Re: Deleting i386 packages

[Chris Bannister]

Please don't top post on the debian-users mailing list

On Tue, Sep 29, 2015 at 07:33:52PM +0300, Eliezer Croitoru wrote:
> Hey Chris,
> 
> It doesn't matter if some would like them to just vanish.
> They do commit to the client but the scale of things might not be understood
> by all in the same level\manner.
> MS doesn't and cannot commit to software maintenance in certain levels.
> I do not know how much money they have and indeed if they are committed to
> PROFIT only with nothing else it would make them something else then what
> they are.

I do know that:

1) They have (had?) a considerable amount of money put aside for
lawsuits against their business practices and their borg like behaviour
to small business where they liked a product of theirs.

2) A story I heard when Windows 95 was launched was they asked REM if
they could use their song 'Man On The Moon' but REM told them to shove
it in no uncertain terms, so they asked The Rolling Stones if they could
use their song "Start Me Up" and they said sure, for some exhorbitant
price, and apparently Microsoft didn't batter an eye, and just said
sure.

although this link:

http://www.networkworld.com/article/2220097/data-center/what-microsoft-paid-the-stones-to-help-launch-windows-95.html

seems to suggest it was way less than the amount I heard it was.

There was also some web pages written by an (ex?) employee of MS on why
he hated MS, (unfortunately, they seemed to buried in the huge number of
search results.) which were very interesting, there was about 9 pages
altogether, each one on a specific area. One in particular was about the
shoddy code. I wish I still had the link. :(

But the main thing I dislike about the MS OS, is the 'reverse logic'
they use in their menu selections and dialog boxes --- who knows it may
have been improved by now, but I've since moved on and have no real urge
to try it again. (once bitten twice shy, etc.)

I recently had a friend come round to use my wifi because her Windows
8.1 laptop, for some reason no longer thought that she was the owner of
it and had shut her out, she had to go through a laborious session of
sending emails and typing in codes she rec on her cell phone into boxes
on the laptop, and believe me; she was getting p***ed off fast. I
actually started to feel sorry for her when the dialog box was asking
for the six character code, but she had been only sent a four character
code. I mean how ridiculous is that? (considering she'd paid good money
for it.)

Nope, I steer clear of MS as much as possible, and just about any time
I've had to use it, I feel mentally drained and frustrated. At least if
a problem crops up in Linux, there's enough reasonable documentation
around so that you can troubleshoot your way through it without pulling
your hair out in the process.

P.S. I wonder if that is why Linux users tend to be hairy and MS users
tend to bald? :)

-- 
"If you're not careful, the newspapers will have you hating the people
who are being oppressed, and loving the people who are doing the 
oppressing." --- Malcolm X

Re: Deleting i386 packages

[Eliezer Croitoru]

On 30/09/2015 14:53, Chris Bannister wrote:


Please don't top post on the debian-users mailing list


It was unintentional.

My main point stays.
An admin and IT manager needs to evaluate their goals and decide on the 
right approach.
Sometimes it can be frustrating to navigate between the drops and decide 
that a product that will survive for 3 years worth it.


I am working with both linux and windows, open-source and non-open 
source software.
I myself wrote a free software(3 clause BSD) but I have not published 
the actual code, and instead I have published the algorithm pesudo.(if 
you have the skills it's like copy and paste)
I have proved that my code is much more efficient then other software 
and which my software(SquidBlocker) gives more then others but still I 
have not seen a dime of it from anyone.

Not even one human contacted me and said "I want to try it".
And more then just that, admins prefer to use a very old product and 
which they say "it just works" when there is no support what so ever to 
the product and it doesn't utilize the resources of the system well 
leaving aside this product doesn't meet basic production systems goals.
And it's amazing that admins still prefer to use a software that was 
designed to be used in a very low load systems on a very high load 
system vs a system that was designed for 99% uptime and very high load.


So well, it's their choice and they will decide whatever they want.

All The Bests,
Eliezer


On Tue, Sep 29, 2015 at 07:33:52PM +0300, Eliezer Croitoru wrote:

Hey Chris,

It doesn't matter if some would like them to just vanish.
They do commit to the client but the scale of things might not be understood
by all in the same level\manner.
MS doesn't and cannot commit to software maintenance in certain levels.
I do not know how much money they have and indeed if they are committed to
PROFIT only with nothing else it would make them something else then what
they are.


