rsync problem

[Pinnerite]

I'm a bit new to rsync so though I have it working it is not as I expected.
Samba is my network manager.

I have two machines named ALBURY and CHURCHILL. I am copying from ALBURY to
CHURCHILL.

I use smb4k to mount the shares established on CHURCHILL. They get mounted
on /home/me/smb4k

As an example one share in smb.conf on CHURCHILL is named b_public.
Its path (on CHURCHILL) is /home/backup/public

What is being copied to b_public from ALBURY are the contents of
/home/me/public

rsync is copying it but is copying the directory as well so on CHURCHILL I
end up with /home/me/public/public.

The command is as follows:

rsync -auv --exclude-from=exclude.txt /home/public (continued on next line)
/home/alan/smb4k/CHURCHILL/b_public

How can I avoid the extra level?
I seem to be tying myself up in knots over this.


--
___________________________________________________

Mageia 2 for x86_64, Kernel: 3.4.34-desktop-1.mga2
KDE version 4.8.5 Running on an AMD 4-core processor

[Nick Leverton]

In article <kimvno$6es$2...@speranza.aioe.org>,

Pinnerite <pinn...@gmail.com> wrote:
>
>As an example one share in smb.conf on CHURCHILL is named b_public.
>Its path (on CHURCHILL) is /home/backup/public
>
>What is being copied to b_public from ALBURY are the contents of
>/home/me/public
>
>rsync is copying it but is copying the directory as well so on CHURCHILL I
>end up with /home/me/public/public.
>
>The command is as follows:
>
>rsync -auv --exclude-from=exclude.txt /home/public (continued on next line)
> /home/alan/smb4k/CHURCHILL/b_public
>
>How can I avoid the extra level?
>I seem to be tying myself up in knots over this.

Try appending / to the two directory names. That causes rsync to copy
the contents, without that it copies the directory itself.

rsync -auv --exclude-from=exclude.txt /home/public/ /home/alan/smb4k/CHURCHILL/b_public/

Incidentally are you using knode ? It has a bug (still, after what must
be about 10 years) which messes up cross posted articles by forcing followups.

Nick
--
"The Internet, a sort of ersatz counterfeit of real life"
-- Janet Street-Porter, BBC2, 19th March 1996

[Aragorn]

On Tuesday 26 March 2013 22:57, Nick Leverton conveyed the following to
uk.comp.os.linux...

> Incidentally are you using knode ? It has a bug (still, after what
> must be about 10 years) which messes up cross posted articles by
> forcing followups.

That's not a bug; that's a feature. As in "intended behavior by the
authors of KNode". ;-)

I will admit that it's annoying, because when no follow-up was set, it
will always set the first of the listed groups as the follow-up group,
and if a follow-up was already set by the previous poster, it will
/only/ send to /that/ group.

It does however have another annoying and /real/ bug, though. It will
fail to send an article when the references header reaches a certain
length. Typically this happens when there are about 20-24 references.
This bug is however specific to the KDE 4 version of KNode and may
already have been fixed in later KDE versions than mine - I'm using KDE
4.6.5 here, and I first noticed the bug in KDE 4.6.3. The bug did not
exist yet in the KDE 3 and earlier versions of KNode.

--
= Aragorn =

http://www.linuxcounter.net - registrant #223157

[alexd]

Nick Leverton (for it is he) wrote:

> Incidentally are you using knode ? It has a bug (still, after what must
> be about 10 years) which messes up cross posted articles by forcing
> followups.

Er, yeah, I sent a reply to this and for some reason, it only went to
alt.os.linux, which I don't actually read [I think Aragorn has explained
why].

Pinnerite (for it is he) wrote:

> How can I avoid the extra level?

Put slash on the end.
Whether the source or destination path
I struggle to remember.

http://defindit.com/readme_files/rsync_backup.html#toc1



--
<http://ale.cx/> (AIM:troffasky) (UnSoEs...@ale.cx)
12:40:37 up 12 days, 4:36, 5 users, load average: 5.67, 3.94, 2.25
Qua illic est reprehendit, illic est a vindicatum

[Chris Davies]

Pinnerite <pinn...@gmail.com> wrote:
> rsync -auv --exclude-from=exclude.txt /home/public \

> /home/alan/smb4k/CHURCHILL/b_public

> How can I avoid the extra level?

Add a trailing slash to the source directory.

It is mentioned in the manpage but IMO you have to read it several times
(and encounter the problem) before you can understand what it's telling
you. The best sentence is this one, « You can think of a trailing / on
a source as meaning "copy the contents of this directory" as opposed to
"copy the directory by name" »

Chris

[chris]

On 27/03/2013 12:42, alexd wrote:
> Nick Leverton (for it is he) wrote:
>
>> Incidentally are you using knode ? It has a bug (still, after what must
>> be about 10 years) which messes up cross posted articles by forcing
>> followups.
>
> Er, yeah, I sent a reply to this and for some reason, it only went to
> alt.os.linux, which I don't actually read [I think Aragorn has explained
> why].
>

Aha, that explains why I never saw my reply...

[The Real Doctor]

On 27/03/13 12:30, Chris Davies wrote:
> It is mentioned in the manpage but IMO you have to read it several times

I still have to read the man page every sodding time to remind myself,
and even then I get it wrong about 1/3 of the time.

