[gentoo-user] Re: [WAY OT] wanna learn networking internals

[James]

Francisco Ares  gmail.com> writes:


> Sorry for such WAY out of topic message, but Gentoo users are also way 
> out of regular computer users.

Far out man. some psychedelic mood music, say Pink Floyd Ummagumma::
Several Species of Small Furry Animals Gathered Together in a Cave...[2]
(now that we got the proper 'grove on'.let's roll!


> I intend to learn more deep details about networking intrinsics, 
> (packets, ports, negotiation, UDP, multicast, unicast, TCP, ethernet, 
> DHCP, protocols, and so on) so I decided to recur to this list.  
> Francisco


Well, if you really want to 'Get Smart'  as in get to 'the edge' of all of
this, why not just jump straight to (SDN) Software Defined Networks and such
things as OpenVswitch on a cluster? [1]. Our good friends at Rackspace are
very progressive in their ideas and systems development. As you learn about
SDN, you can backfill what the old farts have done in the past and be
an active part of the future? Just search out SDN as it's a revolution in
networking that just may be the disruptive thang that is so desperately needed.


PS::I bet McKinnon can build a buffered serial port sniffer out of those old
74LSxx chips! I have a wide selection of those old 74LS chips in
tubes, just in case you get an itch you need to scratch...
The date codes have fadedall I can make out is you old fart.



cheers,
James


[1] https://www.pytexas.org/2015/talk/45

[2] https://www.youtube.com/watch?v=YrRhnaFaBsA

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Alan McKinnon]

On 03/09/2015 16:31, James wrote:
> PS::I bet McKinnon can build a buffered serial port sniffer out of those old
> 74LSxx chips! I have a wide selection of those old 74LS chips in
> tubes, just in case you get an itch you need to scratch...
> The date codes have fadedall I can make out is you old fart.


McKinnon says "I'll have to get back to you on that!" while wandering
into his library to fetch the trusty old copy of Mastering Serial
Communications copyright 1986




-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

On 09/03/2015 07:19 AM, Alan McKinnon wrote:
> On 03/09/2015 15:06, Alec Ten Harmsel wrote:
>> On Wed, Sep 02, 2015 at 02:25:47PM -0600, the...@sys-concept.com wrote:
>>> I've tried to post a "log" file to the http://pastebin.com/ you
>>> suggested but I can only paste the limited size file (not upload it).
>>> Since the txt file is 7.4Mb in size, I can not paste it.
>>> Though, I have compress the file as tar.gz (so it is only 267kB)
>>> here is the link:
>>>
>>> http://www.sysconcept.ca/audacity_error.tar.gz

I've changed the permission on this file to: apache:apache and rw
the link above should work.

>>>
>>> If somebody whats to look at it, I appreciate it.
>>> Meanwhile, I'll be following other folks suggestion and see if I can get
>>> a positive result to this error.
>>>
>>> Thelma
>>
>> I tried to download it but got a 403 Forbidden HTTP error.
>>
>> This might be a bad suggestion, but if you have a lot of time, it might
>> be easier to uninstall audacity and any other packages that have been
>> giving you problems, then `emerge -uDN --with-bdeps=y @world`, then
>> `emerge --depclean', and then install all the stuff you need.
>>
>> I believe someone else mentioned checking /var/lib/portage/world and
>> making sure that it doesn't contain any virtuals - it might also be good
>> to remove from it any software that you do not directly need.
> 
> 
> Even better - Thelma should just post the entire /var/lib/portage/world
> file so we can advise what to take out.
> 
> Most newbies clutter up their world needlessly, it takes a bit of
> practice to grok what should be in their

Yes, the system is few years old.
It make me wonder if the problem might be cause by me emerging:

emerge -avC libjpeg-turbo
emerge -av1 media-libs/jpeg:0 media-libs/jpeg:62

I'm using obsolete "nxclient-3.5.0.7" as I need it and there is no good
replacement alternative on Gentoo.

Here is /var/lib/portage/world

app-admin/gkrellm
app-admin/syslog-ng
app-admin/tmpwatch
app-admin/webapp-config
app-arch/cabextract
app-arch/xarchiver
app-benchmarks/cpuburn
app-cdr/cdrtools
app-cdr/dvd+rw-tools
app-crypt/gnupg
app-crypt/pinentry
app-dicts/myspell-en
app-editors/leafpad
app-editors/nano
app-emulation/dosemu
app-emulation/virtualbox-bin
app-emulation/virtualbox-modules
app-forensics/chkrootkit
app-forensics/rkhunter
app-misc/ca-certificates
app-office/glabels
app-office/gnucash
app-office/gnumeric
app-office/openoffice-bin
app-portage/cfg-update
app-portage/eix
app-portage/genlop
app-portage/gentoolkit
app-portage/portage-utils
app-text/a2ps
app-text/acroread
app-text/dos2unix
app-text/enscript
app-text/evince
app-text/flpsed
app-text/ghostscript-gpl
app-text/gv
app-text/lcdf-typetools
app-text/mpage
app-text/pdfjam
app-text/pdfshuffler
app-text/pdftk
app-text/texi2html
app-text/wdiff
dev-db/mysql
dev-db/phpmyadmin
dev-db/postgresql
dev-db/postgresql:9.0
dev-db/postgresql:9.1
dev-java/icedtea-bin
dev-java/java-config
dev-java/oracle-jdk-bin
dev-lang/lua
dev-lang/php
dev-lang/python
dev-lang/swig
dev-lang/tcl
dev-lang/tk
dev-libs/check
dev-libs/libnl
dev-libs/openssl
dev-perl/DBD-Pg
dev-perl/GStreamer
dev-php/smarty
dev-python/cython
dev-python/dbus-python
dev-python/django
dev-python/pycairo
dev-python/pygobject
dev-python/pygtk
dev-python/pyxml
dev-tcltk/expect
dev-tex/feynmf
dev-util/byacc
dev-util/ccache
dev-util/meld
dev-util/unifdef
dev-vcs/git
dev-vcs/subversion
games-action/supertuxkart
games-arcade/supertux
games-arcade/xscavenger
games-kids/tuxmathscrabble
gnome-base/gconf
mail-client/mutt
mail-client/thunderbird
mail-filter/procmail
mail-filter/spamassassin
mail-mta/postfix
media-fonts/arphicfonts
media-fonts/baekmuk-fonts
media-fonts/corefonts
media-fonts/kochi-substitute
media-fonts/ttf-bitstream-vera
media-gfx/gimp
media-gfx/ristretto
media-gfx/sane-backends
media-gfx/xsane
media-gfx/xv
media-libs/gstreamer
media-libs/libcuefile
media-libs/libdvbpsi
media-libs/libid3tag
media-libs/libmikmod
media-libs/libmodplug
media-libs/libpng
media-libs/libreplaygain
media-libs/libsdl
media-libs/mesa
media-libs/openjpeg
media-libs/schroedinger
media-libs/tiff
media-sound/alsa-utils
media-sound/audacity
media-sound/cdparanoia
media-sound/cmus
media-sound/musepack-tools
media-sound/sox
media-video/dirac
media-video/dvdbackup
media-video/gtk-recordmydesktop
media-video/kino
media-video/motion
media-video/mplayer
media-video/vlc
media-video/xine-ui
net-analyzer/gnu-netcat
net-analyzer/httping
net-analyzer/nagios
net-analyzer/nagios-core
net-analyzer/nmap
net-analyzer/tcpdump
net-dialup/mgetty
net-dns/ddclient
net-fs/nfs-utils
net-fs/samba
net-ftp/gftp
net-libs/adns
net-libs/libvncserver
net-libs/openslp
net-mail/fetchmail
net-mail/tnef
net-misc/asterisk
net-misc/dhcpcd
net-misc/hylafaxplus
net-misc/iperf
net-misc/nx
net-misc/nxclient
net-misc/nxserver-freenx
net-misc/openvpn
net-misc/rdate
net-misc/socat
net-misc/telnet-bsd
net-misc/urlview
net-misc/whois
net-print/cups
net-print/cups-pdf
net-print/fax4cups
net-print/foomatic-db

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

On 09/03/2015 08:28 AM, Neil Bothwick wrote:
> On Thu, 03 Sep 2015 07:56:33 -0600, the...@sys-concept.com wrote:
> 
>> Here is /var/lib/portage/world
> 
> This contains quite a lot of libraries. As a rule of thumb, you should
> rarely have libraries in world as you don't use them - the software you
> have in world uses them. You can quickly clean them out with
> 
> sed -i /libs\//d /var/lib/portage/world

After running above I get:
sed: -e expression #1, char 7: unknown command: `/'

Shouldn't it be:
sed -i /libs\/d /var/lib/portage/world

This one doesn't produce any result.

> 
> Then run emerge --depclean --ask to clear out any that are no longer
> needed.

Thelma

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Neil Bothwick]

On Thu, 03 Sep 2015 07:56:33 -0600, the...@sys-concept.com wrote:

> Here is /var/lib/portage/world

This contains quite a lot of libraries. As a rule of thumb, you should
rarely have libraries in world as you don't use them - the software you
have in world uses them. You can quickly clean them out with

sed -i /libs\//d /var/lib/portage/world

Then run emerge --depclean --ask to clear out any that are no longer
needed.


-- 
Neil Bothwick

Time is the best teacher; unfortunately it kills all its students.


pgpI52Rv7PBwR.pgp
Description: OpenPGP digital signature

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 17:26, the...@sys-concept.com wrote:
> On 09/03/2015 08:28 AM, Neil Bothwick wrote:
>> On Thu, 03 Sep 2015 07:56:33 -0600, the...@sys-concept.com wrote:
>>
>>> Here is /var/lib/portage/world
>>
>> This contains quite a lot of libraries. As a rule of thumb, you should
>> rarely have libraries in world as you don't use them - the software you
>> have in world uses them. You can quickly clean them out with
>>
>> sed -i /libs\//d /var/lib/portage/world
> 
> After running above I get:
> sed: -e expression #1, char 7: unknown command: `/'
> 
> Shouldn't it be:
> sed -i /libs\/d /var/lib/portage/world
> 
> This one doesn't produce any result.

vim plus dd


> 
>>
>> Then run emerge --depclean --ask to clear out any that are no longer
>> needed.
> 
> Thelma
> 
> 
> 


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 15:56, the...@sys-concept.com wrote:
> On 09/03/2015 07:19 AM, Alan McKinnon wrote:
>> On 03/09/2015 15:06, Alec Ten Harmsel wrote:
>>> On Wed, Sep 02, 2015 at 02:25:47PM -0600, the...@sys-concept.com wrote:
 I've tried to post a "log" file to the http://pastebin.com/ you
 suggested but I can only paste the limited size file (not upload it).
 Since the txt file is 7.4Mb in size, I can not paste it.
 Though, I have compress the file as tar.gz (so it is only 267kB)
 here is the link:

 http://www.sysconcept.ca/audacity_error.tar.gz
> 
> I've changed the permission on this file to: apache:apache and rw
> the link above should work.
> 

 If somebody whats to look at it, I appreciate it.
 Meanwhile, I'll be following other folks suggestion and see if I can get
 a positive result to this error.

 Thelma
>>>
>>> I tried to download it but got a 403 Forbidden HTTP error.
>>>
>>> This might be a bad suggestion, but if you have a lot of time, it might
>>> be easier to uninstall audacity and any other packages that have been
>>> giving you problems, then `emerge -uDN --with-bdeps=y @world`, then
>>> `emerge --depclean', and then install all the stuff you need.
>>>
>>> I believe someone else mentioned checking /var/lib/portage/world and
>>> making sure that it doesn't contain any virtuals - it might also be good
>>> to remove from it any software that you do not directly need.
>>
>>
>> Even better - Thelma should just post the entire /var/lib/portage/world
>> file so we can advise what to take out.
>>
>> Most newbies clutter up their world needlessly, it takes a bit of
>> practice to grok what should be in their
> 
> Yes, the system is few years old.
> It make me wonder if the problem might be cause by me emerging:
> 
> emerge -avC libjpeg-turbo
> emerge -av1 media-libs/jpeg:0 media-libs/jpeg:62

Probably. None of that belongs in world

> 
> I'm using obsolete "nxclient-3.5.0.7" as I need it and there is no good
> replacement alternative on Gentoo.
> 
> Here is /var/lib/portage/world

You have a lot of clutter and junk in there. When you add dependant libs
to world, you remove portage's ability to do the right thing, and you
then have to do it all yourself. Humans never get this right - witness
your recent woes.