I do know that:

1) They have (had?) a considerable amount of money put aside for
lawsuits against their business practices and their borg like behaviour
to small business where they liked a product of theirs.

Re: Deleting i386 packages

[Eliezer Croitoru]

Hey Chris,

It doesn't matter if some would like them to just vanish.
They do commit to the client but the scale of things might not be 
understood by all in the same level\manner.

MS doesn't and cannot commit to software maintenance in certain levels.
I do not know how much money they have and indeed if they are committed 
to PROFIT only with nothing else it would make them something else then 
what they are.
I do not need google to understand some fundamentals of business and 
basic software complexity assessment.
As I mentioned before, a complex software can be pretty hard to design 
and maintain.
One of the reasons RedHat and Suse are being used on many critical 
systems is due to the commitment that the vendor gives to supply support 
and updates for a very long period of time and as a result there is a 
feature freeze period.
If you will require services that are not available in the contract they 
will answer you all sort of things.
Sometimes they answer that they cannot do a thing about it and in other 
times they patch the software by themselves.
It is not far from MS but MS just does things in another way then it is 
being done on RedHat or Suse or others(externally).


As a side note, as much as I like google.. it's missing something and it 
will always be missing something.
It cannot replace the basic human understanding of life. With all it's 
CPUs, disks, applications, DBs, statistics etc it cannot and can never 
be more then a sum of about 64bits operations(yes yes maybe later it 
would be more then that).
If someone would ask you "do you want to see the world in 64 bits?" what 
would you prefer? a 64bits based world or a real world?


And no I have not tried to write a kernel or a gui or something similar 
since I think I would be the worst choice to make these and even if just 
to match msdos 6.2 with windows 3.11 capabilities.
And yes google gives me the option to run a search query like "download 
freedos" these days but still this is about it. It cannot give 
affection, warmth and many other things which people who work in MS 
tries to give everyday either to their surrounding MS employes or the 
users. And yes in a weird way which is being presented by a 64bit CPU.


Google, Linux, MS and others just do the same thing but in different shapes.

Eliezer

* I cannot count the number of people who hate google more then MS..

On 29/09/2015 07:51, Chris Bannister wrote:

On Sun, Sep 27, 2015 at 01:21:12PM +0300, Eliezer Croitoru wrote:

>Hey Martin,
>
>I was reading your note and it is not the reality or something that should
>be done but rather another side to consider when working with software
>vendors.
>I do agree that there is a benefit when the sources are open but companies
>like MS(just as an example) do not just vanish.

Some wish they would 'just vanish.'


>The above would be said for many other vendors that are committed to the
>client to support him.

They're not committed to the client, they're committed to PROFIT. It
doesn't take much googling to see some of the abysmal practices MS have
done over the years to see where their heart is.

Re: Deleting i386 packages

[Stuart Longland]

On 23/09/15 00:14, Reco wrote:
> $ dpkg -I teamviewer_10.0.46203_amd64.deb | grep Depe
>  Depends: bash (>= 3.0), libc6-i386 (>= 2.4), lib32asound2, lib32z1,
> libxext6, ia32-libs
> 
> A fine example of non-multiarch package which declared amd64 arch while
> providing i386 binaries only.

It was probably done that way before Debian became true multi-arch.

We had a similar thing in Gentoo, if you did `emerge skype`, that
required 32-bit libraries, that on the "AMD64" architecture, were
provided by packages in the app-emulation package.

Debian (and Gentoo) have since moved to true multi-arch, and so such
kludges are no longer required.  However there are still probably
packages floating around that still rely upon them.
-- 
Stuart Longland (aka Redhatter, VK4MSL)

I haven't lost my mind...
  ...it's backed up on a tape somewhere.



signature.asc
Description: OpenPGP digital signature

Re: Deleting i386 packages

[Chris Bannister]

On Sun, Sep 27, 2015 at 01:21:12PM +0300, Eliezer Croitoru wrote:
> Hey Martin,
> 
> I was reading your note and it is not the reality or something that should
> be done but rather another side to consider when working with software
> vendors.
> I do agree that there is a benefit when the sources are open but companies
> like MS(just as an example) do not just vanish.