Ian

[Tony Houghton]

On 26 Mar 2013 21:32:20 GMT
Huge <Hu...@nowhere.much.invalid> wrote:

> rsync -auv --exclude-from=exclude.txt /home/public/* \
> /home/alan/smb4k/CHURCHILL/b_public"

That would miss hidden files and folders in the top level of
/home/public/.

[Bit Twister]

Heheh, I hear that. I have an ASCII file with some working examples
and key words so I do not have to hit the man page. Example:

$ uh loc push rsync
push_current_location rsync -aAvx --delete $PWD/ $US...@mtv.home.test:$PWD

$ uh get rsync
get_any_location rsync -aAvx --delete $US...@wb.home.test:$PWD/ $PWD

$ uh rsync exa
backup cooker example mount /cooker ; mount /hotbu ; rsync -aAvx --delete /cooker/ /cauldron
restore cooker example mount /cooker ; mount /hotbu ; rsync -aAvx --delete /cauldron/ /cooker

[Nick Leverton]

In article <kiu0ur$kul$1...@dont-email.me>,

Aragorn <str...@telenet.be.invalid> wrote:
>On Tuesday 26 March 2013 22:57, Nick Leverton conveyed the following to
>uk.comp.os.linux...
>
>> Incidentally are you using knode ? It has a bug (still, after what
>> must be about 10 years) which messes up cross posted articles by
>> forcing followups.
>
>That's not a bug; that's a feature. As in "intended behavior by the
>authors of KNode". ;-)
>
>I will admit that it's annoying, because when no follow-up was set, it
>will always set the first of the listed groups as the follow-up group,
>and if a follow-up was already set by the previous poster, it will
>/only/ send to /that/ group.

It is a bug. It breaks the Good News-Keeping Seal of Approval by silently
setting default followups to something other than the original article
itself was posted to. I've seen many non-technical people, innocent knode
users, get accused of trolling, purely because knode sets followups in
this trollish way.

It also lays a boobytrap for whoever follows up, meaning that their
posting will be directed to a group they probably don't subscribe to,
as we have seen even in this thread.

And yes, I checked and it was ten years ago this coming Saturday
that this bug was first reported (KDE bug #56662). Fixes for it
have been posted two or three times to KDE Bugzilla (including by me,
bug #63872). One person even claimed to have applied a fix to the KDE
tree, although apparently not the same tree that goes into releases.
Knode is antisocial in this respect and, let's face it, seems basically
to be buggy abandonware.

>It does however have another annoying and /real/ bug, though. It will
>fail to send an article when the references header reaches a certain
>length. Typically this happens when there are about 20-24 references.
>This bug is however specific to the KDE 4 version of KNode and may
>already have been fixed in later KDE versions than mine - I'm using KDE
>4.6.5 here, and I first noticed the bug in KDE 4.6.3. The bug did not
>exist yet in the KDE 3 and earlier versions of KNode.

Ah, maybe it's not totally abandoned then if they are still adding new
bugs to it. Although I do wonder if that could an old bug but newly
being triggered by increasing References lengths these days, as over the
last few years I've found bugs in INN concerning References headers over
1024 bytes.

By the way I mended your broken followups for you. It would be polite
of you not to set them unless you warn in the posting that you (or your
newsreader) are doing so.

[c...@isbd.net]

Same as me, I feel it ought to default to the other way about, or
something.

--
Chris Green

[Aragorn]

On Wednesday 27 March 2013 21:08, Nick Leverton conveyed the following

to uk.comp.os.linux...

> In article <kiu0ur$kul$1...@dont-email.me>,
> Aragorn <str...@telenet.be.invalid> wrote:
>
>> On Tuesday 26 March 2013 22:57, Nick Leverton conveyed the following
>> to uk.comp.os.linux...
>>
>>> Incidentally are you using knode ? It has a bug (still, after what
>>> must be about 10 years) which messes up cross posted articles by
>>> forcing followups.
>>
>> That's not a bug; that's a feature. As in "intended behavior by the
>> authors of KNode". ;-)
>>
>> I will admit that it's annoying, because when no follow-up was set,
>> it will always set the first of the listed groups as the follow-up
>> group, and if a follow-up was already set by the previous poster, it
>> will /only/ send to /that/ group.
>
> It is a bug. It breaks the Good News-Keeping Seal of Approval by
> silently setting default followups to something other than the
> original article itself was posted to.

If it is intended behavior by the authors of the software, then it is
not a bug, even if it gets reported as such. And that is probably also
the very reason why the bug reports on this issue have been ignored by
the KDE/KNode developer team.

Whether the intended behavior is the correct behavior or not is a
different matter. Not everyone has the same vision on what is correct,
and some people feel that they need to push /their/ vision/agenda.

> I've seen many non-technical people, innocent knode users, get accused
> of trolling, purely because knode sets followups in this trollish way.

Like I said, I will admit to the annoyance factor, but I have never seen
anyone being accused of trolling over this. In fact, I have mentioned
this issue a few times before myself, and I was then told - by non-
KDE/non-KNode users - that this is how it should be, i.e. that follow-
ups /must/ be set and that they /must/ be observed.

Not that I agree with that stance myself, but apparently there is
definitely a whole bunch of people who feel that KNode's behavior is
correct.