Comments inline, edit the world file directly:

> app-admin/gkrellm
> app-admin/syslog-ng
> app-admin/tmpwatch
> app-admin/webapp-config
> app-arch/cabextract
> app-arch/xarchiver
> app-benchmarks/cpuburn
> app-cdr/cdrtools
> app-cdr/dvd+rw-tools
> app-crypt/gnupg
> app-crypt/pinentry
> app-dicts/myspell-en
> app-editors/leafpad
> app-editors/nano
> app-emulation/dosemu
> app-emulation/virtualbox-bin
> app-emulation/virtualbox-modules
> app-forensics/chkrootkit
> app-forensics/rkhunter
> app-misc/ca-certificates
> app-office/glabels
> app-office/gnucash
> app-office/gnumeric
> app-office/openoffice-bin
> app-portage/cfg-update
> app-portage/eix
> app-portage/genlop
> app-portage/gentoolkit
> app-portage/portage-utils
> app-text/a2ps
> app-text/acroread
> app-text/dos2unix
> app-text/enscript
> app-text/evince
> app-text/flpsed
> app-text/ghostscript-gpl
> app-text/gv
> app-text/lcdf-typetools
> app-text/mpage
> app-text/pdfjam
> app-text/pdfshuffler
> app-text/pdftk
> app-text/texi2html
> app-text/wdiff
> dev-db/mysql
> dev-db/phpmyadmin
> dev-db/postgresql
> dev-db/postgresql:9.0
> dev-db/postgresql:9.1
> dev-java/icedtea-bin
> dev-java/java-config
> dev-java/oracle-jdk-bin
> dev-lang/lua
> dev-lang/php
> dev-lang/python
> dev-lang/swig
> dev-lang/tcl
> dev-lang/tk

> dev-libs/check
> dev-libs/libnl
> dev-libs/openssl

remove everything in dev-libs

> dev-perl/DBD-Pg
> dev-perl/GStreamer

Remove. No such package, unless its from an overlay

> dev-php/smarty
> dev-python/cython

> dev-python/dbus-python

Remove. It's a dep of many things

> dev-python/django

> dev-python/pycairo
> dev-python/pygobject
> dev-python/pygtk
> dev-python/pyxml

Remove all this py* stuff. They are deps

> dev-tcltk/expect
> dev-tex/feynmf
> dev-util/byacc
> dev-util/ccache
> dev-util/meld
> dev-util/unifdef
> dev-vcs/git
> dev-vcs/subversion
> games-action/supertuxkart
> games-arcade/supertux
> games-arcade/xscavenger
> games-kids/tuxmathscrabble

> gnome-base/gconf

Remove, this is a common dep

> mail-client/mutt
> mail-client/thunderbird
> mail-filter/procmail
> mail-filter/spamassassin
> mail-mta/postfix
> media-fonts/arphicfonts
> media-fonts/baekmuk-fonts
> media-fonts/corefonts
> media-fonts/kochi-substitute
> media-fonts/ttf-bitstream-vera
> media-gfx/gimp
> media-gfx/ristretto
> media-gfx/sane-backends
> media-gfx/xsane
> media-gfx/xv

> media-libs/gstreamer
> media-libs/libcuefile
> media-libs/libdvbpsi
> media-libs/libid3tag
> media-libs/libmikmod
> media-libs/libmodplug
> media-libs/libpng
> media-libs/libreplaygain
> media-libs/libsdl
> media-libs/mesa
> media-libs/openjpeg
> media-libs/schroedinger
> media-libs/tiff

Remove everything from media-libs

Re: [gentoo-user] system uptime

[Philip Webb]

Thanks for the responses to my inquiry re how long systems stay up.

My feeling for electronics is that equipment needs stability :
temperature + humidity should not vary much
& electrons need to flow thro' wires regularly.
That suggests a machine will do better if left running.

I don't like the idea of phoning my ISP's help desk
with an ill-defined problem : I can't be sure it will recur
if they ask me to go thro' a few restart cycles ;
the staff are probably trained to offer a few simple solutions ;
they may simply offer to send me a new router
or worse to send me a modem instead, which is harder to set up.
The problem probably lies in their server :
my router enjoys constant conditions & shows no other signs of wear ;
if that's the case, ISP help staff may not even know what has changed.

I'm planning to build a new machine in the next few weeks,
so I also now plan to set it up to do suspend, hibernate etc ;
I've grabbed a few dox which explain what's involved.
Mick mentioned 'S3' = 'sleep' & 'S4' = 'hibernation' in this connection :
what do they mean ?  Do they correspond to saving to RAM/disk ?

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

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

On 09/03/2015 09:26 AM, the...@sys-concept.com wrote:
> On 09/03/2015 08:28 AM, Neil Bothwick wrote:
>> On Thu, 03 Sep 2015 07:56:33 -0600, the...@sys-concept.com wrote:
>>
>>> Here is /var/lib/portage/world
>>
>> This contains quite a lot of libraries. As a rule of thumb, you should
>> rarely have libraries in world as you don't use them - the software you
>> have in world uses them. You can quickly clean them out with
>>
>> sed -i /libs\//d /var/lib/portage/world
> 
> After running above I get:
> sed: -e expression #1, char 7: unknown command: `/'
> 
> Shouldn't it be:
> sed -i /libs\/d /var/lib/portage/world
> 
> This one doesn't produce any result.

Yes, it worked got rid of all "libs" form world.

Thelma

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alexander Kapshuk]

On Thu, Sep 3, 2015 at 6:33 PM,   wrote:
>
> On 09/03/2015 09:26 AM, the...@sys-concept.com wrote:
>> On 09/03/2015 08:28 AM, Neil Bothwick wrote:
>>> On Thu, 03 Sep 2015 07:56:33 -0600, the...@sys-concept.com wrote:
>>>
 Here is /var/lib/portage/world
>>>
>>> This contains quite a lot of libraries. As a rule of thumb, you should
>>> rarely have libraries in world as you don't use them - the software you
>>> have in world uses them. You can quickly clean them out with
>>>
>>> sed -i /libs\//d /var/lib/portage/world
>>
>> After running above I get:
>> sed: -e expression #1, char 7: unknown command: `/'
>>
>> Shouldn't it be:
>> sed -i /libs\/d /var/lib/portage/world

Enclosing the 'sed' commands in single quotes would do it:

sed -i '/libs\//d' /var/lib/portage/world

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alec Ten Harmsel]

On Thu, Sep 03, 2015 at 04:27:26PM +0200, Alan McKinnon wrote:
> On 03/09/2015 15:56, the...@sys-concept.com wrote:
> > dev-libs/check
> > dev-libs/libnl
> > dev-libs/openssl
> 
> remove everything in dev-libs

Not necessarily - dev-libs/check is for unit testing. It's not a direct
dependency of any packages (or at least, shouldn't be), but it comes in
useful for developers. It's in my world.

Alec

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-02 21:30 GMT-03:00 David M. Fellows :

> On Wed, 2 Sep 2015 14:19:24 -0300
> Francisco Ares wrote -
> > --089e013a029e929a39051ec6e045
> > Content-Type: text/plain; charset=UTF-8
> >
> > Hi,
> >
> > Sorry for such WAY out of topic message, but Gentoo users are also way
> out
> > of regular computer users.
> >
> > I intend to learn more deep details about networking intrinsics,
> (packets,
> > ports, negotiation, UDP, multicast, unicast, TCP, ethernet, DHCP,
> > protocols, and so on) so I decided to recur to this list.  Googling the
> > terms, just gets me to network administration and equipment
> interconnection.
> >
> > Any hints on web resources for this research?
>
> W. Richard Stevens wrote *the* books on internet protocols back in the days
> when the world was young.  He died in 1999, so they are a bit dated, but
> most
> are available online as free pdfs. They are voluminous.
>
> Try google searchs
>   "Richard Stevens" tcp
>   and "Richard Stevens" unix network programming
> for links.
> Dave F
>


That looks promissing, thanks Dave

Best Regards,
Francisco

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-03 6:51 GMT-03:00 Alan McKinnon :

> On 03/09/2015 03:16, James wrote:
> > Alan McKinnon  gmail.com> writes:
> >
> >
> >>> Last, I suggest a parallel learning of C/C++ as it really helps
> >
> >> ^this^, after the basics are fully mastered.
> >
> >> netmasks make no sense at all until bitwise operators are fully
> >> understood. Even CIDR notation is not really obvious until you
> >> understand what languages like C do with the 32 bit words we call IP
> >> addresses. All x10 when IPv6 comes into play
> >
> >
> > Huh. I find teaching networking, including the intricacies  of advanced
> > protocol design, implementation and debugging, are far simpler if
> > folks know at least one programming language. Bit manipulations
> > are but one part of logic, sequential circuits timing and such
> > of the Computer Engineer's domain.  In my experience, if folks read too
> > much, but do not play with some codes  on actual hardware, it all
> becomes a
> > giant nebula. I guess I just like the practical side of these issues, to
> get
> > folks hooked on hardware.
>
> Yes, knowing at least one language is key
>
> >
> >
> > How a serial port (rs_232) works and the putting ppp over that is very
> > keen for teaching networking. ymmv. You can also use a protocol analyzer
> to
> > see some cool things. Many codes are published and looking at how a
> > microprocessor handles basic packets is very stimulating and encouraging.
> > Too bad most kids now days do not get to work on embedded hardware and
> build
> > up an executive or state machine and send/recieve data over interfaces.
> > Granted I worked in the world where assembler was
> > king (embedded) and assembler folks learning C and tcp/ip were easily
> amazed
> > and happy to migrate from assembler to C.
>
> I think the critical thing is to have a good grasp of what the
> technology you use is built on.
>
> C is a thin wrapper around assembly so to master C you should know cpu
> instructions, logic and at lest something about hardware. I don't
> believe it's possible or desirable to completely abstract something like
> C away from those things and still use it well. It's C, not the ISO 7
> layer model
>
> >
> > As Joost pointed out, I guess it really depends on the background of
> > the student. Being a hardware guy, I guess my focus is tainted
>
> I'm a hardware and electronics guy too. I spent hours in college
> building circuits with breadboards, 74xx TTL chips and bits of wire
> before they'd let me move onto the next thing
>
> >
> > So, fair enough, but how long (exactly what are the basics) do you
> > read before you go to the lab and play? Labs are always more fun
> > than classrooms, lectures and stuffy old farts.(gotcha!) ?
>
> The raspberry pi was specifically built to get back to those old days,
> the main designers were from the BBC micro era.
>
> Which I think is a wonderful idea.
>
> >
> >
> > cheers,
> > James
> >
> >
> >
> >
> >
> >
>
>
> --
> Alan McKinnon
> alan.mckin...@gmail.com
>
>
>
Me, too, a hardware guy, but having to learn high level stuff. Here at the
company that work for, we had a programmer a couple of years ago, that has
gone for a better opportunity. So I got his load.

Blinking a bunch of LEDs is where I started. The first ones with simple
transistors, resistors and capacitors, TTLs were next, and then, finally, a
Z80 with an UV EPROM, having to be programmed at the university lab in a
terribly monstrous gig - there was a teletype (remember those?) where a
paper tape had to be punched with the byte codes, hand assembled from
mnemonics, the tape transferred to another part where it was read while the
bytes been burnt to the EPROM; if one missed or twisted a byte, everything
had to be done again, program tapes being literally patched over and over.
Nowadays it all look very funny, but not at all on those days with a final
degree project deadline approaching ;-)

Thanks for the opportunity for an old story to be remembered.
Francisco

Re: [gentoo-user] system uptime

[Philip Webb]

150903 Neil Bothwick wrote:
> On Thu, 3 Sep 2015 11:32:52 -0400, Philip Webb wrote:
>> Mick mentioned 'S3' = 'sleep' & 'S4' = 'hibernation' in this connection:
>> what do they mean ?  Do they correspond to saving to RAM/disk ?
> Short answer: Yes
> Long answer:
> https://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface#Power_states

Thanks.

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

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-02 17:23 GMT-03:00 Alan McKinnon :

> On 02/09/2015 21:43, J. Roeleveld wrote:
> > On Wednesday, September 02, 2015 02:19:24 PM Francisco Ares wrote:
> >> Hi,
> >>
> >> Sorry for such WAY out of topic message, but Gentoo users are also way
> out
> >> of regular computer users.
> >>
> >> I intend to learn more deep details about networking intrinsics,
> (packets,
> >> ports, negotiation, UDP, multicast, unicast, TCP, ethernet, DHCP,
> >> protocols, and so on) so I decided to recur to this list.  Googling the
> >> terms, just gets me to network administration and equipment
> interconnection.
> >>
> >> Any hints on web resources for this research?
> >
> > It would depend on the level you are at now. :)
> >
> > Generally, I know more than enough about how it all works to do my job
> and
> > keep my own systems running reliably.
> >
> > But generally I simply listen when the likes of Alan McKinnon start
> talking
> > about networking.
>
> Hey, that's me!
>
> As it turns out, I got a call last week from an old mate who needed
> someone to deliver his 2-day TCP/IP course on short notice. I had 2 days
> free anyway so I help out.
>
> It all went well till we got into the dirty details of TCP header
> fields. You know how that stuff works - a whole bunch of fields that we
> mostly ignore and concentrate on just the few we know are important.
> Anyway, there was me standing in front of a class going down the list.
> And all I could think of was "WTF is most of this stuff??? Half of these
> fields I've never heard of!"
>
> There was more fun to come. Someone asked to clarify the exact
> differences between unicast, multicast, anycast and any other *cast that
> happens to be. Holy cow. Try explain that off the cuff without having
> time to think the answer through first :-)
>
> To the OP:
>
> Someone suggested RUTE. That's a good one, it may be 14 years old, but
> networking basics have not changed. The Linux Network Administrator's
> Guide available at tldp.org is also worth reading.
>
> And then wikipedia too. Technical facts are usually reliable there and
> most articles give you nice pictures and tables without assuming you
> already know it all anyway.
>
> Finally you already have Gentoo, which is probably the best tool you
> could have to find out such stuff. Read up on a topic, grasp the basic
> theory, then follow it all through on Gentoo seeing how the bits fit
> together.
>
> For the full picture in strict technical language, nothing beats the
> proper Internet RFCs. They are not for the faint-hearted though.
>
> I don't want to scare you off but working in spare time it probably
> takes something like a year to go from networking user to having a
> decent depth of knowledge about it. It's all logical, all the info is
> there, and it can be understood. There's just so much of it :-)
>
>
> >
> > You could start with sites like:
> >
> >
> http://web.stanford.edu/class/msande91si/www-spr04/readings/week1/InternetWhitepaper.htm
> >
> > --
> > Joost
> >
>
>
> --
> Alan McKinnon
> alan.mckin...@gmail.com
>
>
>
Thanks, Alan.