Some wish they would 'just vanish.'

> The above would be said for many other vendors that are committed to the
> client to support him.

They're not committed to the client, they're committed to PROFIT. It
doesn't take much googling to see some of the abysmal practices MS have
done over the years to see where their heart is.

-- 
"If you're not careful, the newspapers will have you hating the people
who are being oppressed, and loving the people who are doing the 
oppressing." --- Malcolm X

Re: Deleting i386 packages

[Martin Read]

On 27/09/15 08:06, Eliezer Croitoru wrote:

Like any other job the programmers need money and software authors are
not obligated  to publish their work to be available to all humanity(or
at-least these parts of humanity that are connected to the WWW).
The above is something I think is right and it is right especially for
security and health related software.


Security-related software is very *precisely* the kind that should not 
be closed-source proprietary software, because when your security 
software is proprietary, only the copyright holder has the right to 
publish and distribute a fix for that piece of software when it turns 
out to have a vulnerability.


And, of course, on an Internet-connected computer *most* software turns 
out to be security-related.

Re: Deleting i386 packages

[Eliezer Croitoru]

Hey Reco,

I must admit that this is not the first time I was confused as a 
trolling creature.


And responding to the above mentioned arguments\ideas\thoughts.
I know some might disagree with me about my point of view and I do not 
have any obligations to change my mind but I can clarify my thoughts.


You have asked if I am a windows user and the answer is that I do use 
windows and I find it a very nice piece of software.
But I will need to clarify couple things since I am almost sure you 
misinterpreted what I was writing.


There are couple things to consider with software.
Like any other job the programmers need money and software authors are 
not obligated  to publish their work to be available to all humanity(or 
at-least these parts of humanity that are connected to the WWW).
The above is something I think is right and it is right especially for 
security and health related software.
As opposed to what some think that software should be available for all 
I am on the side which thinks that secrecy or confidentiality is a value 
which is either required or wanted by people.


So the above doesn't implies that anyone should use any software. And 
also in any case of software usage there is a great need to consider the 
pros and cos and to see if it matches the requirements and needs.
I am not writing and talking about debian specifically since this is not 
the question(at-least from my side of the glass).


If some vendors supply compiled software that was audited by their 
programmers, QA team and security personnel it is OK for me.
If I pay for the software then it's fine from my side to get a packaged 
product.
I had a talk with a friend about the dangerous things on the Internet 
and the conclusion was that some might not understand that the Internet 
is just a "reflection" of the real world and there is no magic there.

The same crooks can be found both in the real world and the Internet.
In a case that a software vendor is suspected to be violating basic 
security requirements intentionally it would black list the name brand 
and the software.
If we are talking about a complex piece of software then it is possible 
that flaws do exist and it is also applies to any Linux and open source 
software the same way(statistically) as non open source.


Since I do have some experience with health care related programming 
subjects and I do also have couple medical facilities that runs a 
software with critical code I wrote and designed I can give a scenario.
When a sysadmin decides on what software to use for these mission 
critical human life related systems he needs to fully understand the 
weight of using software(open source and non open source, free or non-free).
He needs to be able to run a command such as "apt-get install squid" 
without any fear that human life's would be affected by it.
It means that he will have someone that he can trust to test it 
externally and verify it fits the purpose or that the software was 
tested enough by the developers and by the distribution team.
Free and non Free, open source or non open source is not the question at 
all in this case.
For some, when looking at a non-free software in banking or health care 
the question is not if it is more dangerous since the main question and 
almost only question is "is this piece of software fit for this role I 
requrie?" and it contains kind of "recursively" security and operations 
sides.


The only dangerous software is a "dangerous" software!
The definition of dangerous is not by free or non free, open source or 
none open source.
It's a subject by itself that requires a new definition in any software 
adoption steps.
Would a bug in exim, a fatal one, makes exim a dangerous piece of 
software? The answer is that in some cases it would indeed do that but 
in other cases it would be dangerous the same way as postfix or MS exchange.