> It also lays a boobytrap for whoever follows up, meaning that their
> posting will be directed to a group they probably don't subscribe to,
> as we have seen even in this thread.

Yes, I have already been bitten by that myself... :-/

> And yes, I checked and it was ten years ago this coming Saturday
> that this bug was first reported (KDE bug #56662). Fixes for it
> have been posted two or three times to KDE Bugzilla (including by me,
> bug #63872). One person even claimed to have applied a fix to the KDE
> tree, although apparently not the same tree that goes into releases.
> Knode is antisocial in this respect and, let's face it, seems
> basically to be buggy abandonware.

It is definitely not abandonware, but I will agree that the KDE
developers seem mostly interested in adding new and shiny features
rather than in fixing bugs.

>> It does however have another annoying and /real/ bug, though. It
>> will fail to send an article when the references header reaches a
>> certain length. Typically this happens when there are about 20-24
>> references. This bug is however specific to the KDE 4 version of
>> KNode and may already have been fixed in later KDE versions than mine
>> - I'm using KDE 4.6.5 here, and I first noticed the bug in KDE 4.6.3.
>> The bug did not exist yet in the KDE 3 and earlier versions of KNode.
>
> Ah, maybe it's not totally abandoned then if they are still adding new
> bugs to it. Although I do wonder if that could an old bug but newly
> being triggered by increasing References lengths these days, as over
> the last few years I've found bugs in INN concerning References
> headers over 1024 bytes.

I think that the issue with KNode here also centers around the 1024 byte
range.

> By the way I mended your broken followups for you. It would be polite
> of you not to set them unless you warn in the posting that you (or
> your newsreader) are doing so.

You are making a logical error there. Yes, "it would be polite of me
not to set them", except that _I_ am not the one setting them, and I
have to pay special attention when replying to a post because KNode sets
the follow-ups all by itself, as discussed above.

I /try/ paying attention to the follow-ups, but I can't always guarantee
that I'll notice it. I would therefore postulate that it's /impolite/
to shove the blame off on me for what my newsreader is doing without my
consent.

[Nick Leverton]

In article <kj0n7p$f44$1...@dont-email.me>,

Aragorn <str...@telenet.be.invalid> wrote:
>>
>> It is a bug. It breaks the Good News-Keeping Seal of Approval by
>> silently setting default followups to something other than the
>> original article itself was posted to.
>
>If it is intended behavior by the authors of the software, then it is
>not a bug, even if it gets reported as such. And that is probably also
>the very reason why the bug reports on this issue have been ignored by
>the KDE/KNode developer team.

That is just such a strange idea. Programmers are quite capable of
designing and planning behaviour that turns out to be incorrect and a bug.
There is nothing superhuman or omniscient about us.

I'm grateful that you did fix the followups this time, but given the
above and your apparent contention that you can't be expected to take
any responsibility for the behaviour of the software you choose to use,
even though you use it in full knowledge of that behaviour, *and* even
though the source is available to you to alter it, I think I'm going to
engage my personal First Law of Usenet and leave you to your beliefs.

[Aragorn]

On Thursday 28 March 2013 10:14, Nick Leverton conveyed the following to
uk.comp.os.linux...

> In article <kj0n7p$f44$1...@dont-email.me>,
> Aragorn <str...@telenet.be.invalid> wrote:
>
>>> It is a bug. It breaks the Good News-Keeping Seal of Approval by
>>> silently setting default followups to something other than the
>>> original article itself was posted to.
>>
>> If it is intended behavior by the authors of the software, then it is
>> not a bug, even if it gets reported as such. And that is probably
>> also the very reason why the bug reports on this issue have been
>> ignored by the KDE/KNode developer team.
>
> That is just such a strange idea. Programmers are quite capable of
> designing and planning behaviour that turns out to be incorrect and a
> bug. There is nothing superhuman or omniscient about us.

I never said that, but it /would/ appear that (at least some of) the KDE
developers feel that they may make decisions on what's good and what's
bad on account of someone else.

And it isn't just the KDE developers who are guilty of this. In fact,
they are probably the least of everyone's worries in the FLOSS
community. RedHat is constantly making decisions which anger even Linus
Torvalds and cause lots of breakage, "just because they can", which they
then justify via their propaganda machine freedesktop.org, disdainfully
mocking other distributions "because they are irrelevant and they will
simply have to follow our lead".

> I'm grateful that you did fix the followups this time, but given the
> above and your apparent contention that you can't be expected to take
> any responsibility for the behaviour of the software you choose to

> use, [...

That is not what I contended, so kindly refrain from putting words in my
mouth.

I /try/ to avoid the situation with my newsreader where it sets a
follow-up, but as much as it may shock you, I find that my body is still
human and that even my eidetic memory tends to overlook or forget
something every once in a while. So go ahead and shoot me if I
accidentally overlook the follow-ups which my newsreader, in its
politically correct fit, imposes on my posts.

> ...] even though you use it in full knowledge of that behaviour, *and*

> even though the source is available to you to alter it, I think I'm
> going to engage my personal First Law of Usenet and leave you to your
> beliefs.

Do you have any idea how pompous that sounds?