Well, I have noticed that, for the few details I got an eye on, it will
take a good time for an deep dive in.

I will start to look into some RFCs and see how much can be digested. Also,
downloaded RUTE to read during lunch, alternating with some RFCs ;-)

Best Regards,
Francisco

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Francisco Ares]

Thanks, James.

I already know C and a bit of C++, and that's why I want to understand
deeply about networking protocols, there are some tricks that I see some
device's closed source SDKs doing that I would like to reproduce. And
probably - if they worth it - publish as open source, of course.

Best Regards,
Francisco



2015-09-02 17:57 GMT-03:00 James :

> Francisco Ares  gmail.com> writes:
>
> >
> > Hi,
> > Sorry for such WAY out of topic message, but Gentoo users are also way
> out
> of regular computer users.
> >
> > I intend to learn more deep details about networking intrinsics,
> (packets,
> ports, negotiation, UDP, multicast, unicast, TCP, ethernet, DHCP,
> protocols,
> and so on) so I decided to recur to this list.  Googling the terms, just
> gets me to network administration and equipment interconnection.
> >
> > Any hints on web resources for this research?
> >
> > Thanks a lot and
> >
> > Best Regards,
> > Francisco
> >
>
>
> You can always build your own router, and learn about those internals
> as they come up while setting up your home router. All you need
> are some old PC parts laying around, gentoo and this guide [1].
> Iptables if a really cool network applications as is Network Address
> Translation (NAT).
>
>
> Also reading key "RFCs" is the way to go [2]. But try not to get hung up
> On the really cool RFCs like OSPF or SIP, as they are ever evolving
> and looking at sources it actually better.
>
>
> Last, I suggest a parallel learning of C/C++ as it really helps with
> unix/linux/networking if can look at software sources and see what is going
> on. Bash and Python are really important too. It's a lifelong journey, so
> relax and enjoy the experiences and try not to get frustrated.
>
>
> hth,
> James
>
> [1] https://wiki.gentoo.org/wiki/Home_Router
>
> [2] https://en.wikipedia.org/wiki/List_of_RFCs

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Mick]

On Thursday 03 Sep 2015 20:11:40 the...@sys-concept.com wrote:
> When I boot the computer I see some complain message about "jpeg" but it
> scroll too fast, couldn't read it. Is there a way to pause the boot
> process?

Configure your /etc/rc.conf temporarily to capture a log file and reboot 
before you check the output.
-- 
Regards,
Mick


signature.asc
Description: This is a digitally signed message part.

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

On 09/03/2015 01:30 PM, Mick wrote:
> On Thursday 03 Sep 2015 20:11:40 the...@sys-concept.com wrote:
>> When I boot the computer I see some complain message about "jpeg" but it
>> scroll too fast, couldn't read it. Is there a way to pause the boot
>> process?
> 
> Configure your /etc/rc.conf temporarily to capture a log file and reboot 
> before you check the output.

I solved this one an old fashion way :-) capture the boot process with
video camera.

It was nothing serious, "slim" was complaining about jpeg library; I
re-emerge slim.

Thelma

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Fernando Rodriguez]

On Thursday, September 03, 2015 4:27:26 PM Alan McKinnon wrote:
> Back to jpeg. One of your first recent posts was about the conflict
> between jpeg and jpeg-turbo. You had nxclient requiring jpeg and
> something else requiring jpeg-turbo. These conflict and cannot co-exist.
> 
> You are going to have to give up one of those apps, or run them on
> separate hosts. No way past this.

Just changing the dep on the ebuild to the virtual and rebuilding may do it. 
Since it's no longer in the tree it should be put on a overlay anyways.

The problem though, is that media-libs/libjpeg-turbo as it's built on gentoo 
is not exactly ABI compatible with media-libs/jpeg:62. It's may be close 
enough that it may even fool revdep-rebuild but media-libs/jpeg:62 does not 
contain versioned symbols, that's why the linker complains about the 
@LIBJPEG_6.2 symbols. He may need a command to rebuild everything that depends 
on jpeg manually. Or fix that ebuild, and switch back to jpeg-turbo.


-- 
Fernando Rodriguez

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Fernando Rodriguez]

On Thursday, September 03, 2015 9:10:26 AM walt wrote:
> On Wed, 2 Sep 2015 15:34:43 +0200
> Alan McKinnon  wrote:
> 
> > On 02/09/2015 15:04, Neil Bothwick wrote:
> > > On Wed, 2 Sep 2015 05:24:33 -0700, walt wrote:
> > >   
> > >> If the devs can't explain slots to their
> > >> users then they don't understand it themselves.  (Hm.  That phrase
> > >> sounds familiar.  Where did I get that?)  
> > > 
> > > I think it is an Einstein quote that says something like "if you
> > > can't explain it in simple terms, you don't understand it". He was
> > > probably having a pop at Niels Bohr and quantum theory at the time.
> > > 
> > > Bohr said something like "if thinking about quantum theory doesn't
> > > give you a headache, you don't understand it".
> > > 
> > >   
> > 
> > And Feynmann said something along the lines of "Anyone who claims to
> > understand quantum mechanics, doesn't".
> > 
> > Back to subslots and not replying to Neil directly: They aren't that
> > hard to grasp, they look like this:
> > 
> > cat/pkg/pkg-1.2:3/4
> > 
> > The SLOT is 3 and the subslot is 4. As usual, different versions of
> > the same package in different SLOTs can co-exist. Subslots are a
> > different matter, and it's an unfortunate choice of name, as they are
> > *not* a subset of a SLOT. Look at ncurses:
> > 
> > [I] sys-libs/ncurses
> >  Available versions:
> >  (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
> >  (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)
> > 
> > There's 2 SLOTs (0 and 5), and both have versions of subslot 5 and 6.
> > Subslots are most useful for things like api/abi versions where
> > upstream breaks these but don't increment the major version, this is
> > why we had endless issues in the past where emerge world broke stuff
> > horribly and it only got fixed much later when we could run
> > revdep-rebuild. Nowadays we have better tools, if the subslot changes
> > for a consumed library, then all consuming packages need to be
> > rebuilt.
> > 
> > Describing and defining subslots is not hard, neither are the
> > operators. The problem with subslots is the usual one - you have to
> > deal with real life, and in real life upstreams sometimes do peculiar
> > things to their code that doesn't exactly match the effect of a
> > subslot operation.
> > 
> > Or put another way: subslot docs describe the effect you should end up
> > with, it's not always the same thing as what you *do* end up with.
> > Finding that out means testing every possible circumstances and seeing
> > the results, but there's an infinite variety of those.
> 
> I just updated my virtualbox ~amd64 guest and all went well when I ran
> emerge ncurses:5/5, so I'm encouraged but not fearless about doing the
> same on my real machine.  I'm going to be quickpkged to the max before
> I try it.
> 
> BTW, emerge world on the vbox guest did not offer to touch ncurses in
> any way.  I had to do it manually as I just said.  
> 
> Leveraging Neil's quote:  thinking about slots (and their misnamed
> subslots) gives me a 4-dimensional headache.

I don't think they're misnamed, the problem is in our heads. As end users 
we've come to accept that they can be co-installed as the defining property of 
slots. But they're about the same in every other way AFAIU.

-- 
Fernando Rodriguez

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

[snip]

Thank you for your helping hand. Yes, did perform the surgery (removing the 
packages you suggested) and run: emerge --depclean -p analyzed every package it 
wants to remove and put it back in world.
 
Here are the results
> Comments inline, edit the world file directly:
> 
[snip]
> 
>> dev-libs/check
>> dev-libs/libnl
>> dev-libs/openssl
> 
> remove everything in dev-libs

gone

> 
>> dev-perl/DBD-Pg
>> dev-perl/GStreamer
> 
> Remove. No such package, unless its from an overlay

DBD-Pg is needed by SQL-ledger application I run
dev-perl/GStreamer is gone

> 
>> dev-php/smarty

gone

>> dev-python/cython
>> dev-python/dbus-python
> 
> Remove. It's a dep of many things

above gone

>> dev-python/django

gone

>> dev-python/pycairo
>> dev-python/pygobject
>> dev-python/pygtk
>> dev-python/pyxml
> 
> Remove all this py* stuff. They are deps

gone

[snip]
>> gnome-base/gconf
> 
> Remove, this is a common dep

gone

[snip]
> 
>> media-libs/gstreamer
>> media-libs/libcuefile
>> media-libs/libdvbpsi

above gone

>> media-libs/libid3tag

equery d media-libs/libid3tag
 * These packages depend on media-libs/libid3tag:
media-sound/audacity-2.0.2 (id3tag ? media-libs/libid3tag)
media-sound/sox-14.4.2 (id3tag ? media-libs/libid3tag)

>> media-libs/libmikmod
gone

>> media-libs/libmodplug

equery d media-libs/libmodplug
 * These packages depend on media-libs/libmodplug:
media-libs/sdl-mixer-1.2.12-r4 (modplug ? 
>=media-libs/libmodplug-0.8.8.4-r1[abi_x86_32(-)?,abi_x86_64(-)?,abi_x86_x32(-)?,abi_mips_n32(-)?,abi_mips_n64(-)?,abi_mips_o32(-)?,abi_ppc_32(-)?,abi_ppc_64(-)?,abi_s390_32(-)?,abi_s390_64(-)?])
media-libs/xine-lib-1.2.6-r1 (modplug ? >=media-libs/libmodplug-0.8.8.1)
media-sound/cmus-2.5.0-r1 (modplug ? >=media-libs/libmodplug-0.7)
media-video/ffmpeg-2.6.3 (modplug ? 
>=media-libs/libmodplug-0.8.8.4-r1[abi_x86_32(-)?,abi_x86_64(-)?,abi_x86_x32(-)?,abi_mips_n32(-)?,abi_mips_n64(-)?,abi_mips_o32(-)?,abi_ppc_32(-)?,abi_ppc_64(-)?,abi_s390_32(-)?,abi_s390_64(-)?])
media-video/vlc-2.1.5-r1 (modplug ? >=media-libs/libmodplug-0.8.8.1:0)

>> media-libs/libpng
>> media-libs/libreplaygain
>> media-libs/libsdl
>> media-libs/mesa
>> media-libs/openjpeg
>> media-libs/schroedinger
>> media-libs/tiff
> 
> Remove everything from media-libs
all media-libs gone

>> media-sound/alsa-utils
>> media-sound/audacity
>> media-sound/cdparanoia
>> media-sound/cmus
>> media-sound/musepack-tools
>> media-sound/sox
>> media-video/dirac
>> media-video/dvdbackup
>> media-video/gtk-recordmydesktop
>> media-video/kino
>> media-video/motion
>> media-video/mplayer
>> media-video/vlc
>> media-video/xine-ui
>> net-analyzer/gnu-netcat
>> net-analyzer/httping
>> net-analyzer/nagios
>> net-analyzer/nagios-core
>> net-analyzer/nmap
>> net-analyzer/tcpdump
>> net-dialup/mgetty
>> net-dns/ddclient
>> net-fs/nfs-utils
>> net-fs/samba
>> net-ftp/gftp
>> net-libs/adns
>> net-libs/libvncserver
>> net-libs/openslp
>> net-mail/fetchmail
>> net-mail/tnef
>> net-misc/asterisk
> 
> why are you running asterisk on a machine that is obviously a
> workstation?...

Yes, I do. It might not be the correct way of doing things but I find it 
practical to run server and workstation as one machine.
Easy to manage and only one computer working (running 7/24).

The way I manage it, I have several computers configures similarly (older ones) 
as backup.  I upgrade older one first, if most of the major programs I run are 
still running without problems I upgrade main server.  If something goes wrong, 
it is easier (less downtime) to just boot older machine point IP to it in 
firewall and I'm back and running.  
Sometimes troubleshoot something takes time.  
The problem could be harder as well (power supply, fan on CPU etc); so running 
server/workstation combo (as one machine) is easier.  If something goes wrong, 
I just boot older machine point IP address in firewall to older machine and I'm 
back in business in 10min.

[snip]
> 
>> net-print/foomatic-db
>> net-print/foomatic-db-engine
>> net-print/foomatic-db-ppds
> 
> You can probably remove foomatic, it's a dep on cups-filters

above gone

[snip]
>> sys-kernel/genkernel
>> sys-kernel/gentoo-sources
>> sys-kernel/gentoo-sources:3.10.17
>> sys-kernel/gentoo-sources:3.5.7
> 
> Wee bit behind on kernel versions...