You can have arguments about a specific vendor\provider\programmer 
software or a repository based on it's characteristics and statistically 
it might fit debian non-free repo but it cannot make all non-free 
software out-there to be dangerous and it's not an argument from debian 
non-free to other software or from other software to debian non-free.


A part of the health care facility sysadmin duties is to make sure that 
he can maintain the software or at-least to make sure that there are 
out-there developers, testers and others who can answer the facility 
requirements.
The above is one of the main reasons that many sysadmins prefer to use 
RedHat and Windows despite the fact that both companies cannot always be 
aware of very critical bugs.

This is not what makes these companies and their software dangerous.

If you do feel that this is what makes software dangerous indeed it's an 
argument but it might not meet reality or reality requirements in many 
facilities.


Then, what do you mean by dangerous? it was not really clear from your 
words.


Thanks,
Eliezer

On 25/09/2015 15:26, 

Re: Deleting i386 packages

[Eliezer Croitoru]

Hey Martin,

I was reading your note and it is not the reality or something that 
should be done but rather another side to consider when working with 
software vendors.
I do agree that there is a benefit when the sources are open but 
companies like MS(just as an example) do not just vanish.
The above would be said for many other vendors that are committed to the 
client to support him.

I took an example and I stick with it:
The sysadmin and IT department needs to consider and evaluate what is 
their relationship with the software vendor and decide.
Sometimes they decide to open the source but only in-house due to the 
demand but it is still unrelated to "dangerous".
I agree that if we do not trust the vendor to test and patch it's 
software then it is a risk and when the sysadmin types "apt-get install 
tzdata" then he should understand that it will be updated.. and if not 
he(or somebody) can compile and update the tzdata files.


I have seen more then once that an open source distribution did not 
updated critical updates and admins was required to run some errands to 
make the software work.
I took tzdata package since it was a very major issue on many systems I 
have seen.


I still think that an institute small enough to not build it's own OS 
can asses it's requirements and decide that for example Debian is not 
for them and they prefer a specific vendor.
It's not dangerous and not reckless but a decision which considers what 
is good for the institute from couple aspects.

Many admins feels safe enough with Windows and not with Debian.
I have couple servers and desktops and I have seen bugs that was not 
fixed in Debian and the effort it will take from me to fix them will be 
more then to buy an Hyper-v or Vmware license.
So what if they are the only ones that can patch the software? they meet 
the institute global goals with a good price. is it that bad? no!


I remember that some admin I met showed me what he did to disable the 
apache server version advertisement.
Will it secure the service against some attacks? no, but F5, RADWARE and 
other companies products will indeed do that and in some cases it's 
cheaper then patching or upgrading a running system.


So still the argument that it's dangerous is not really an argument.
The state stays exactly the same: there is a risk that needs to be 
assessed and evaluated like in any software product and like any other 
chair in the planet.


All The Bests,
Eliezer

On 27/09/2015 11:47, Martin Read wrote:

On 27/09/15 08:06, Eliezer Croitoru wrote:

Like any other job the programmers need money and software authors are
not obligated  to publish their work to be available to all humanity(or
at-least these parts of humanity that are connected to the WWW).
The above is something I think is right and it is right especially for
security and health related software.


Security-related software is very *precisely* the kind that should not
be closed-source proprietary software, because when your security
software is proprietary, only the copyright holder has the right to
publish and distribute a fix for that piece of software when it turns
out to have a vulnerability.

And, of course, on an Internet-connected computer *most* software turns
out to be security-related.

Re: [OT] Free software vs non-free, here we go again (was: Deleting i386 packages)

[Reco]

 Hi.

On Sun, 27 Sep 2015 10:06:37 +0300
Eliezer Croitoru  wrote:

> Hey Reco,
> 
> I must admit that this is not the first time I was confused as a 
> trolling creature.

For the record - I did not 'confuse' you as a troll and did not call
you one. I could not care less about it, actually.


> And responding to the above mentioned arguments\ideas\thoughts.
> I know some might disagree with me about my point of view and I do not 
> have any obligations to change my mind but I can clarify my thoughts.
> 
> You have asked if I am a windows user and the answer is that I do use 
> windows and I find it a very nice piece of software.