[unruh]

On 2013-03-28, Nick Leverton <ni...@leverton.org> wrote:
> In article <kj0n7p$f44$1...@dont-email.me>,
> Aragorn <str...@telenet.be.invalid> wrote:
>>>
>>> It is a bug. It breaks the Good News-Keeping Seal of Approval by
>>> silently setting default followups to something other than the
>>> original article itself was posted to.
>>
>>If it is intended behavior by the authors of the software, then it is
>>not a bug, even if it gets reported as such. And that is probably also
>>the very reason why the bug reports on this issue have been ignored by
>>the KDE/KNode developer team.
>
> That is just such a strange idea. Programmers are quite capable of
> designing and planning behaviour that turns out to be incorrect and a bug.
> There is nothing superhuman or omniscient about us.

No. BEcause of their design, the system may do things that they did not
want it to do. That is a bug. But if it does what they want, it is not a
bug. It is a feature. If it is incorrect, it may be stupid feature, or
an antisocial feature, or a non-compliance with some standard.

>
> I'm grateful that you did fix the followups this time, but given the
> above and your apparent contention that you can't be expected to take
> any responsibility for the behaviour of the software you choose to use,
> even though you use it in full knowledge of that behaviour, *and* even
> though the source is available to you to alter it, I think I'm going to
> engage my personal First Law of Usenet and leave you to your beliefs.

Excellent choice.
If his beliefs bother you enough you could even killfile him.

[Richard Kettlewell]

unruh <un...@invalid.ca> writes:

> Nick Leverton <ni...@leverton.org> wrote:
>> Aragorn <str...@telenet.be.invalid> wrote:
>>>> It is a bug. It breaks the Good News-Keeping Seal of Approval by
>>>> silently setting default followups to something other than the
>>>> original article itself was posted to.
>>>
>>>If it is intended behavior by the authors of the software, then it is
>>>not a bug, even if it gets reported as such. And that is probably also
>>>the very reason why the bug reports on this issue have been ignored by
>>>the KDE/KNode developer team.
>>
>> That is just such a strange idea. Programmers are quite capable of
>> designing and planning behaviour that turns out to be incorrect and a bug.
>> There is nothing superhuman or omniscient about us.
>
> No. BEcause of their design, the system may do things that they did not
> want it to do. That is a bug. But if it does what they want, it is not a
> bug. It is a feature. If it is incorrect, it may be stupid feature, or
> an antisocial feature, or a non-compliance with some standard.

I agree with Nick (as they say). Just because the author of a program
intended something doesn’t stop it being a bug. Sometimes programmers
are just wrong.

--
http://www.greenend.org.uk/rjk/

[Martin]

As long as you people don't agree on teh semantics of the term "bug" you
can argue until the kine come home. ;)

however, the software industry even looks at bugs that are introduced as
early as in the user story and product specification stage (ie. even
before the formal software specification, if that is ever done). Those
are particularly expensive if uncaught.

> Just because the author of a program
> intended something doesn’t stop it being a bug. Sometimes programmers
> are just wrong.

Btw software development is a holistic process and i am not aware of
anywhere in the industry where you still have "programmers" these days,
but that may be just my environment.

On another note, experience shows, the more product managers, solition
managers, solution architects, product owners and other parasites you
add to the process, the poorer the result will be. :p

[unruh]

["Followup-To:" header set to alt.os.linux.]

> intended something doesn???t stop it being a bug. Sometimes programmers
> are just wrong.

This is getting pedantic, but I would call it a bug only if the
programmer did not intend the action the program takes. To say that
sometimes the programmer is just wrong suggests the narcicistic view
that your views are right and anyone who disagrees is wrong, which to me
is a hijacking of the word "bug". I do not mind you saying that the
programmer was wrong in his design-- that makes it clear that the issue
is a difference of opinion. But to say that it is a bug, which must be
eradicated without compunction by any person able to do so, takes it out
of the arena of disagreement of design.
That Mandriva always changes /etc/inittab to make ctrl-alt-del do a
reboot rather than a halt I consider to be completely wrongheaded,
dictatorial, and causing problems, and can see no reason for,
but it is not a bug, because it is exactly what they intended.

>

[unruh]

On 2013-03-29, Martin <n...@spam.invalid> wrote:
> On 03/29/2013 09:53 AM, Richard Kettlewell wrote:
>> unruh <un...@invalid.ca> writes:
>>> Nick Leverton <ni...@leverton.org> wrote:
>>>> Aragorn <str...@telenet.be.invalid> wrote:
>>>>>> It is a bug. It breaks the Good News-Keeping Seal of Approval by
>>>>>> silently setting default followups to something other than the
>>>>>> original article itself was posted to.
>>>>>
>>>>> If it is intended behavior by the authors of the software, then it is
>>>>> not a bug, even if it gets reported as such. And that is probably also
>>>>> the very reason why the bug reports on this issue have been ignored by
>>>>> the KDE/KNode developer team.
>>>>
>>>> That is just such a strange idea. Programmers are quite capable of
>>>> designing and planning behaviour that turns out to be incorrect and a bug.
>>>> There is nothing superhuman or omniscient about us.
>>>
>>> No. BEcause of their design, the system may do things that they did not
>>> want it to do. That is a bug. But if it does what they want, it is not a
>>> bug. It is a feature. If it is incorrect, it may be stupid feature, or
>>> an antisocial feature, or a non-compliance with some standard.
>>
>> I agree with Nick (as they say). Just because the author of a program

>> intended something doesn???t stop it being a bug. Sometimes programmers

>> are just wrong.
>>
>
> As long as you people don't agree on teh semantics of the term "bug" you
> can argue until the kine come home. ;)
>
> however, the software industry even looks at bugs that are introduced as
> early as in the user story and product specification stage (ie. even
> before the formal software specification, if that is ever done). Those
> are particularly expensive if uncaught.