Yes, I'm a bit behind. Sometimes upgrading to the latest/newest kernel tent to 
break something.
My idea is "if it ain't broke don't fix it" :-/

> 
>> sys-kernel/module-rebuild
> 
> Remove, no such package. Whatever it was, portage now does it internally

gone

[snip]
> 
>> virtual/ghostscript

gone

>> virtual/jdk
> 
> Remove both virtuals, replace with the actual ghostscript and jdk
> implementations you actually use

emerge -avq jdk
put back "virtual/jdk" to world

[snip]
> 
>> x11-apps/xdm
> 
> XDM? Really?
OK, gone 

> 
>> x11-base/xorg-server
> 
> Remove. The server is a dep for just about everything GUI-related
gone

[snip]
> 
>> x11-drivers/xf86-input-evdev
>> x11-drivers/xf86-input-keyboard
>> 

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-02 18:14 GMT-03:00 Alan McKinnon :

> On 02/09/2015 22:57, James wrote:
> > Last, I suggest a parallel learning of C/C++ as it really helps with
> > unix/linux/networking if can look at software sources and see what is
> going
> > on.
>
> ^this^, after the basics are fully mastered.
>
> netmasks make no sense at all until bitwise operators are fully
> understood. Even CIDR notation is not really obvious until you
> understand what languages like C do with the 32 bit words we call IP
> addresses. All x10 when IPv6 comes into play
>
>
> --
> Alan McKinnon
> alan.mckin...@gmail.com
>
>
>
Hi again, Alan.

Those basics are already well understood, and for now IPv6 is not even
being considered ;-)  Let me get the second level first.

Thanks.

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-03 3:20 GMT-03:00 J. Roeleveld :

> On Thursday, September 03, 2015 01:16:47 AM James wrote:
> > Alan McKinnon  gmail.com> writes:
> > > > Last, I suggest a parallel learning of C/C++ as it really helps
> > >
> > > ^this^, after the basics are fully mastered.
> > >
> > > netmasks make no sense at all until bitwise operators are fully
> > > understood. Even CIDR notation is not really obvious until you
> > > understand what languages like C do with the 32 bit words we call IP
> > > addresses. All x10 when IPv6 comes into play
> >
> > Huh. I find teaching networking, including the intricacies  of advanced
> > protocol design, implementation and debugging, are far simpler if
> > folks know at least one programming language. Bit manipulations
> > are but one part of logic, sequential circuits timing and such
> > of the Computer Engineer's domain.  In my experience, if folks read too
> > much, but do not play with some codes  on actual hardware, it all
> becomes a
> > giant nebula. I guess I just like the practical side of these issues, to
> get
> > folks hooked on hardware.
> >
> >
> > How a serial port (rs_232) works and the putting ppp over that is very
> > keen for teaching networking. ymmv. You can also use a protocol analyzer
> to
> > see some cool things. Many codes are published and looking at how a
> > microprocessor handles basic packets is very stimulating and encouraging.
> > Too bad most kids now days do not get to work on embedded hardware and
> build
> > up an executive or state machine and send/recieve data over interfaces.
> > Granted I worked in the world where assembler was
> > king (embedded) and assembler folks learning C and tcp/ip were easily
> amazed
> > and happy to migrate from assembler to C.
> >
> > As Joost pointed out, I guess it really depends on the background of
> > the student. Being a hardware guy, I guess my focus is tainted
> >
> > So, fair enough, but how long (exactly what are the basics) do you
> > read before you go to the lab and play? Labs are always more fun
> > than classrooms, lectures and stuffy old farts.(gotcha!) ?
> >
> >
> > cheers,
> > James
>
> If you want to base it on programming, I would recommend the following as
> well:
> http://beej.us/guide/bgnet/
>
> --
> Joost
>
>
Thanks again, Joost, that also looks quite promising.

Best Regards,
Francisco

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Mick]

On Thursday 03 Sep 2015 15:27:26 Alan McKinnon wrote:
> On 03/09/2015 15:56, the...@sys-concept.com wrote:

> > sys-kernel/gentoo-sources
> > sys-kernel/gentoo-sources:3.10.17
> > sys-kernel/gentoo-sources:3.5.7
> 
> Wee bit behind on kernel versions...

Yes, only the first is needed, *unless* thelma wants to keep 3.10.17 and 3.5.7 
for some reason.


> > sys-kernel/module-rebuild
> 
> Remove, no such package. Whatever it was, portage now does it internally

You can use sets to emerge ... sets of packages.

In this case:

 emerge -a @module-rebuild

will do what you're after.  To see list of all sets available look at the 
output of:

$ emerge --list-sets

(Hmm ... Some of them don't make much sense to me, e.g. what is 
selected/selected-packages/selected-sets?)


> > x11-base/xorg-x11
> > x11-drivers/nvidia-drivers
> > 
> > x11-drivers/xf86-input-evdev
> > x11-drivers/xf86-input-keyboard
> > x11-drivers/xf86-input-mouse
> > x11-drivers/xf86-video-fbdev
> > x11-drivers/xf86-video-nv
> > x11-drivers/xf86-video-tdfx
> > x11-drivers/xf86-video-vesa
> 
> Remove all xf86 drivers. They are controlled by INPUT_DEVICES and
> VIDEO_CARDS in make.conf

and if for some reason you want to re-emerge them you can: @x11-module-rebuild

-- 
Regards,
Mick


signature.asc
Description: This is a digitally signed message part.

[gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[walt]

On Wed, 2 Sep 2015 15:34:43 +0200
Alan McKinnon  wrote:

> On 02/09/2015 15:04, Neil Bothwick wrote:
> > On Wed, 2 Sep 2015 05:24:33 -0700, walt wrote:
> >   
> >> If the devs can't explain slots to their
> >> users then they don't understand it themselves.  (Hm.  That phrase
> >> sounds familiar.  Where did I get that?)  
> > 
> > I think it is an Einstein quote that says something like "if you
> > can't explain it in simple terms, you don't understand it". He was
> > probably having a pop at Niels Bohr and quantum theory at the time.
> > 
> > Bohr said something like "if thinking about quantum theory doesn't
> > give you a headache, you don't understand it".
> > 
> >   
> 
> And Feynmann said something along the lines of "Anyone who claims to
> understand quantum mechanics, doesn't".
> 
> Back to subslots and not replying to Neil directly: They aren't that
> hard to grasp, they look like this:
> 
> cat/pkg/pkg-1.2:3/4
> 
> The SLOT is 3 and the subslot is 4. As usual, different versions of
> the same package in different SLOTs can co-exist. Subslots are a
> different matter, and it's an unfortunate choice of name, as they are
> *not* a subset of a SLOT. Look at ncurses:
> 
> [I] sys-libs/ncurses
>  Available versions:
>  (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
>  (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)
> 
> There's 2 SLOTs (0 and 5), and both have versions of subslot 5 and 6.
> Subslots are most useful for things like api/abi versions where
> upstream breaks these but don't increment the major version, this is
> why we had endless issues in the past where emerge world broke stuff
> horribly and it only got fixed much later when we could run
> revdep-rebuild. Nowadays we have better tools, if the subslot changes
> for a consumed library, then all consuming packages need to be
> rebuilt.
> 
> Describing and defining subslots is not hard, neither are the
> operators. The problem with subslots is the usual one - you have to
> deal with real life, and in real life upstreams sometimes do peculiar
> things to their code that doesn't exactly match the effect of a
> subslot operation.
> 
> Or put another way: subslot docs describe the effect you should end up
> with, it's not always the same thing as what you *do* end up with.
> Finding that out means testing every possible circumstances and seeing
> the results, but there's an infinite variety of those.

I just updated my virtualbox ~amd64 guest and all went well when I ran
emerge ncurses:5/5, so I'm encouraged but not fearless about doing the
same on my real machine.  I'm going to be quickpkged to the max before
I try it.

BTW, emerge world on the vbox guest did not offer to touch ncurses in
any way.  I had to do it manually as I just said.  

Leveraging Neil's quote:  thinking about slots (and their misnamed
subslots) gives me a 4-dimensional headache.

Anyway, thanks for the helpful explanation.

Re: [gentoo-user] system uptime

[Neil Bothwick]

On Thu, 3 Sep 2015 11:32:52 -0400, Philip Webb wrote:

> I'm planning to build a new machine in the next few weeks,
> so I also now plan to set it up to do suspend, hibernate etc ;
> I've grabbed a few dox which explain what's involved.
> Mick mentioned 'S3' = 'sleep' & 'S4' = 'hibernation' in this
> connection : what do they mean ?  Do they correspond to saving to
> RAM/disk ?

Short answer: Yes

Long ansder:
https://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface#Power_states


-- 
Neil Bothwick

** I'm not going to get married again **
** I'll just find a woman I don't like and give her a house **


pgpzxM3XOq_NC.pgp
Description: OpenPGP digital signature

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Neil Bothwick]

On Thu, 3 Sep 2015 19:16:21 +0300, Alexander Kapshuk wrote:

> >>> sed -i /libs\//d /var/lib/portage/world  
> >>
> >> After running above I get:
> >> sed: -e expression #1, char 7: unknown command: `/'
> >>
> >> Shouldn't it be:
> >> sed -i /libs\/d /var/lib/portage/world  
> 
> Enclosing the 'sed' commands in single quotes would do it:
> 
> sed -i '/libs\//d' /var/lib/portage/world

Yes, I escaped the / with \ for sed but forgot to then escape the \
for the shell. The reason for deleting all lines containing libs/ rather
than just libs is to only remove entries from a *-libs category.

Sometimes you can try too hard :(


-- 
Neil Bothwick

Never ask a geek why, just nod your head and slowly back away


pgpEl0D9eDBhc.pgp
Description: OpenPGP digital signature

[gentoo-user] broken seamonkey :(

[lee]

Hi,

since quite a while, seamonkey and its relatives are completely broken
when it comes to use self-signed certificates.  They just refuse the
connection to the server, blocking you from accessing your email.

Is there still no solution for this problem?  I'm totally fed up with it
by now.  At work, I have frozen seamonkey at version 2.31 and
thunderbird at some outdated version that still works with the
certificates.  Googling for a solution doesn't reveal one, either.

Now I need seamonkey to access the email, and I can't very well turn it
back to an outdated version just for that.

BTW, if this won't be fixed, what are the replacements?


-- 
Again we must be afraid of speaking of daemons for fear that daemons
might swallow us.  Finally, this fear has become reasonable.

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 19:22, Alec Ten Harmsel wrote:
> On Thu, Sep 03, 2015 at 04:27:26PM +0200, Alan McKinnon wrote:
>> On 03/09/2015 15:56, the...@sys-concept.com wrote:
>>> dev-libs/check
>>> dev-libs/libnl
>>> dev-libs/openssl
>>
>> remove everything in dev-libs
> 
> Not necessarily - dev-libs/check is for unit testing. It's not a direct
> dependency of any packages (or at least, shouldn't be), but it comes in
> useful for developers. It's in my world.


OK. Such anomolies do happen, I have a few myself. Usually for my own
code which obviously has no ebuild.

The only truism is still true - a Gentoo user must know their system
well enough to know *why* they put something in world


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[Alan McKinnon]

On 03/09/2015 21:09, Francisco Ares wrote:
> 
> 
> 2015-09-02 17:23 GMT-03:00 Alan McKinnon  >:
> 
> On 02/09/2015 21:43, J. Roeleveld wrote:
> > On Wednesday, September 02, 2015 02:19:24 PM Francisco Ares wrote:
> >> Hi,
> >>
> >> Sorry for such WAY out of topic message, but Gentoo users are also way 
> out
> >> of regular computer users.
> >>
> >> I intend to learn more deep details about networking intrinsics, 
> (packets,
> >> ports, negotiation, UDP, multicast, unicast, TCP, ethernet, DHCP,
> >> protocols, and so on) so I decided to recur to this list.  Googling the
> >> terms, just gets me to network administration and equipment 
> interconnection.
> >>
> >> Any hints on web resources for this research?
> >
> > It would depend on the level you are at now. :)
> >
> > Generally, I know more than enough about how it all works to do my job 
> and
> > keep my own systems running reliably.
> >
> > But generally I simply listen when the likes of Alan McKinnon start 
> talking
> > about networking.
> 
> Hey, that's me!
> 
> As it turns out, I got a call last week from an old mate who needed
> someone to deliver his 2-day TCP/IP course on short notice. I had 2 days
> free anyway so I help out.
> 
> It all went well till we got into the dirty details of TCP header
> fields. You know how that stuff works - a whole bunch of fields that we
> mostly ignore and concentrate on just the few we know are important.
> Anyway, there was me standing in front of a class going down the list.
> And all I could think of was "WTF is most of this stuff??? Half of these
> fields I've never heard of!"
> 
> There was more fun to come. Someone asked to clarify the exact
> differences between unicast, multicast, anycast and any other *cast that
> happens to be. Holy cow. Try explain that off the cuff without having
> time to think the answer through first :-)
> 
> To the OP:
> 
> Someone suggested RUTE. That's a good one, it may be 14 years old, but
> networking basics have not changed. The Linux Network Administrator's
> Guide available at tldp.org  is also worth reading.
> 
> And then wikipedia too. Technical facts are usually reliable there and
> most articles give you nice pictures and tables without assuming you
> already know it all anyway.
> 
> Finally you already have Gentoo, which is probably the best tool you
> could have to find out such stuff. Read up on a topic, grasp the basic
> theory, then follow it all through on Gentoo seeing how the bits fit
> together.
> 
> For the full picture in strict technical language, nothing beats the
> proper Internet RFCs. They are not for the faint-hearted though.
> 
> I don't want to scare you off but working in spare time it probably
> takes something like a year to go from networking user to having a
> decent depth of knowledge about it. It's all logical, all the info is
> there, and it can be understood. There's just so much of it :-)
> 
> 
> >
> > You could start with sites like:
> >
> >
> 
> http://web.stanford.edu/class/msande91si/www-spr04/readings/week1/InternetWhitepaper.htm
> >
> > --
> > Joost
> >
> 
> 
> --
> Alan McKinnon
> alan.mckin...@gmail.com 
> 
> 
> 
> Thanks, Alan.
> 
> Well, I have noticed that, for the few details I got an eye on, it will
> take a good time for an deep dive in.
> 
> I will start to look into some RFCs and see how much can be digested.
> Also, downloaded RUTE to read during lunch, alternating with some RFCs ;-)