Dear Eliezer. I don't *need* to ask you whenever you're using Windows
or not. E-mail headers 'User-Agent' and 'Content-Type' from your e-mails
leave me absolutely no doubt about you use Windows.

Also I must apologize in advance if I mistook your last name for the
first one. No offense meant.


> But I will need to clarify couple things since I am almost sure you 
> misinterpreted what I was writing.
> 
> There are couple things to consider with software.
> Like any other job the programmers need money and software authors are 
> not obligated  to publish their work to be available to all humanity(or 
> at-least these parts of humanity that are connected to the WWW).

True. To reduce free software concept to the availability of the source
code is gross oversimplification (and misses the point almost
completely btw), but we'll leave it here for a moment.


> The above is something I think is right and it is right especially for 
> security and health related software.

False. Please take your time to research 'Bugtraq' and
'Full-Disclosure' maillists to observe the falsehood of this statement.
Security by obscurity never works, and to promote such point of view
means to deny the truth.


> As opposed to what some think that software should be available for all 
> I am on the side which thinks that secrecy or confidentiality is a value 
> which is either required or wanted by people.

Here you've got it all wrong. You do not attach 'confidentiality' (let
along 'secrecy') labels to the software itself. You need to attach said
labels to the data which the software in question deals with.


> So the above doesn't implies that anyone should use any software. And 
> also in any case of software usage there is a great need to consider the 
> pros and cos and to see if it matches the requirements and needs.
> I am not writing and talking about debian specifically since this is not 
> the question(at-least from my side of the glass).
> 
> If some vendors supply compiled software that was audited by their 
> programmers, QA team and security personnel it is OK for me.

... And in real life that's not enough at all. Internal security audit
is a must, I agree. But unless the software in question is only used
internally by completely trusted personel - good, secure software
should be audited externally.
Inability to access the source code of said software is a serious
disadvantage in such a case.
Inability to *change* the source of said software equals inability to
use security-fixed software for large amounts of time (can be
sometimes remedied with assorted klugdes, true).
Inability to *deploy* the changed software (aka tivoized software) means
the same.


> If I pay for the software then it's fine from my side to get a packaged 
> product.

Ok. No arguments here. Note that free software does not equals zero
cost or lack of said 'packaged product'. Case study - Red Hat.


> I had a talk with a friend about the dangerous things on the Internet 
> and the conclusion was that some might not understand that the Internet 
> is just a "reflection" of the real world and there is no magic there.
> The same crooks can be found both in the real world and the Internet.

Ok. No arguments here too.


> In a case that a software vendor is suspected to be violating basic 
> security requirements intentionally it would black list the name brand 
> and the software.

False. There are numerous examples of the contrary. Starting with the
Microsoft, but not ending here.


> If we are talking about a complex piece of software then it is possible 
> that flaws do exist and it is also applies to any Linux and open source 
> software the same way(statistically) as non open source.

Oh, do I see 'They are bad too' argument here, right?


> Since I do have some experience with health care related programming 
> subjects and I do also have couple medical facilities that runs a 
> software with critical code I wrote and designed I can give a scenario.
> When a sysadmin decides on what software to use for these mission 
> critical human life related systems he needs to fully understand the 
> weight of using software(open source and non open source, free or non-free).
> He needs to be able to run a command such as "apt-get install squid" 
> without any fear that human life's would be affected by it.

Re: Deleting i386 packages

[Reco]

On Fri, Sep 25, 2015 at 09:54:05PM +1200, Chris Bannister wrote:
> On Thu, Sep 24, 2015 at 01:24:45PM +0300, Reco wrote:
> > Hi.
> > 
> > On Thu, Sep 24, 2015 at 07:36:32AM +0300, Eliezer Croitoru wrote:
> > > I will not argue since truth can be seen from more then one side.
> > > Proprietary software usage is normal in all cases.
> > 
> > No surprise in such position here, since apparently you're using Windows.
> > And you came to the wrong place to promote such view of things.
> 
> I believe you have just been trolled! :)

Let's recount.

Number of personal attacks - 0.
Number of naughty words - 0.
Number of off-list e-mails with kill threats/suicide propositions - 0.
Number of doxxings - 0.
Number of kittens harmed - 0.
Windoze users fended off - 1.