The question is whether, if the program does exactly what the
designer/programmer/... wanted it to do (no matter how stupid that
intent was), it can be called a bug.

>
>
>> Just because the author of a program

>> intended something doesn???t stop it being a bug. Sometimes programmers

>> are just wrong.
>
> Btw software development is a holistic process and i am not aware of
> anywhere in the industry where you still have "programmers" these days,
> but that may be just my environment.

So, we can add a discussion about the definition of "programmer" to the
mix as well:-)

[Martin]

On 03/29/2013 04:30 PM, unruh wrote:
> On 2013-03-29, Martin <n...@spam.invalid> wrote:
>> however, the software industry even looks at bugs that are introduced as
>> early as in the user story and product specification stage (ie. even
>> before the formal software specification, if that is ever done). Those
>> are particularly expensive if uncaught.
>
> The question is whether, if the program does exactly what the
> designer/programmer/... wanted it to do (no matter how stupid that
> intent was), it can be called a bug.
>

in the statement you quote I am answering this question in the affirmative.

[Martin Gregorie]

On Fri, 29 Mar 2013 15:30:24 +0000, unruh wrote:

>
> The question is whether, if the program does exactly what the
> designer/programmer/... wanted it to do (no matter how stupid that
> intent was), it can be called a bug.
>

IMO, if the program does exactly what its author intended but most users
consider that choice to be annoying or stupid, its more properly
described as as 'feature' than a 'bug'.

There is another meaning of 'feature' that both ICL and IBM used from
time to time. If a program did something that was annoyingly different
from what the manual's wording said it should do, the next release of the
program would turn out to do exactly the same thing but now the manual
had been altered agree with what the program did. If you queried this
with your assigned sales-droid or an SE you got told that it was a
feature and how dare you call it a bug since the program did what the
manual said.


--
martin@ | Martin Gregorie
gregorie. | Essex, UK
org |

[Jasen Betts]

On 2013-03-29, Martin Gregorie <mar...@address-in-sig.invalid> wrote:
> On Fri, 29 Mar 2013 15:30:24 +0000, unruh wrote:
>
>>
>> The question is whether, if the program does exactly what the
>> designer/programmer/... wanted it to do (no matter how stupid that
>> intent was), it can be called a bug.
>>
> IMO, if the program does exactly what its author intended but most users
> consider that choice to be annoying or stupid, its more properly
> described as as 'feature' than a 'bug'.

How about "misfeature" or "wart"

> There is another meaning of 'feature' that both ICL and IBM used from
> time to time. If a program did something that was annoyingly different
> from what the manual's wording said it should do, the next release of the
> program would turn out to do exactly the same thing but now the manual
> had been altered agree with what the program did.

I've had something like that happen to me, but in the interim I had prepared
a source patch and everything. I wanted the old documented behaviour.

--
⚂⚃ 100% natural

--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---

[Martin Gregorie]

So did we.

The bug in question was in ICL's IDMSX network database. It was enscribed
in the manual that, if an ordered set contained more than one type of
member record, traversing it in ordered sequence would follow strict key
sequence regardless of record type. However, in practice records were
fetched grouped by record type with each group ordered in key sequence,
which was NOT what we needed.

So, we raised a bug and heard nothing. When the next release appeared, we
discovered the behavior we'd bugged and that the manual had been
rewritten to say that this type of set would return records grouped by
type with records in key sequence within each type.

Now doubt it was cheaper to rewrite that paragraph in the manual than to
fix IDMXS ordered fetch behavior, so that was what some PHB had decreed
should be done. Grrrrrr!

I forget what we did to overcome the problem: probably respecified the
set to use a single member record type with REDEFINEs to remap all non-
key fields in that record into a common area. The system and record
definitions were written in COBOL and the IDMSX bindings formed a COBOL
module generated from the schema definition, that sort of trick worked
quite well.

[Daniel James]

In article <kj0n7p$f44$1...@dont-email.me>, Aragorn wrote:
> If it is intended behavior by the authors of the software, then it is

> not a bug ...

Not an implementation bug, but it can still be a conceptual bug.

In this case the authors of the software are clearly wrong. The user
should be able at least to override the default behaviour, and
preferably to be able to change the default.

There's no single right answer where cross-posting is concerned. In
almost all cases the correct answer would be not to cross-post in the
first place, but once a thing is cross-posted it will have been seen by
a number of people each of whom follows only one of the original groups
and restricting the circulation with a follow-up is a disservice to
some/all of those people.

Sometimes a cross-post is erroneous and one group is not interested in
the post, sometime a thread drifts and becomes less relevant to one or
more of the original groups or becomes of interest to another group
that was not a destination of the original thread -- in these cases it
may make sense to change the list of groups in the "Newsgroups" header,
but it seldom makes sense to set a followup.

What KNode does *is* broken. Worse: the broken behaviour has been
implemented as per spec!

Cheers,
Daniel.