The absolute best network RFCs are the ones about coffeepot over HTTP,
and IP by carrier pigoen (or is it Avain IP? something like that)


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Fernando Rodriguez]

On Thursday, September 03, 2015 10:44:49 PM Alan McKinnon wrote:
> On 03/09/2015 21:09, Fernando Rodriguez wrote:
> > On Thursday, September 03, 2015 9:10:26 AM walt wrote:
> >> On Wed, 2 Sep 2015 15:34:43 +0200
> >> Alan McKinnon  wrote:
> >>
> >>> On 02/09/2015 15:04, Neil Bothwick wrote:
>  On Wed, 2 Sep 2015 05:24:33 -0700, walt wrote:
>    
> > If the devs can't explain slots to their
> > users then they don't understand it themselves.  (Hm.  That phrase
> > sounds familiar.  Where did I get that?)  
> 
>  I think it is an Einstein quote that says something like "if you
>  can't explain it in simple terms, you don't understand it". He was
>  probably having a pop at Niels Bohr and quantum theory at the time.
> 
>  Bohr said something like "if thinking about quantum theory doesn't
>  give you a headache, you don't understand it".
> 
>    
> >>>
> >>> And Feynmann said something along the lines of "Anyone who claims to
> >>> understand quantum mechanics, doesn't".
> >>>
> >>> Back to subslots and not replying to Neil directly: They aren't that
> >>> hard to grasp, they look like this:
> >>>
> >>> cat/pkg/pkg-1.2:3/4
> >>>
> >>> The SLOT is 3 and the subslot is 4. As usual, different versions of
> >>> the same package in different SLOTs can co-exist. Subslots are a
> >>> different matter, and it's an unfortunate choice of name, as they are
> >>> *not* a subset of a SLOT. Look at ncurses:
> >>>
> >>> [I] sys-libs/ncurses
> >>>  Available versions:
> >>>  (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
> >>>  (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)
> >>>
> >>> There's 2 SLOTs (0 and 5), and both have versions of subslot 5 and 6.
> >>> Subslots are most useful for things like api/abi versions where
> >>> upstream breaks these but don't increment the major version, this is
> >>> why we had endless issues in the past where emerge world broke stuff
> >>> horribly and it only got fixed much later when we could run
> >>> revdep-rebuild. Nowadays we have better tools, if the subslot changes
> >>> for a consumed library, then all consuming packages need to be
> >>> rebuilt.
> >>>
> >>> Describing and defining subslots is not hard, neither are the
> >>> operators. The problem with subslots is the usual one - you have to
> >>> deal with real life, and in real life upstreams sometimes do peculiar
> >>> things to their code that doesn't exactly match the effect of a
> >>> subslot operation.
> >>>
> >>> Or put another way: subslot docs describe the effect you should end up
> >>> with, it's not always the same thing as what you *do* end up with.
> >>> Finding that out means testing every possible circumstances and seeing
> >>> the results, but there's an infinite variety of those.
> >>
> >> I just updated my virtualbox ~amd64 guest and all went well when I ran
> >> emerge ncurses:5/5, so I'm encouraged but not fearless about doing the
> >> same on my real machine.  I'm going to be quickpkged to the max before
> >> I try it.
> >>
> >> BTW, emerge world on the vbox guest did not offer to touch ncurses in
> >> any way.  I had to do it manually as I just said.  
> >>
> >> Leveraging Neil's quote:  thinking about slots (and their misnamed
> >> subslots) gives me a 4-dimensional headache.
> > 
> > I don't think they're misnamed, the problem is in our heads. As end users 
> > we've come to accept that they can be co-installed as the defining property 
of 
> > slots. But they're about the same in every other way AFAIU.
> > 
> 
> 
> [I] sys-libs/ncurses
>  Available versions:
>  (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
>  (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)
> 
> 
> I still maintain the name could be better and less confusing. The SLOTs
> for ncurses are simply. But what kind of sub-thing are subslots 5 & 6?
> And what are they a subset of?

I agree that it is confusing, and the ncurses case makes it more so because 
they're being used in an unusual way. But when you think about it makes some 
sense at least. The functionality of a subslot is a subset of the 
functionality of a slot. Another way to think of it is slots are major 
versions and subslots are minor versions that still have some ABI change...or 
sub-versions :P

-- 
Fernando Rodriguez

Re: [gentoo-user] audacity-2.0.5 failed to compile

[thelma]

On 09/03/2015 01:24 PM, Mick wrote:
> On Thursday 03 Sep 2015 15:27:26 Alan McKinnon wrote:
>> On 03/09/2015 15:56, the...@sys-concept.com wrote:
> 
>>> sys-kernel/gentoo-sources
>>> sys-kernel/gentoo-sources:3.10.17
>>> sys-kernel/gentoo-sources:3.5.7
>>
>> Wee bit behind on kernel versions...
> 
> Yes, only the first is needed, *unless* thelma wants to keep 3.10.17 and 
> 3.5.7 
> for some reason.

Keeping in world only:
sys-kernel/gentoo-sources

will keep the newest kernel in world and unmerge 3.10.17 the one I'm using

Thelma

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 21:11, the...@sys-concept.com wrote:
> [snip]
> 
> Thank you for your helping hand. Yes, did perform the surgery (removing the 
> packages you suggested) and run: emerge --depclean -p analyzed every package 
> it wants to remove and put it back in world.
>  


I see you are getting the hang of it but I still see some oddities in
your reply. There's a few things to keep in mind about Gentoo, that are
worth repeating. Eventually, I does all imprint in your brain :-)

Every old-timer here has gone through this learning process - even the
old farts like me (and James...) - and it took us many years to figure
out as a community how to deal with world.

Cleanups: Sometimes when cleaning up you'll delete something you really
do need, and you forget why you put it in world. Maybe emerge world puts
it back, but more likely stuff just breaks. Keep a list of all removals
so if you find breakage you can add things back. The classic case is
libs you need for your own code - you probably don't have an ebuild for
that and therefore no deps for portage to use.

Virtuals: you don't add the virtual to world. A virtual is a collection
of packages that all do the same thing and you can pick which one you
want. So you add oracle-jdk-bin to world, and the ebuild depends on
virtual/jdk. oracle-jdk-bin satisfied the virtual, so all is good. If
you unmerge oracle-jdk-bin and use icedtea instead, everything still
works. If you add a virtual to world, portage tends to just pick the
first one in the list which might not be what you want. Rather be explicit.

Sets: You have a many-purpose machine so you might find sets useful,
mostly because you can't add comments to world. You can with sets. They
are just files in /etc/portage/sets/ that list packages. You add them to
the system with emerge @. Here is one of mine:

$ cat /etc/portage/sets/alan-kde
kde-apps/kdeadmin-meta
kde-apps/kdeartwork-meta
kde-apps/kdebase-meta
kde-apps/kdebase-runtime-meta
kde-apps/kdegraphics-meta
kde-apps/kdemultimedia-meta
kde-apps/kdenetwork-meta
kde-apps/kdeutils-meta
#kde-base/kde-meta:4
#kde-base/kdeaccessibility-meta:4
#kde-base/kdebindings-meta:4
#kde-base/kdeedu-meta:4
#kde-base/kdegames-meta:4
#kde-base/kdepim-meta:4
#kde-base/kdesdk-meta:4
#kde-base/kdetoys-meta:4
#kde-base/kdewebdev-meta:4

and install using emerge -av@alan-kde

portage expands the name to the contents of the file and merges them.
Easy peasy.

You edit set files by hand so you can comment them. Perhaps you could
create a set for each major thing you do with that computer and make a
set for each one. This way you can easily keep track of major packages
types and comment *why* you did it.

jpeg: if it works the best is to remove all jpeg packages from world and
let portage deal with it. None of my gentoo machines have a jpeg package
in world. But you have nxclient which needs an old jpeg. We you can get
away with automagic, we'll have to see what happens when you do a full
emerge world

soxr: I'm not sure why this is giving a problem. If ffmpeg needs it, it
should be pulled in directly

jpeg on boot: Depends when the error happens, if you use openrc and it
happens during runlevel start, you can read /var/log/rc.log. Before that
point there's dmesg. We'd need to have more detail to answer better.

USE: There's no such thing as a correct USE :-) It's all just user
choices. So if it does what you want, it's all good. The only thing you
should not do is start USE with -*. A few people here do that and swear
by it, but it comes with a massive maintenance load on you, and vast
potential for side effects as you remove things you may need, and you
don't know you need them. That's why we have profiles, to set up a
minimally correct USE




> Here are the results
>> Comments inline, edit the world file directly:
>>
> [snip]
>>
>>> dev-libs/check
>>> dev-libs/libnl
>>> dev-libs/openssl
>>
>> remove everything in dev-libs
> 
> gone
> 
>>
>>> dev-perl/DBD-Pg
>>> dev-perl/GStreamer
>>
>> Remove. No such package, unless its from an overlay
> 
> DBD-Pg is needed by SQL-ledger application I run
> dev-perl/GStreamer is gone
> 
>>
>>> dev-php/smarty
> 
> gone
> 
>>> dev-python/cython
>>> dev-python/dbus-python
>>
>> Remove. It's a dep of many things
> 
> above gone
> 
>>> dev-python/django
> 
> gone
> 
>>> dev-python/pycairo
>>> dev-python/pygobject
>>> dev-python/pygtk
>>> dev-python/pyxml
>>
>> Remove all this py* stuff. They are deps
> 
> gone
> 
> [snip]
>>> gnome-base/gconf
>>
>> Remove, this is a common dep
> 
> gone
> 
> [snip]
>>
>>> media-libs/gstreamer
>>> media-libs/libcuefile
>>> media-libs/libdvbpsi
> 
> above gone
> 
>>> media-libs/libid3tag
> 
> equery d media-libs/libid3tag
>  * These packages depend on media-libs/libid3tag:
> media-sound/audacity-2.0.2 (id3tag ? media-libs/libid3tag)
> media-sound/sox-14.4.2 (id3tag ? media-libs/libid3tag)
> 
>>> media-libs/libmikmod
> gone
> 
>>> media-libs/libmodplug
> 
> equery d media-libs/libmodplug
>  * These packages depend on 

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 21:56, the...@sys-concept.com wrote:
> On 09/03/2015 01:24 PM, Mick wrote:
>> On Thursday 03 Sep 2015 15:27:26 Alan McKinnon wrote:
>>> On 03/09/2015 15:56, the...@sys-concept.com wrote:
>>
 sys-kernel/gentoo-sources
 sys-kernel/gentoo-sources:3.10.17
 sys-kernel/gentoo-sources:3.5.7
>>>
>>> Wee bit behind on kernel versions...
>>
>> Yes, only the first is needed, *unless* thelma wants to keep 3.10.17 and 
>> 3.5.7 
>> for some reason.
> 
> Keeping in world only:
> sys-kernel/gentoo-sources
> 
> will keep the newest kernel in world and unmerge 3.10.17 the one I'm using


Kernel sources work differently. They are the only package where every
version must be it's own SLOT, and one of very few that only unpacks
sources and have no equivalent to a make step.

gentoo-sources will always add the latest sources when you update world,
and then they stay there forever. You must manually build a kernel and
install it, and unmerging the package only removes the original sources.