Nope, I have to disagree with your assertion :)

Therefore, my point stands unchallenged - non-free software users should
suffer anyway.

Reco

Re: Deleting i386 packages

[Chris Bannister]

On Thu, Sep 24, 2015 at 01:24:45PM +0300, Reco wrote:
>   Hi.
> 
> On Thu, Sep 24, 2015 at 07:36:32AM +0300, Eliezer Croitoru wrote:
> > I will not argue since truth can be seen from more then one side.
> > Proprietary software usage is normal in all cases.
> 
> No surprise in such position here, since apparently you're using Windows.
> And you came to the wrong place to promote such view of things.

I believe you have just been trolled! :)

-- 
"If you're not careful, the newspapers will have you hating the people
who are being oppressed, and loving the people who are doing the 
oppressing." --- Malcolm X

Re: Deleting i386 packages

[Reco]

Hi.

On Thu, Sep 24, 2015 at 07:36:32AM +0300, Eliezer Croitoru wrote:
> I will not argue since truth can be seen from more then one side.
> Proprietary software usage is normal in all cases.

No surprise in such position here, since apparently you're using Windows.
And you came to the wrong place to promote such view of things.

But I disagree with you. It boils down to the question of control. As
long as you cannot change the software in question if needed - you're
not in control. They do.

As for 'all cases':

Is it OK to do banking using non-free software?
Is it OK to trust healthcare of personal information to non-free
software?


> It is as dangerous as the usage of open source software.

No it's more dangerous. See above about control.


> It might limit but it gives something that not all open source software can
> give.

Which is? What's that magical thing?


> It doesn't limit freedom but just a mere functionality "limit" like any
> other software that exists and will exist ever.

No. Existence or absence of functionality is the question of change.
With the free software you can make the needed change.
With the non-free software you cannot.


> Software is Software and if someone requires something it doesn't matter if
> it fits the needs.

Nope. Non-free software usually comes with restrictions in field of
usage or other completely artificial restrictions.

Reco

Re: Deleting i386 packages

[Eliezer Croitoru]

I will not argue since truth can be seen from more then one side.
Proprietary software usage is normal in all cases.
It is as dangerous as the usage of open source software.
It might limit but it gives something that not all open source software 
can give.
It doesn't limit freedom but just a mere functionality "limit" like any 
other software that exists and will exist ever.
Software is Software and if someone requires something it doesn't matter 
if it fits the needs.


All The Bests,
Eliezer

On 22/09/2015 17:01, Reco wrote:

Using proprietary software is not normal. It's dangerous. It limits
one's freedom. And IMO it definitely should not be encouraged on
*Debian*  user maillist.

Reco

Re: Deleting i386 packages

[Reco]

Hi.

On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:
> For historical reasons, my x86-64 architecture computers have a large
> number of i386 packages on them that I'd just as soon be rid of.  is
> there a good way to simply tell a package manager that I want everything
> involving that architecture deleted?  The best answer I've found on my
> own has been to use dpkg and grep to find everything with :i386, and
> then construct a huge dpkg --purge command to get rid of them all.
> Hoping for something a little simpler...

Try this:

apt-get remove --purge .*:i386

Reco

Re: Deleting i386 packages

[mudongliang]

On 09/22/2015 02:49 PM, Reco wrote:

Hi.

On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:

For historical reasons, my x86-64 architecture computers have a large
number of i386 packages on them that I'd just as soon be rid of.  is
there a good way to simply tell a package manager that I want everything
involving that architecture deleted?  The best answer I've found on my
own has been to use dpkg and grep to find everything with :i386, and
then construct a huge dpkg --purge command to get rid of them all.
Hoping for something a little simpler...

Try this:

apt-get remove --purge .*:i386
Before you do this dangerous command , please check all your 64bit 
software and make sure they don't need any i386 package.

Or some of 64bit software will not work if you don't check.
For example , skype ,teamviewer all need i386 packages.
- mudongliang


Reco

Re: Deleting i386 packages

[Reco]

 Hi.

