentoo-user] Alternatives to knutclient

[Rich Freeman]

On Tue, Oct 31, 2017 at 9:56 AM, Wols Lists  wrote:
>
> The reason British Telecom made obscene profits for YEARS after they
> privatised was because they inherited terrible infrastructure that cost
> a lot of money just to keep going.
>

A similar kind of system is the reason that Bell Labs made so many
inventions back in the day.  The phone company worked on a cost-plus
model, so the more they spent on anything remotely telecom-related,
the more money they made.  So they just did everything they can to
increase their costs.

At least something good came out of it in that case.  Modern
businesses have gotten better at ensuring the graft goes to the
executives and owners without the risk of too much public benefit
trickling down, heaven forbid...

-- 
Rich

Re: [gentoo-user] Alternatives to knutclient

[Wols Lists]

On 30/10/17 23:42, Rich Freeman wrote:
> If the big banks thought that investing for the long term would make
> them more money they would do it.  They have no loyalty to the
> companies they invest in.  If they can invest in a company one month,
> and make more money by investing in a competitor that will put them
> out of business the next month, they will.  I'm not sure why a "long
> term investor" wouldn't do the same if they could make money doing it.
> They have money for the long-term, but that doesn't mean that they
> have to keep it in once place.

The banks have nothing to do with it.

The fund managers are rewarded for beating the index. They are NOT
punished for missing the index. That means a fund manager who is bang on
the average will get a bonus every second year. Paid for with investors'
money.

The quickest way for day traders to make a profit is to predict what the
big boys are going to do (easy if you understand that most of them are
index trackers who have little choice where to invest) and beat them to
it. Buy shares that you have good reason are going to join the index
next week, and dump them as soon as they do. Sell shares short that you
think are going to be dumped from the index, and buy them back
afterwards. Easy money!

The thing about investing is to remember that the big boys are simply
gambling with your money. If you can, do a Buffet, buy stocks you
understand, and hang on to them. They'll go up.

Imho the quickest way to stabilise the market and kill a lot of these
shenanigans would be to demand that pension funds invest in companies
that pay good dividends, well covered. In other words, if a company
earns 50p/share and pays a 10p dividend, then the dividend "is covered 5
times". That means the pensions can't invest in speculative, highly
geared companies. Which means those companies ARE going to invest in
themselves, and will be good, solid companies unlikely to go
spectacularly bust.

Cheers,
Wol

Re: [gentoo-user] Alternatives to knutclient

[Wols Lists]

On 30/10/17 23:28, Mick wrote:
> The regulator does not hold the budget, central government departments do and 
> the regulator cannot (or will not) control abnormal profits privatised 
> utilities are making year after year.  However, the regulator will engage 
> enthusiastically in the a theatre of regular reviews of market conditions, in 
> an attempt to convince consumers and tax payers the most gratuitous abuse of 
> power is kept in check.

Quite often that's the government's fault too ...

The reason British Telecom made obscene profits for YEARS after they
privatised was because they inherited terrible infrastructure that cost
a lot of money just to keep going.

And the reason it cost so much was that every time the General Post
Office tried to put money aside to replace said infrastructure, the
Government (the sole shareholder) declared a large dividend and took it
away. The alternative was for the GPO to borrow, but every time they did
that the Treasury said "that money is part of Government Borrowing.
We're committed to reducing government borrowing so you're not allowed
to borrow".

So BT made gazillions because a brand new exchange, with a ten year
warranty, probably cost about two years' maintenance of the exchange it
replaced. And because BT were quite visibly cutting prices (not by
much!) but the new infrastructure was reducing costs so much faster,
their profits soared.

Cheers,
Wol