By all means have kernel sources with versions or slots in world, it
only means that you want to keep the original sources around so portage
won't remove them. To fully delete a kernel version you have to unmerge
the package, rm the source directory in /usr/src, rm the modules dir in
/lib/modules, rm the files in /boot and edit grub.conf


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Francisco Ares]

2015-09-03 11:31 GMT-03:00 James :

> Francisco Ares  gmail.com> writes:
>
>
> > Sorry for such WAY out of topic message, but Gentoo users are also way
> > out of regular computer users.
>
> Far out man. some psychedelic mood music, say Pink Floyd Ummagumma::
> Several Species of Small Furry Animals Gathered Together in a Cave...[2]
> (now that we got the proper 'grove on'.let's roll!
>
>
> > I intend to learn more deep details about networking intrinsics,
> > (packets, ports, negotiation, UDP, multicast, unicast, TCP, ethernet,
> > DHCP, protocols, and so on) so I decided to recur to this list.
> > Francisco
>
>
> Well, if you really want to 'Get Smart'  as in get to 'the edge' of all of
> this, why not just jump straight to (SDN) Software Defined Networks and
> such
> things as OpenVswitch on a cluster? [1]. Our good friends at Rackspace are
> very progressive in their ideas and systems development. As you learn about
> SDN, you can backfill what the old farts have done in the past and be
> an active part of the future? Just search out SDN as it's a revolution in
> networking that just may be the disruptive thang that is so desperately
> needed.
>
>
> PS::I bet McKinnon can build a buffered serial port sniffer out of those
> old
> 74LSxx chips! I have a wide selection of those old 74LS chips in
> tubes, just in case you get an itch you need to scratch...
> The date codes have fadedall I can make out is you old fart.
>
>
>
> cheers,
> James
>
>
> [1] https://www.pytexas.org/2015/talk/45
>
> [2] https://www.youtube.com/watch?v=YrRhnaFaBsA



Thanks for the tip and for the video, those glory days of progressive rock
still amazes old and new generations.

To jump right now on a project like the ones you mentioned is, for now, out
of scope. But as more depth can be reached, well, who knows?

Best Regards,
Francisco

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Dale]

the...@sys-concept.com wrote:
> Thelma 

Suggestion.  To avoid that world file getting cluttered up again, why
not add -1 or --oneshot to your make.conf for emerge defaults?  That way
you can emerge away and not worry about it being added to your world
file.  If you really want something added to the world file, just add
--select y to the command line and it will override the -1 option set in
the defaults. 

Basically, if something gets added to the world file, you put effort
into it being added.  I've done mine that way for a long time now and it
is rare that something gets added to the world file that I don't know
about.  Also, if you install something just to play around with but
don't know if you really want it or not, it will remind you on the next
--depclean run.  It will be listed to remove since it's not in the world
file.  So, if you decide not to keep it around, it will be removed for
you when you clean the unneeded stuff out. 

Just a thought.

Dale

:-)  :-)

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Stefan G. Weichinger]

>> I think it is an Einstein quote 

[..]

> And Feynmann said something 

[..]

>>> I don't think they're misnamed, the problem is in our heads.

[..]

tl;dr

I don't care.
I want to be a gentoo-USER (see name of ml).

Don't get me wrong, I like fuzzing around with details etc (and most
people subscribed here know that from various threads I triggered ...).

But don't you agree that it *should* be possible to simply pull upgrades
from the distro of choice without solving "problems in my head" ?

IMO it's the definition of maintainers to keep that level of complexity
away from the plain users. Especially for users of the stable "branch"
(the OP of this thread wrote "stable" = amd64 machine).

I admit: maybe I miss some point here because I didn't read the whole
thread.

IMO "stable" should result in a (mostly ...) carefree and maintained
experience for the user.

opinions welcome.

Stefan

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Alan McKinnon]

On 03/09/2015 21:46, Francisco Ares wrote:
> Me, too, a hardware guy, but having to learn high level stuff. Here at
> the company that work for, we had a programmer a couple of years ago,
> that has gone for a better opportunity. So I got his load.
> 
> Blinking a bunch of LEDs is where I started. The first ones with simple
> transistors, resistors and capacitors, TTLs were next, and then,
> finally, a Z80 with an UV EPROM, 

Z80? A fine CPU. Built by a bunch of guys who left Intel early on,
convinced a full 8 bit cpu with 16 address lines was possible!

That chip powered so many home pcs in the late 70s and early 80s. That
and the 6502

# having to be programmed at the
> university lab in a terribly monstrous gig

A giant thing with a ZIL socket and huge UV tubes to blank the EPROM>
Yup, I remember them well

> - there was a teletype
> (remember those?) 

Golf balls. Gods, those things made a racket.
But worse still was line printers with 136 disks, one for each character
position. If you printed just the right things, you could get them to
play out a tune :-)

> where a paper tape had to be punched with the byte
> codes, hand assembled from mnemonics, the tape transferred to another
> part where it was read while the bytes been burnt to the EPROM; if one
> missed or twisted a byte, everything had to be done again, program tapes
> being literally patched over and over. Nowadays it all look very funny,
> but not at all on those days with a final degree project deadline
> approaching ;-)

Oh no, not punched paper tape. All I remember is thousands of tiny
yellow punched shards that floated everywhere like confetti

> Thanks for the opportunity for an old story to be remembered.

We old farts here reminisce about every 6 months or so. It usually
starts when someone asks a question like: did you ever work with those
original 8 inch floppies? and the thread goes on for days :-)


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[J. Roeleveld]

On Thursday, September 03, 2015 01:16:47 AM James wrote:
> Alan McKinnon  gmail.com> writes:
> > > Last, I suggest a parallel learning of C/C++ as it really helps
> > 
> > ^this^, after the basics are fully mastered.
> > 
> > netmasks make no sense at all until bitwise operators are fully
> > understood. Even CIDR notation is not really obvious until you
> > understand what languages like C do with the 32 bit words we call IP
> > addresses. All x10 when IPv6 comes into play
> 
> Huh. I find teaching networking, including the intricacies  of advanced
> protocol design, implementation and debugging, are far simpler if
> folks know at least one programming language. Bit manipulations
> are but one part of logic, sequential circuits timing and such
> of the Computer Engineer's domain.  In my experience, if folks read too
> much, but do not play with some codes  on actual hardware, it all becomes a
> giant nebula. I guess I just like the practical side of these issues, to get
> folks hooked on hardware.
> 
> 
> How a serial port (rs_232) works and the putting ppp over that is very
> keen for teaching networking. ymmv. You can also use a protocol analyzer to
> see some cool things. Many codes are published and looking at how a
> microprocessor handles basic packets is very stimulating and encouraging.
> Too bad most kids now days do not get to work on embedded hardware and build
> up an executive or state machine and send/recieve data over interfaces.
> Granted I worked in the world where assembler was
> king (embedded) and assembler folks learning C and tcp/ip were easily amazed
> and happy to migrate from assembler to C.
> 
> As Joost pointed out, I guess it really depends on the background of
> the student. Being a hardware guy, I guess my focus is tainted
> 
> So, fair enough, but how long (exactly what are the basics) do you
> read before you go to the lab and play? Labs are always more fun
> than classrooms, lectures and stuffy old farts.(gotcha!) ?
> 
> 
> cheers,
> James

If you want to base it on programming, I would recommend the following as 
well:
http://beej.us/guide/bgnet/

--
Joost

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[Mihamina Rakotomandimby]

In order to complete other answers, I woul like to point out a test 
environment that would be nice:

https://en.wikipedia.org/wiki/Marionnet

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Dale]

walt wrote:
> On Wed, 02 Sep 2015 10:53:00 -0500
> Dale  wrote:
>
>> walt wrote:
>>> Thanks also to wabe and Fernando for your replies. Just for the
>>> record I did the update this morning, which completed without
>>> errors. qlop shows that both updates completed, but eix shows that
>>> I now have only ncurses-5.9-r5 installed. (This is apparently the
>>> desired result, but I'm only guessing what the desired result
>>> really is.) I think every portage tool should announce very clearly
>>> whether a package is slotted/subslotted, and exactly which slot and
>>> subslot the package belongs in. The subject of slots is way too
>>> confusing to withhold such information. If the devs can't explain
>>> slots to their users then they don't understand it themselves. (Hm.
>>> That phrase sounds familiar. Where did I get that?)   
>>
>> I did this the other day, after skipping it several times while it got
>> sorted out, and I have this now.
>>
>> [IP-] [  ] sys-libs/ncurses-5.9-r5:0/5
>> [IP-] [  ] sys-libs/ncurses-5.9-r99:5/5
>>  
>>
>> Now, which of us is somewhat off the mark here?  :/ 
> Both of us have been off the mark so many times I can't count that
> high :p   I'm going to come back tomorrow when I'm more awake and
> re-read this entire thread because every reply so far is full of good
> info and advice.
>
> I'm assuming you did that ncurses update on some flavor of gentoo
> "unstable"?  (I know what a daredevil you are.)
>
> Just for laughs, what does your "qlop -l ncurses" tell you about what
> date you did the ncurses-5.9 update?
>
>
>
>


This is what I get:

root@fireball / # qlop -l ncurses
Mon Dec 15 17:17:17 2014 >>> sys-libs/ncurses-5.9-r3
Sun Mar 22 22:37:40 2015 >>> sys-libs/ncurses-5.9-r3
Sun Mar 29 19:28:46 2015 >>> sys-libs/ncurses-5.9-r3
Sat May 23 02:03:55 2015 >>> sys-libs/ncurses-5.9-r3
Sat Aug 29 05:35:33 2015 >>> sys-libs/ncurses-5.9-r5
Sat Aug 29 05:35:40 2015 >>> sys-libs/ncurses-5.9-r99
root@fireball / # 

What I find odd is that some updated with no problems, some had minor
issues and then some had some serious issues.  It seems a dev may have
had to fix a few things but it seems differences between everyone else's
system then complicated things further. 

I do have some unstable here.  It usually starts when I want the latest
KDE due to a bug fix.  It seems to roll downhill at that point.  Anyway,
it works.  ;-)

Dale

:-)  :-) 

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[J. Roeleveld]

On Thursday, September 03, 2015 09:28:05 AM Mihamina Rakotomandimby wrote:
> In order to complete other answers, I woul like to point out a test
> environment that would be nice:
> https://en.wikipedia.org/wiki/Marionnet

Cool! :)

I wrote something like this at uni.
The version the professor had was old and unstable.

In order to prove I knew my stuff better than him, he challenged me to rewrite 
it for him ;)

But Marionnet looks far more advanced.

--
Joost

Re: [gentoo-user] Re: [WAY OT] wanna learn networking internals

[Alan McKinnon]

On 03/09/2015 03:16, James wrote:
> Alan McKinnon  gmail.com> writes:
> 
> 
>>> Last, I suggest a parallel learning of C/C++ as it really helps 
> 
>> ^this^, after the basics are fully mastered.
> 
>> netmasks make no sense at all until bitwise operators are fully
>> understood. Even CIDR notation is not really obvious until you
>> understand what languages like C do with the 32 bit words we call IP
>> addresses. All x10 when IPv6 comes into play
> 
> 
> Huh. I find teaching networking, including the intricacies  of advanced
> protocol design, implementation and debugging, are far simpler if
> folks know at least one programming language. Bit manipulations
> are but one part of logic, sequential circuits timing and such
> of the Computer Engineer's domain.  In my experience, if folks read too
> much, but do not play with some codes  on actual hardware, it all becomes a
> giant nebula. I guess I just like the practical side of these issues, to get
> folks hooked on hardware.

Yes, knowing at least one language is key

> 
> 
> How a serial port (rs_232) works and the putting ppp over that is very
> keen for teaching networking. ymmv. You can also use a protocol analyzer to
> see some cool things. Many codes are published and looking at how a
> microprocessor handles basic packets is very stimulating and encouraging.
> Too bad most kids now days do not get to work on embedded hardware and build
> up an executive or state machine and send/recieve data over interfaces. 
> Granted I worked in the world where assembler was
> king (embedded) and assembler folks learning C and tcp/ip were easily amazed
> and happy to migrate from assembler to C.

I think the critical thing is to have a good grasp of what the
technology you use is built on.

C is a thin wrapper around assembly so to master C you should know cpu
instructions, logic and at lest something about hardware. I don't
believe it's possible or desirable to completely abstract something like
C away from those things and still use it well. It's C, not the ISO 7
layer model

> 
> As Joost pointed out, I guess it really depends on the background of
> the student. Being a hardware guy, I guess my focus is tainted

I'm a hardware and electronics guy too. I spent hours in college
building circuits with breadboards, 74xx TTL chips and bits of wire
before they'd let me move onto the next thing

> 
> So, fair enough, but how long (exactly what are the basics) do you
> read before you go to the lab and play? Labs are always more fun
> than classrooms, lectures and stuffy old farts.(gotcha!) ?

The raspberry pi was specifically built to get back to those old days,
the main designers were from the BBC micro era.

Which I think is a wonderful idea.

> 
> 
> cheers,
> James
> 
> 
> 
> 
> 
> 


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Alan McKinnon]

On 04/09/2015 00:05, Stefan G. Weichinger wrote:
> 
>>> I think it is an Einstein quote 
> 
> [..]
> 
>> And Feynmann said something 
> 
> [..]
> 
 I don't think they're misnamed, the problem is in our heads.
> 
> [..]
> 
> tl;dr
> 
> I don't care.
> I want to be a gentoo-USER (see name of ml).
> 
> Don't get me wrong, I like fuzzing around with details etc (and most
> people subscribed here know that from various threads I triggered ...).
> 
> But don't you agree that it *should* be possible to simply pull upgrades
> from the distro of choice without solving "problems in my head" ?
> 
> IMO it's the definition of maintainers to keep that level of complexity
> away from the plain users. Especially for users of the stable "branch"
> (the OP of this thread wrote "stable" = amd64 machine).
> 
> I admit: maybe I miss some point here because I didn't read the whole
> thread.
> 
> IMO "stable" should result in a (mostly ...) carefree and maintained
> experience for the user.