On Tue, Sep 22, 2015 at 09:06:47PM +0800, mudongliang wrote:
> Or some of 64bit software will not work if you don't check.
> For example , skype ,teamviewer all need i386 packages.
> 
> 1) Users of non-free software (especially users of non-free wine-embedded
> software) should suffer anyway.
> 
> 2) Which part of teamviewer is 64bit?
> 
> Its webiste shows 32-Bit / 64-Bit Multiarch! If you install multiarch, the 
> dpkg shows you :
> 
> $ dpkg -l | grep teamviewer
> ii  teamviewer  10.0.41499
>   amd64    TeamViewer (Remote Control Application)

$ dpkg -I teamviewer_10.0.46203_amd64.deb | grep Depe
 Depends: bash (>= 3.0), libc6-i386 (>= 2.4), lib32asound2, lib32z1,
libxext6, ia32-libs

A fine example of non-multiarch package which declared amd64 arch while
providing i386 binaries only.

Also it won't be affected by removal of i386 packages as it depends on
amd64 packages only.

Of course, they may have changed it from version 10.0.41499 I've just
downloaded, but I doubt it.


> And skype :
> 
> $ dpkg -l | grep skype
> ii  skype    4.3.0.37-0ubuntu0.12.04.1
>     amd64    client for Skype VOIP and instant messaging service
> ii  skype-bin    4.3.0.37-0ubuntu0.12.04.1
>     i386 client for Skype VOIP and instant messaging service - 
> binary
> files

In this case I agree that 'apt-get remove .*:i386' should get rid of
skype-bin. Which will be clearly expressed by apt, giving the user a
chance to bail out.

So again, no custom checks are needed.

Reco

Re: Deleting i386 packages

[Reco]

On Tue, Sep 22, 2015 at 03:43:48PM +0800, mudongliang wrote:
> 
> 
> On 09/22/2015 02:49 PM, Reco wrote:
> > Hi.
> >
> >On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:
> >>For historical reasons, my x86-64 architecture computers have a large
> >>number of i386 packages on them that I'd just as soon be rid of.  is
> >>there a good way to simply tell a package manager that I want everything
> >>involving that architecture deleted?  The best answer I've found on my
> >>own has been to use dpkg and grep to find everything with :i386, and
> >>then construct a huge dpkg --purge command to get rid of them all.
> >>Hoping for something a little simpler...

> >Try this:
> >
> >apt-get remove --purge .*:i386

> Before you do this dangerous command , please check all your 64bit software
> and make sure they don't need any i386 package.

No stock Debian amd64 package should require a i386 package.


> Or some of 64bit software will not work if you don't check.
> For example , skype ,teamviewer all need i386 packages.

1) Users of non-free software (especially users of non-free wine-embedded
software) should suffer anyway.

2) Which part of teamviewer is 64bit?

Reco

Re: Deleting i386 packages

[mudongliang]

On 09/22/2015 08:38 PM, Reco wrote:

On Tue, Sep 22, 2015 at 03:43:48PM +0800, mudongliang wrote:


On 09/22/2015 02:49 PM, Reco wrote:

Hi.

On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:

For historical reasons, my x86-64 architecture computers have a large
number of i386 packages on them that I'd just as soon be rid of.  is
there a good way to simply tell a package manager that I want everything
involving that architecture deleted?  The best answer I've found on my
own has been to use dpkg and grep to find everything with :i386, and
then construct a huge dpkg --purge command to get rid of them all.
Hoping for something a little simpler...

Try this:

apt-get remove --purge .*:i386

Before you do this dangerous command , please check all your 64bit software
and make sure they don't need any i386 package.

No stock Debian amd64 package should require a i386 package.

Yes ,you're right!




Or some of 64bit software will not work if you don't check.
For example , skype ,teamviewer all need i386 packages.

1) Users of non-free software (especially users of non-free wine-embedded
software) should suffer anyway.