[Daniel James]

In article <kj7649$cgn$1...@localhost.localdomain>, Martin Gregorie wrote:
> ... When the next release appeared, we discovered the behavior we'd

> bugged and that the manual had been rewritten to say that this type
> of set would return records grouped by type with records in key
> sequence within each type.
>
> Now doubt it was cheaper to rewrite that paragraph in the manual

> than to fix IDMXS ordered fetch behavior ...

Cheaper in two senses ... implementing the behaviour you wanted would
probably have involved an additional sort (or, at least, merge)
operation after fetching the set of records, which would have imposed a
runtime penalty in both time and storage. Anyone performing the
operation who didn't need the strict ordering would have been annoyed
at the additional overhead.

In this case, at least, I'd say the decision was probably the correct
one.

Cheers,
Daniel.

[The Real Doctor]

On 28/03/13 14:29, Aragorn wrote:
> And it isn't just the KDE developers who are guilty of this.

Indeed. It's an endemic issue. Look at how the Gnome people decided that
it would be a good idea to change their desktop completely, and how
rapidly that catapulted them from "leading desktop environment" to
"collapsing also-ran".

Ian

[Martin]

On 04/02/2013 12:41 AM, The Real Doctor wrote:
> On 28/03/13 14:29, Aragorn wrote:
>> And it isn't just the KDE developers who are guilty of this.
>
> Indeed. It's an endemic issue.

oh really. the people putting in their own effort and their own life
time to create something and giving it away for free turn out to be the
ones calling the shots? outrageous. people, get over it, or start
programming yourselves.

[unruh]

Yes, they do call the shots. Unfortunately or otherwise, they often do
so so that their work will actually become popular and used by lots of
people. Thus, their changes, which then proceed to destroy many many
other people's work, can be irresponsible, and harmful. It is like
saying that if a investment manager skims off one's life savings, since
it is he who is putting in the work to invest the money, this is
acceptable behaviour. If you do not want that to happen, you should
learn economics and do your own investing.
Life is too short to do everything oneself, but that means you have to
trust others to do things you cannot competently and responsibly. That
trust is a valuable commodity in making society work, and just because
you donate your time and effort does not relieve you from the
responsibility.


>

[J G Miller]

On Tuesday, April 2nd, 2013, at 15:36:48h +0000, UnRuh wrote:

> but that means you have to trust others to do things
> you cannot competently and responsibly.

And also hire a good "claims" lawyer.

[The Real Doctor]

On 02/04/13 15:27, Martin wrote:
> oh really. the people putting in their own effort and their own life
> time to create something and giving it away for free turn out to be the
> ones calling the shots?

Of course they can. They did, after all. But Gnome is a dire example of
what happens when programmers - who one presumes want other people to
use their work - decide to tell their users what they want.

One of the joys and one of the freedoms of Linux is the ability to say
"fuck off" to software teams who do that without any financial
investment being lost.

Ian

[The Real Doctor]

On 02/04/13 16:36, unruh wrote:
> Yes, they do call the shots. Unfortunately or otherwise, they often do
> so so that their work will actually become popular and used by lots of
> people. Thus, their changes, which then proceed to destroy many many
> other people's work, can be irresponsible, and harmful.

It is also worth noting that those who fucked up Gnome by pushing
through an ugly and unwanted interface change thereby nullified the work
of the many other and previous programmers who had contributed to the
project.

Ian

[Martin Gregorie]

...but it was already stuffed up thanks to the stuff they did to final
version of Gnome 2 - was that 2.3? They'd done a partial implementation
whereby all the options etc were there but in many cases did nothing and
documented as such. Did they finish that job? Nah, it got kicked into the
long grass and forgotten for a few months before they introduced the
Gnome 3 mess.

I used to like Gnome, but I am now using XFCE to show my appreciation of
the new, improved Gnome 3.

[Martin Gregorie]

With all due respect, you're wrong here.

IDMSX is/was a network database. Unlike the relational model where
relationships don't exist except, maybe, as unordered constraints, in
IDMSX they are implemented as a pointer chain that links members of a set
to their owner. Any ordering is therefore embodied in the linkage chain:
the DBMS never sorts records stored in the database or needs to because
the each member is added to the set by inserting it in the place needed
to preserve the required key sequence. IOW, the time it takes to insert
records into a sorted set or to retrieve them is independent of whether a
set contains one record type or many.

[unruh]

A lawyer is not substitute for trust.

[Martin]

On 04/02/2013 07:14 PM, The Real Doctor wrote:
> On 02/04/13 15:27, Martin wrote:
>> oh really. the people putting in their own effort and their own life
>> time to create something and giving it away for free turn out to be the
>> ones calling the shots?
>
> Of course they can. They did, after all. But Gnome is a dire example of
> what happens when programmers - who one presumes want other people to
> use their work - decide to tell their users what they want.

I agree to some extent, but the point is that the user base is not
homogeneous. Some like a product direction, some are indifferent, others
dislike it, and the latter ones usually make more noise than the former
ones. So far so good.

What ticks me off are delusional paranoid people who

(a) rant against every major change of every software product (the same
people who formerly condemned KDE 3 now praise it because according to
them KDE 4 is the antichrist)

(b) attribute these changes to a microsoft conspiracy, or at least
condemn those changes as an attempt to indoctrinate the open source
world with microsoft concepts

(c) have never contributed to any project in a constructive way

(d) assume it is their god-given right to place their demands on other
people

(e) are often foul-mouthed

> One of the joys and one of the freedoms of Linux is the ability to say
> "fuck off" to software teams who do that without any financial
> investment being lost.