Sure. That's a valid POV.

Gentoo config should be declarative, not imperative. You should tell it
*what* you want, not *how* to do it. Packages and SLOTs are
user-visible, they define what you want. Sub-slots, although I think I
understand them, are a poorly though out implementation as they expose
part of the how to the user. Worse, they sometimes make the user decide.

And the name is confusing, it's all indicative of the concept not being
thought all the way through.


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] broken seamonkey :(

[lee]

Fernando Rodriguez  writes:

> On Thursday, September 03, 2015 9:53:39 PM lee wrote:
>> Hi,
>> 
>> since quite a while, seamonkey and its relatives are completely broken
>> when it comes to use self-signed certificates.  They just refuse the
>> connection to the server, blocking you from accessing your email.
>> 
>> Is there still no solution for this problem?  I'm totally fed up with it
>> by now.  At work, I have frozen seamonkey at version 2.31 and
>> thunderbird at some outdated version that still works with the
>> certificates.  Googling for a solution doesn't reveal one, either.
>> 
>> Now I need seamonkey to access the email, and I can't very well turn it
>> back to an outdated version just for that.
>> 
>> BTW, if this won't be fixed, what are the replacements?
>
> This[1] is for firefox but should work similarly. Scroll all the way down to 
> "bypassing the warning". There's also an about:config option, I *think* it's 
> this one[2].

Thank you.  The problem is that it doesn't let me add an exception. Only
the older versions do that.  All options to add an exception are
disabled.

There is 'browser.ssl_override_behavior', the value of which is
2. Guessing by what that means from [2], that should allow me to add an
exception.

'browser.xul.error_pages.enabled' is enabled.  There's also
'browser.xul.error_pages.expert_bad_cert', which is disabled.  Let's see
what that does ...  still cannot add an exception when I enable it.  [3]
would indicate that it's advisable to set it to "true".

Restarting seamonkey after changing it doesn't help.

There's nothing wrong with the certificate, either.  Older version work
just fine with it.  Mutt works fine with it.  Gnus works fine with it.
Evolution works fine with it.  All of those are more recent than
seamonkey 2.31.

I could resort to unencrypted connections on the LAN to be able to
upgrade the browsers and MUAs --- for security reasons, ironically ---
but some ppl with laptops need to be able to connect from anywhere over
the internet.  So omit all security and use VPN for those to make things
more secure by not using self-signed certificates but insecure
connections?


[3]: http://kb.mozillazine.org/Browser.xul.error_pages.expert_bad_cert

>
> [1] https://support.mozilla.org/en-US/kb/connection-untrusted-error-message
> [2] http://kb.mozillazine.org/Browser.ssl_override_behavior

-- 
Again we must be afraid of speaking of daemons for fear that daemons
might swallow us.  Finally, this fear has become reasonable.

[gentoo-user] dhcpd always shows "crashed" even though it's running

[Mike Edenfield]

For some reason, whenever I check the status of my startup scripts, 
dhcpd registers as "crashed". However, dhcpd is up and running and 
working fine. Normally I don't worry about it, but on those occasions 
where dhcpd does stop working, it's hard to tell if it's "fixed" or not.


What makes rc-status think something is crashed, and how can I fix this?

basement log # rc-status -v | grep crashed
 dhcpd [  crashed  ]
basement log # ps aux | grep dhcpd
root  2214  0.0  0.0   8268   876 pts/0S+   19:47   0:00 grep 
--colour=auto dhcpd
dhcp  2648  0.0  0.6  30028 12136 ?Ss   Aug29   0:00 
/usr/sbin/dhcpd -cf /etc/dhcp/dhcpd.conf -q -pf /var/run/dhcp/dhcpd.pid 
-lf /var/lib/dhcp/dhcpd.leases -user dhcp -group dhcp -chroot 
/chroot/dhcp enp0s7

Re: [gentoo-user] broken seamonkey :(

[Fernando Rodriguez]

On Friday, September 04, 2015 1:39:46 AM lee wrote:
> Fernando Rodriguez  writes:
> 
> > On Thursday, September 03, 2015 9:53:39 PM lee wrote:
> >> Hi,
> >> 
> >> since quite a while, seamonkey and its relatives are completely broken
> >> when it comes to use self-signed certificates.  They just refuse the
> >> connection to the server, blocking you from accessing your email.
> >> 
> >> Is there still no solution for this problem?  I'm totally fed up with it
> >> by now.  At work, I have frozen seamonkey at version 2.31 and
> >> thunderbird at some outdated version that still works with the
> >> certificates.  Googling for a solution doesn't reveal one, either.
> >> 
> >> Now I need seamonkey to access the email, and I can't very well turn it
> >> back to an outdated version just for that.
> >> 
> >> BTW, if this won't be fixed, what are the replacements?
> >
> > This[1] is for firefox but should work similarly. Scroll all the way down 
to 
> > "bypassing the warning". There's also an about:config option, I *think* 
it's 
> > this one[2].
> 
> Thank you.  The problem is that it doesn't let me add an exception. Only
> the older versions do that.  All options to add an exception are
> disabled.
> 
> There is 'browser.ssl_override_behavior', the value of which is
> 2. Guessing by what that means from [2], that should allow me to add an
> exception.
> 
> 'browser.xul.error_pages.enabled' is enabled.  There's also
> 'browser.xul.error_pages.expert_bad_cert', which is disabled.  Let's see
> what that does ...  still cannot add an exception when I enable it.  [3]
> would indicate that it's advisable to set it to "true".
> 
> Restarting seamonkey after changing it doesn't help.
> 
> There's nothing wrong with the certificate, either.  Older version work
> just fine with it.  Mutt works fine with it.  Gnus works fine with it.
> Evolution works fine with it.  All of those are more recent than
> seamonkey 2.31.
> 
> I could resort to unencrypted connections on the LAN to be able to
> upgrade the browsers and MUAs --- for security reasons, ironically ---
> but some ppl with laptops need to be able to connect from anywhere over
> the internet.  So omit all security and use VPN for those to make things
> more secure by not using self-signed certificates but insecure
> connections?
> 
> 
> [3]: http://kb.mozillazine.org/Browser.xul.error_pages.expert_bad_cert
> 
> >
> > [1] https://support.mozilla.org/en-US/kb/connection-untrusted-error-message
> > [2] http://kb.mozillazine.org/Browser.ssl_override_behavior
> 
> 

I got the same settings as yours on firefox 40 and it lets me add exceptions so 
it must be a seamonkey thing. I think I remember this happening with an 
earlier firefox, I don't remember how I fixed it or if it fixed itself after an 
update. Maybe you can add your cert to mozilla's store:

http://www.cyberciti.biz/faq/firefox-adding-trusted-ca/



-- 
Fernando Rodriguez

Re: [gentoo-user] dhcpd always shows "crashed" even though it's running

[Fernando Rodriguez]

On Thursday, September 03, 2015 8:09:02 PM Mike Edenfield wrote:
> For some reason, whenever I check the status of my startup scripts, 
> dhcpd registers as "crashed". However, dhcpd is up and running and 
> working fine. Normally I don't worry about it, but on those occasions 
> where dhcpd does stop working, it's hard to tell if it's "fixed" or not.
> 
> What makes rc-status think something is crashed, and how can I fix this?
> 
> basement log # rc-status -v | grep crashed
>   dhcpd [  crashed  ]
> basement log # ps aux | grep dhcpd
> root  2214  0.0  0.0   8268   876 pts/0S+   19:47   0:00 grep 
> --colour=auto dhcpd
> dhcp  2648  0.0  0.6  30028 12136 ?Ss   Aug29   0:00 
> /usr/sbin/dhcpd -cf /etc/dhcp/dhcpd.conf -q -pf /var/run/dhcp/dhcpd.pid 
> -lf /var/lib/dhcp/dhcpd.leases -user dhcp -group dhcp -chroot 
> /chroot/dhcp enp0s7
> 
> 

This is just a guess but it could be the permissions on the pid file on 
/chroot/dhcp/var/run/dhcp/. So stop the daemon, delete the file, check that the 
directory is owned by dhcp:dhcp and start the daemon again.

-- 
Fernando Rodriguez

Re: [gentoo-user] broken seamonkey :(

[Fernando Rodriguez]

On Thursday, September 03, 2015 9:53:39 PM lee wrote:
> Hi,
> 
> since quite a while, seamonkey and its relatives are completely broken
> when it comes to use self-signed certificates.  They just refuse the
> connection to the server, blocking you from accessing your email.
> 
> Is there still no solution for this problem?  I'm totally fed up with it
> by now.  At work, I have frozen seamonkey at version 2.31 and
> thunderbird at some outdated version that still works with the
> certificates.  Googling for a solution doesn't reveal one, either.
> 
> Now I need seamonkey to access the email, and I can't very well turn it
> back to an outdated version just for that.
> 
> BTW, if this won't be fixed, what are the replacements?

This[1] is for firefox but should work similarly. Scroll all the way down to 
"bypassing the warning". There's also an about:config option, I *think* it's 
this one[2].

[1] https://support.mozilla.org/en-US/kb/connection-untrusted-error-message
[2] http://kb.mozillazine.org/Browser.ssl_override_behavior

-- 
Fernando Rodriguez

Re: [gentoo-user] Re: Portage is proposing an ncurses update and I don't understand what it means

[Alan McKinnon]

On 03/09/2015 21:09, Fernando Rodriguez wrote:
> On Thursday, September 03, 2015 9:10:26 AM walt wrote:
>> On Wed, 2 Sep 2015 15:34:43 +0200
>> Alan McKinnon  wrote:
>>
>>> On 02/09/2015 15:04, Neil Bothwick wrote:
 On Wed, 2 Sep 2015 05:24:33 -0700, walt wrote:
   
> If the devs can't explain slots to their
> users then they don't understand it themselves.  (Hm.  That phrase
> sounds familiar.  Where did I get that?)  

 I think it is an Einstein quote that says something like "if you
 can't explain it in simple terms, you don't understand it". He was
 probably having a pop at Niels Bohr and quantum theory at the time.

 Bohr said something like "if thinking about quantum theory doesn't
 give you a headache, you don't understand it".

   
>>>
>>> And Feynmann said something along the lines of "Anyone who claims to
>>> understand quantum mechanics, doesn't".
>>>
>>> Back to subslots and not replying to Neil directly: They aren't that
>>> hard to grasp, they look like this:
>>>
>>> cat/pkg/pkg-1.2:3/4
>>>
>>> The SLOT is 3 and the subslot is 4. As usual, different versions of
>>> the same package in different SLOTs can co-exist. Subslots are a
>>> different matter, and it's an unfortunate choice of name, as they are
>>> *not* a subset of a SLOT. Look at ncurses:
>>>
>>> [I] sys-libs/ncurses
>>>  Available versions:
>>>  (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
>>>  (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)
>>>
>>> There's 2 SLOTs (0 and 5), and both have versions of subslot 5 and 6.
>>> Subslots are most useful for things like api/abi versions where
>>> upstream breaks these but don't increment the major version, this is
>>> why we had endless issues in the past where emerge world broke stuff
>>> horribly and it only got fixed much later when we could run
>>> revdep-rebuild. Nowadays we have better tools, if the subslot changes
>>> for a consumed library, then all consuming packages need to be
>>> rebuilt.
>>>
>>> Describing and defining subslots is not hard, neither are the
>>> operators. The problem with subslots is the usual one - you have to
>>> deal with real life, and in real life upstreams sometimes do peculiar
>>> things to their code that doesn't exactly match the effect of a
>>> subslot operation.
>>>
>>> Or put another way: subslot docs describe the effect you should end up
>>> with, it's not always the same thing as what you *do* end up with.
>>> Finding that out means testing every possible circumstances and seeing
>>> the results, but there's an infinite variety of those.
>>
>> I just updated my virtualbox ~amd64 guest and all went well when I ran
>> emerge ncurses:5/5, so I'm encouraged but not fearless about doing the
>> same on my real machine.  I'm going to be quickpkged to the max before
>> I try it.
>>
>> BTW, emerge world on the vbox guest did not offer to touch ncurses in
>> any way.  I had to do it manually as I just said.  
>>
>> Leveraging Neil's quote:  thinking about slots (and their misnamed
>> subslots) gives me a 4-dimensional headache.
> 
> I don't think they're misnamed, the problem is in our heads. As end users 
> we've come to accept that they can be co-installed as the defining property 
> of 
> slots. But they're about the same in every other way AFAIU.
> 


[I] sys-libs/ncurses
 Available versions:
 (0)5.9-r3 (~)5.9-r4 5.9-r5(0/5) (~)6.0-r1(0/6)
 (5)5.9-r99(5/5) (~)5.9-r101(5/5) (~)6.0(5/6)


I still maintain the name could be better and less confusing. The SLOTs
for ncurses are simply. But what kind of sub-thing are subslots 5 & 6?
And what are they a subset of?