2) Which part of teamviewer is 64bit?
Its webiste shows 32-Bit / 64-Bit Multiarch! If you install multiarch, 
the dpkg shows you :


$ dpkg -l | grep teamviewer
ii  teamviewer 10.0.41499  amd64 
TeamViewer (Remote Control Application)


And skype :

$ dpkg -l | grep skype
ii  skype 4.3.0.37-0ubuntu0.12.04.1amd64client for 
Skype VOIP and instant messaging service
ii  skype-bin 4.3.0.37-0ubuntu0.12.04.1i386 client 
for Skype VOIP and instant messaging service - binary files


- mudongliang


Reco

Re: Deleting i386 packages

[Mirko Parthey]

On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:
> For historical reasons, my x86-64 architecture computers have a large
> number of i386 packages on them that I'd just as soon be rid of.  is
> there a good way to simply tell a package manager that I want everything
> involving that architecture deleted?  The best answer I've found on my
> own has been to use dpkg and grep to find everything with :i386, and
> then construct a huge dpkg --purge command to get rid of them all.
> Hoping for something a little simpler...

# aptitude --purge-unused markauto "~ri386"

This should only remove packages that are not needed anymore,
assuming that nobody messed up the auto markings of your installed
packages.

At the yes/no prompt, you might want to check the list of packages to be
purged.
You could also run a first pass without "--purge-unused", check if you are
happy with the results, then do a second pass with this option included.

Regards,
Mirko

Re: Deleting i386 packages

[Martin Read]

On 22/09/15 13:38, Reco wrote:

1) Users of non-free software (especially users of non-free wine-embedded
software) should suffer anyway.


It speaks ill of you that you cite this as a reason for not offering 
cautionary advice to users of proprietary software.


If such people *do* in fact deserve suffering, I submit that their use 
of proprietary software will provide all the suffering they deserve, and 
that their suffering need not be compounded by the behaviour of 
advocates for libre software.

Re: Deleting i386 packages

[Reco]

On Tue, Sep 22, 2015 at 02:25:20PM +0100, Martin Read wrote:
> On 22/09/15 13:38, Reco wrote:
> >1) Users of non-free software (especially users of non-free wine-embedded
> >software) should suffer anyway.
> 
> It speaks ill of you that you cite this as a reason for not offering
> cautionary advice to users of proprietary software.

Probably. Still, if OP wishes to purify his amd64 installation by
removing unneeded i386 parts by clearly expressing his wish to do so (in
the e-mail), it seems rude to assume that those i386 parts were serving
as a mere compatibility shim to some proprietary blob.

Using i386 packages on amd64 install is not limited by this usecase.


> If such people *do* in fact deserve suffering, I submit that their use of
> proprietary software will provide all the suffering they deserve, and that
> their suffering need not be compounded by the behaviour of advocates for
> libre software.

Call it 'raising the awareness' then. Or 'zero tolerance policy'. Or
whatever newspeak term they invented for this.

Using proprietary software is not normal. It's dangerous. It limits
one's freedom. And IMO it definitely should not be encouraged on
*Debian* user maillist.

Reco

Deleting i386 packages

[Joe Pfeiffer]

For historical reasons, my x86-64 architecture computers have a large
number of i386 packages on them that I'd just as soon be rid of.  is
there a good way to simply tell a package manager that I want everything
involving that architecture deleted?  The best answer I've found on my
own has been to use dpkg and grep to find everything with :i386, and
then construct a huge dpkg --purge command to get rid of them all.
Hoping for something a little simpler...

Re: Deleting i386 packages

[Himanshu Shekhar]

You should probably avoid doing so. We are using systems based on amd64 (64
bit) architecture, still there are many applications that yet depend on the
i386 (32 bit) model. 64 bit processors allow 32 apps to run, which lets
them function properly on modern computers too.
i386 packages should not worry you much, and you may need some of them
badly, if I am not wrong.

Happy computing!

On Tue, Sep 22, 2015 at 8:48 AM, Joe Pfeiffer  wrote:

> For historical reasons, my x86-64 architecture computers have a large
> number of i386 packages on them that I'd just as soon be rid of.  is
> there a good way to simply tell a package manager that I want everything
> involving that architecture deleted?  The best answer I've found on my
> own has been to use dpkg and grep to find everything with :i386, and
> then construct a huge dpkg --purge command to get rid of them all.
> Hoping for something a little simpler...
>
>


-- 
Himanshu Shekhar
IIIT-Allahabad
IRM2015006