Totally agree. It is quite natural that there are conflicting opinions
and all sorts of resolution strategies occur, including the "f*** off"
approach on both sides of an argument. It shouldn't be the preferred way
of settling an argument, though.

[The Real Doctor]

On 02/04/13 21:36, Martin Gregorie wrote:
> I used to like Gnome, but I am now using XFCE to show my appreciation of
> the new, improved Gnome 3.

Me too, on both counts. I liked LXDE, but it doesn't handle dual
monitors well and XFCE does.

Ian

[J G Miller]

On Wednesday, April 7th 2013, at 12:29:43h +0200, Martin ranted:

> What ticks me off are delusional paranoid people who

One has to remember though that no matter how hard one tries,
some people are always going to bitch, so to preserve one's
own sanity, just say "there, there", and refer them to help
from the "Care in the Community" program.

Just think, if they were not provided with something to complain
about, they would be unhappy, because they are only happy
when they have something to complain about, so the programmers
are actually providing them with an important service.

Meanwhile over on channel 43, stunning new evidence that it
was extra-terrestrials from Alpha Centauri who planted the
bombs which brought down Building 7 ...

[J G Miller]

On Tuesday, April 2nd, 2013, at 23:58:12h +0000, UnRuh wrote:

> A lawyer is not substitute for trust.

Can one really trust a lawyer? ;)

"Trust, but verify."
-- President Ronald Reagan

[Aragorn]

On Wednesday 03 April 2013 12:29, Martin conveyed the following to
uk.comp.os.linux...

> On 04/02/2013 07:14 PM, The Real Doctor wrote:
>
>> On 02/04/13 15:27, Martin wrote:
>>
>>> oh really. the people putting in their own effort and their own life
>>> time to create something and giving it away for free turn out to be
>>> the ones calling the shots?
>>
>> Of course they can. They did, after all. But Gnome is a dire example
>> of what happens when programmers - who one presumes want other people
>> to use their work - decide to tell their users what they want.
>
> I agree to some extent, but the point is that the user base is not
> homogeneous. Some like a product direction, some are indifferent,
> others dislike it, and the latter ones usually make more noise than
> the former ones. So far so good.
>
> What ticks me off are delusional paranoid people who
>
> (a) rant against every major change of every software product (the
> same people who formerly condemned KDE 3 now praise it because
> according to them KDE 4 is the antichrist)

I personally feel that KDE has /always/ received quite a lot of
undeserved flak.

While it is true that the distribution vendors were pushing KDE 4 (and
in some cases, /deliberately/ breaking KDE 3 so as to push the
migration) long before KDE 4 was production-ready, I have to say that
KDE 4 /is/ a great desktop environment, even though it /does/ come with
quite a lot of Microsoft'isms "out of the box". But then again, KDE
being KDE - and this is one of the reasons why I like it so much - the
"out of the box" look and feel is easily changed into something of your
own liking.

My own desktop is organized in a way that makes sense to myself, and
that includes several GUI concepts from the Apple OSX and OS9 user
interfaces, such as a desktop menu at the top of the screen - embedded
in a thin panel at the top of the screen, which also holds a few icons,
among which the KDE main menu - and having my icons on the right side of
the screen instead of on the left. I've always liked that feature, and
I also have the "window close" title bar button on the left hand side,
not on the right. Many traditional UNIX window managers also had it on
the left, and it was also featured in two of the stock window
decorations for KDE 2.

> (b) attribute these changes to a microsoft conspiracy, or at least
> condemn those changes as an attempt to indoctrinate the open source
> world with microsoft concepts

I don't think there's a conspiracy going on to indoctrinate the FLOSS
world with Microsoft concepts, but one cannot deny the fact that more
and more Windows'isms are creeping into the GNU/Linux GUIs, and
especially so through the influence of freedesktop.org (and thus
RedHat).

Fact is that we are now in the second decade of the 21st century. GNU
has its origins back in 1983 and Linux back in 1991, and most of the
GNU/Linux developers back then were guys from the UNIX world. Meanwhile
we're over 20 years later and most of the new FLOSS developers today are
people who've come from the Microsoft Windows world. They bring their
own influences with them, and it shows in their work.

One of such Microsoft'isms which annoys me to death is the move towards
a volume-oriented approach to disk storage via the "Places" menus in the
Dolphin, Konqueror and Nautilus file managers - even GNOME 2 had a
"Places" menu. That may all look swell if you have your entire
operating system installed in a single partition, and perhaps it makes
even more sense for people who dual-boot between GNU/Linux and Microsoft
Windows. But if you, like me, have your system distributed across
multiple separate partitions - with different mount options et al for
reasons of security and fault tolerance - and you have two GNU/Linux
distributions installed in this manner on one and the same computer,
then it becomes a nightmare having to navigate through that mess.

The UNIX filesystem hierarchy has always been clean and uncluttered.
One could navigate through the entire structure in a logical and
sensible way, and that is what I myself prefer as well.