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] broken seamonkey :(

[Dale]

lee wrote:
> Fernando Rodriguez  writes:
>
>> On Thursday, September 03, 2015 9:53:39 PM lee wrote:
>>> Hi,
>>>
>>> since quite a while, seamonkey and its relatives are completely broken
>>> when it comes to use self-signed certificates.  They just refuse the
>>> connection to the server, blocking you from accessing your email.
>>>
>>> Is there still no solution for this problem?  I'm totally fed up with it
>>> by now.  At work, I have frozen seamonkey at version 2.31 and
>>> thunderbird at some outdated version that still works with the
>>> certificates.  Googling for a solution doesn't reveal one, either.
>>>
>>> Now I need seamonkey to access the email, and I can't very well turn it
>>> back to an outdated version just for that.
>>>
>>> BTW, if this won't be fixed, what are the replacements?
>> This[1] is for firefox but should work similarly. Scroll all the way down to 
>> "bypassing the warning". There's also an about:config option, I *think* it's 
>> this one[2].
> Thank you.  The problem is that it doesn't let me add an exception. Only
> the older versions do that.  All options to add an exception are
> disabled.
>
> There is 'browser.ssl_override_behavior', the value of which is
> 2. Guessing by what that means from [2], that should allow me to add an
> exception.
>
> 'browser.xul.error_pages.enabled' is enabled.  There's also
> 'browser.xul.error_pages.expert_bad_cert', which is disabled.  Let's see
> what that does ...  still cannot add an exception when I enable it.  [3]
> would indicate that it's advisable to set it to "true".
>
> Restarting seamonkey after changing it doesn't help.
>
> There's nothing wrong with the certificate, either.  Older version work
> just fine with it.  Mutt works fine with it.  Gnus works fine with it.
> Evolution works fine with it.  All of those are more recent than
> seamonkey 2.31.
>
> I could resort to unencrypted connections on the LAN to be able to
> upgrade the browsers and MUAs --- for security reasons, ironically ---
> but some ppl with laptops need to be able to connect from anywhere over
> the internet.  So omit all security and use VPN for those to make things
> more secure by not using self-signed certificates but insecure
> connections?
>
>
> [3]: http://kb.mozillazine.org/Browser.xul.error_pages.expert_bad_cert
>
>> [1] https://support.mozilla.org/en-US/kb/connection-untrusted-error-message
>> [2] http://kb.mozillazine.org/Browser.ssl_override_behavior


I'm using seamonkey-2.33.1-r1 and I have not run into this problem.  If
you would like, I can check some settings and compare them to mine to
see if it would help fix yours.  Just let me know what you need. 

Surely this can be fixed.  Maybe I did something and don't recall it??

Dale

:-)  :-) 

Re: [gentoo-user] dhcpd always shows "crashed" even though it's running

[Rich Freeman]

On Thu, Sep 3, 2015 at 8:09 PM, Mike Edenfield  wrote:
> For some reason, whenever I check the status of my startup scripts, dhcpd
> registers as "crashed". However, dhcpd is up and running and working fine.
> Normally I don't worry about it, but on those occasions where dhcpd does
> stop working, it's hard to tell if it's "fixed" or not.
>
> What makes rc-status think something is crashed, and how can I fix this?
>

Not to sidetrack the discussion, but this is one of the areas where
systemd does shine, in a sense.  If systemd thinks the service is
down, it is definitely down, because by default when systemd thinks a
service is down it kills anything it ever spawned (and it can
auto-restart if configured to do so).  So, this forces to you
configure the unit correctly so that you don't have these kinds of
maybe-running-maybe-not situations.

It is a bit like having strong types and stricter build-time error
checking.  It makes it a little harder to be lazier but saves you as
the user from the lazy developer.

As far as openrc goes, I suspect it is a pid file issue of some kind.
If a process goes and forks without putting the right pid in the file
then there is no way for openrc to detect this.

-- 
Rich

Re: [gentoo-user] [WAY OT] wanna learn networking internals

[J. Roeleveld]

On Thursday, September 03, 2015 10:47:45 PM Alan McKinnon wrote:
> The absolute best network RFCs are the ones about coffeepot over HTTP,
> and IP by carrier pigoen (or is it Avain IP? something like that)

Which, if I'm not mistaken, was actually proven to provide a higher bandwidth 
than wired internet available between offices in a big city in South Africa? :)
(Can't find the link quickly)

You just have a higher latency and in some regions a higher chance of packet-
loss ;)

--
Joost

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alan McKinnon]

On 03/09/2015 15:06, Alec Ten Harmsel wrote:
> On Wed, Sep 02, 2015 at 02:25:47PM -0600, the...@sys-concept.com wrote:
>> I've tried to post a "log" file to the http://pastebin.com/ you
>> suggested but I can only paste the limited size file (not upload it).
>> Since the txt file is 7.4Mb in size, I can not paste it.
>> Though, I have compress the file as tar.gz (so it is only 267kB)
>> here is the link:
>>
>> http://www.sysconcept.ca/audacity_error.tar.gz
>>
>> If somebody whats to look at it, I appreciate it.
>> Meanwhile, I'll be following other folks suggestion and see if I can get
>> a positive result to this error.
>>
>> Thelma
> 
> I tried to download it but got a 403 Forbidden HTTP error.
> 
> This might be a bad suggestion, but if you have a lot of time, it might
> be easier to uninstall audacity and any other packages that have been
> giving you problems, then `emerge -uDN --with-bdeps=y @world`, then
> `emerge --depclean', and then install all the stuff you need.
> 
> I believe someone else mentioned checking /var/lib/portage/world and
> making sure that it doesn't contain any virtuals - it might also be good
> to remove from it any software that you do not directly need.


Even better - Thelma should just post the entire /var/lib/portage/world
file so we can advise what to take out.

Most newbies clutter up their world needlessly, it takes a bit of
practice to grok what should be in their


-- 
Alan McKinnon
alan.mckin...@gmail.com

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alec Ten Harmsel]

On Wed, Sep 02, 2015 at 08:52:22PM +0200, Andreas K. Huettel wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA512
> 
> On Wednesday 02 September 2015 03:22:44 Alec Ten Harmsel wrote:
> 
> > > This kind of language does not belong here. If you can not refrain
> > > yourself making sarcastic remarks don't make any comments.
> > 
> > This language occasionally belongs on this list.
> > 
> 
> No. 
> 
> I would like to respectfully point you to the Gentoo Code of Conduct:
> https://wiki.gentoo.org/wiki/Project:Council/Code_of_conduct

Interesting, I have never seen that before. Is it possible to get that
linked to on the mailing list page[1]?

Alec

[1] https://gentoo.org/get-involved/mailing-lists/

Re: [gentoo-user] audacity-2.0.5 failed to compile

[Alec Ten Harmsel]

On Wed, Sep 02, 2015 at 02:25:47PM -0600, the...@sys-concept.com wrote:
> I've tried to post a "log" file to the http://pastebin.com/ you
> suggested but I can only paste the limited size file (not upload it).
> Since the txt file is 7.4Mb in size, I can not paste it.
> Though, I have compress the file as tar.gz (so it is only 267kB)
> here is the link:
> 
> http://www.sysconcept.ca/audacity_error.tar.gz
> 
> If somebody whats to look at it, I appreciate it.
> Meanwhile, I'll be following other folks suggestion and see if I can get
> a positive result to this error.
> 
> Thelma

I tried to download it but got a 403 Forbidden HTTP error.

This might be a bad suggestion, but if you have a lot of time, it might
be easier to uninstall audacity and any other packages that have been
giving you problems, then `emerge -uDN --with-bdeps=y @world`, then
`emerge --depclean', and then install all the stuff you need.

I believe someone else mentioned checking /var/lib/portage/world and
making sure that it doesn't contain any virtuals - it might also be good
to remove from it any software that you do not directly need.

Alec

Re: [gentoo-user] dhcpd always shows "crashed" even though it's running

[Dale]

Fernando Rodriguez wrote:
> On Thursday, September 03, 2015 8:47:34 PM Dale wrote:
>
> I don't know if this will help or not but don't forget the zap option. 
>
> root@fireball / # /etc/init.d/dhcpcd  
> broken   ineediuse needsme  pauserestart  startstatus  
> stop usesme   zap 
> root@fireball / # /etc/init.d/dhcpcd 
>
> It's been a long time since I used it but if I recall correctly that
> resets the status.  I think it stops and deletes any files that stores
> its run status. 
>
> If that doesn't apply, just ignore me.  Heck, a lot of people ignore
> me.  lol 
>
> Dale
>
> :-)  :-) 
>
> It does. I knew about it but never used it and didn't know what it'll do if 
> the permissions are wrong so I thought I'd minimize the chances of being 
> wrong.
>

I wasn't sure if it would help or not.  I know it has worked in the past
for me but it has been while.  I usually tell the service to stop, make
sure any processes are dead, with kill command if needed, and then use
the zap thing.  Maybe try that as a last resort if nothing else. 

I just didn't want to not mention it and it turn out to be just what was
needed.  ;-)

Dale

:-)  :-)

Re: [gentoo-user] dhcpd always shows "crashed" even though it's running

[Dale]

Fernando Rodriguez wrote:
> On Thursday, September 03, 2015 8:09:02 PM Mike Edenfield wrote:
>> For some reason, whenever I check the status of my startup scripts, 
>> dhcpd registers as "crashed". However, dhcpd is up and running and 
>> working fine. Normally I don't worry about it, but on those occasions 
>> where dhcpd does stop working, it's hard to tell if it's "fixed" or not.
>>
>> What makes rc-status think something is crashed, and how can I fix this?
>>
>> basement log # rc-status -v | grep crashed
>>   dhcpd [  crashed  ]
>> basement log # ps aux | grep dhcpd
>> root  2214  0.0  0.0   8268   876 pts/0S+   19:47   0:00 grep 
>> --colour=auto dhcpd
>> dhcp  2648  0.0  0.6  30028 12136 ?Ss   Aug29   0:00 
>> /usr/sbin/dhcpd -cf /etc/dhcp/dhcpd.conf -q -pf /var/run/dhcp/dhcpd.pid 
>> -lf /var/lib/dhcp/dhcpd.leases -user dhcp -group dhcp -chroot 
>> /chroot/dhcp enp0s7
>>
>>
> This is just a guess but it could be the permissions on the pid file on 
> /chroot/dhcp/var/run/dhcp/. So stop the daemon, delete the file, check that 
> the 
> directory is owned by dhcp:dhcp and start the daemon again.
>


I don't know if this will help or not but don't forget the zap option. 

root@fireball / # /etc/init.d/dhcpcd  
broken   ineediuse needsme  pauserestart  startstatus  
stop usesme   zap 
root@fireball / # /etc/init.d/dhcpcd 

It's been a long time since I used it but if I recall correctly that
resets the status.  I think it stops and deletes any files that stores
its run status. 

If that doesn't apply, just ignore me.  Heck, a lot of people ignore
me.  lol 

Dale

:-)  :-) 

Re: [gentoo-user] dhcpd always shows "crashed" even though it's running

[Fernando Rodriguez]

On Thursday, September 03, 2015 8:47:34 PM Dale wrote:
> Fernando Rodriguez wrote:
> > On Thursday, September 03, 2015 8:09:02 PM Mike Edenfield wrote:
> >> For some reason, whenever I check the status of my startup scripts, 
> >> dhcpd registers as "crashed". However, dhcpd is up and running and 
> >> working fine. Normally I don't worry about it, but on those occasions 
> >> where dhcpd does stop working, it's hard to tell if it's "fixed" or not.
> >>
> >> What makes rc-status think something is crashed, and how can I fix this?
> >>
> >> basement log # rc-status -v | grep crashed
> >>   dhcpd [  crashed  ]
> >> basement log # ps aux | grep dhcpd
> >> root  2214  0.0  0.0   8268   876 pts/0S+   19:47   0:00 grep 
> >> --colour=auto dhcpd
> >> dhcp  2648  0.0  0.6  30028 12136 ?Ss   Aug29   0:00 
> >> /usr/sbin/dhcpd -cf /etc/dhcp/dhcpd.conf -q -pf /var/run/dhcp/dhcpd.pid 
> >> -lf /var/lib/dhcp/dhcpd.leases -user dhcp -group dhcp -chroot 
> >> /chroot/dhcp enp0s7
> >>
> >>
> > This is just a guess but it could be the permissions on the pid file on 
> > /chroot/dhcp/var/run/dhcp/. So stop the daemon, delete the file, check that 
the 
> > directory is owned by dhcp:dhcp and start the daemon again.
> >
> 
> 
> I don't know if this will help or not but don't forget the zap option. 
> 
> root@fireball / # /etc/init.d/dhcpcd  
> broken   ineediuse needsme  pauserestart  startstatus  
> stop usesme   zap 
> root@fireball / # /etc/init.d/dhcpcd 
> 
> It's been a long time since I used it but if I recall correctly that
> resets the status.  I think it stops and deletes any files that stores
> its run status. 
> 
> If that doesn't apply, just ignore me.  Heck, a lot of people ignore
> me.  lol 
> 
> Dale
> 
> :-)  :-) 
> 

It does. I knew about it but never used it and didn't know what it'll do if 
the permissions are wrong so I thought I'd minimize the chances of being 
wrong.

-- 
Fernando Rodriguez