(And no, I'm not one of those old UNIX dinosaurs from the 1970s, even
though I do have some (proprietary) UNIX experience [*]. I was
primarily an OS/2 user in the years before I used GNU/Linux, and I also
had some brief experience with Windows NT 4.0 (2 years) and Windows 3.1
on DOS (6 months) in that time. And I have always liked the unified
directory tree of UNIX. It makes perfect sense, and it's transparent to
the users.)


[*] Sperry OS/3 on a Unisys mini and DEC OSF/1 on an NSR server, but in
neither case as the root user.

--
= Aragorn =
GNU/Linux user #223157 - http://www.linuxcounter.net

[Martin]

On 04/03/2013 08:04 PM, Aragorn wrote:

>
> [vindication of KDE 4]
>

I've always liked KDE to some extent, too. Only the resource consumption
per user account made me turn to non-KDE applications, and when I
stopped using a compositing WM (to free up GPU resources) I switched to
XFCE altogether.

>
> [criticism of volume-oriented representation of disk storage]
>

I would agree with that - not having seen it on my desktop a such. What
I do find helpful is when Thunar shows auto-mounted memory sticks as
root nodes in the left panel.

However, a lot of the supposed "microsoft'isms" neither originate from
microsoft (they just copy ideas) nor are they part of a conspiracy. Some
might even be useful.

Sometimes people go as far as associating unwelcome ideas (eg. systemd)
with microsoft for the sole purpose of discrediting them further (which
in the case of systemd is not even necessary).

The problem with all that is that it blinds us from the real issue which
is microsoft's hegemonial business behaviour (including FUD, lock-in of
hardware vendors and OEMs, DRM, UEFI).

Lastly, should microsoft disappear over night, there would be other
proprietary vendors taking over seamlessly.

[Martin]

On 04/03/2013 01:50 PM, J G Miller wrote:
> On Wednesday, April 7th 2013, at 12:29:43h +0200, Martin ranted:
>
>> What ticks me off are delusional paranoid people who
>
> One has to remember though that no matter how hard one tries,
> some people are always going to bitch, so to preserve one's
> own sanity, just say "there, there", and refer them to help
> from the "Care in the Community" program.

hehe, I usually just turn the other way.

>
> Just think, if they were not provided with something to complain
> about, they would be unhappy, because they are only happy
> when they have something to complain about, so the programmers
> are actually providing them with an important service.

true. and don't we all like to rant about something at some point. ;)

>
> Meanwhile over on channel 43, stunning new evidence that it
> was extra-terrestrials from Alpha Centauri who planted the
> bombs which brought down Building 7 ...
>

did they say the extra-terrestrial are also responsible for oklahoma city?

[J G Miller]

On Thursday, April 4th, 2013, at 10:44:30h +0200, Martin asked:

> did they say the extra-terrestrial are also responsible for oklahoma city?

Well everybody "knows" that it was Timothy McVeigh who did the job, but one
does have to remember that he was not wearing protective headgear so he
may have come under the influence of the mind control rays of a few
surviving angry Martians who obviously have a grudge against capitalism.

"I have always said, heard, that it would not be strange
that there had been civilization on Mars, but maybe capitalism
arrived there, imperialism arrived and finished off the planet."

President Chavez in speech to mark World Water Day, March 2011

<http://www.reuters.COM/article/2011/03/22/us-venezuela-chavez-mars-idUSTRE72L61D20110322>

[J G Miller]

On Thursday, April 4th, 2013, at 10:43:53h +0200, Martin wrote:

> Lastly, should microsoft disappear over night, there would be other
> proprietary vendors taking over seamlessly.

Apple already has.

[Martin]

On 04/04/2013 01:35 PM, J G Miller wrote:
> On Thursday, April 4th, 2013, at 10:44:30h +0200, Martin asked:
>
>> did they say the extra-terrestrial are also responsible for oklahoma city?
>
> Well everybody "knows" that it was Timothy McVeigh who did the job, but one
> does have to remember that he was not wearing protective headgear so he
> may have come under the influence of the mind control rays of a few
> surviving angry Martians who obviously have a grudge against capitalism.

maybe it was those mind control rays that got reflected and took out the
internal pillars not reachable by a fertilizer bomb exploding out in the
open?

[Martin]

lol, that's exactly who I had in mind when writing the comment above

[J G Miller]

On Thursday, April 4th, 2013, at 17:01:48h +0200, Martin wrote:

> maybe it was those mind control rays that got reflected and took out the
> internal pillars not reachable by a fertilizer bomb exploding out in the
> open?

Well according to

<http://www.rickross.COM/reference/mcveigh/mcveigh8.html>

which claims to be a reproduction of an article in in The Independent
newspaper (published in the London Borough of Kensington and Chelsea)
of May 11, 2001

QUOTE

There was one claim that the bombing was a federal sting operation
gone horribly wrong; another that there were explosive packs strapped
to the internal pillars of the Murrah building, timed to go off at the
same time as the fertiliser bomb.

There is no credible evidence for either claim.

UNQUOTE


I would suggest however, that is highly unlikely that McVeigh acted on his own.

[Norman Hull]

unruh wrote:

> It is like
> saying that if a investment manager skims off one's life savings, since
> it is he who is putting in the work to invest the money, this is
> acceptable behaviour. If you do not want that to happen, you should
> learn economics and do your own investing.

Sorry folks, but this remark is pure bollocks

--
Norman
If at first you don't succeed - give someone else a go