Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[akhiezer]

> From: Bruce Dubbs 
> Date: Tue, 29 Aug 2017 12:32:38 -0500
>
> akhiezer wrote:
>
> > The matter is clearly, obviously, not about how long a single error
> > persists.
> >
> >
> > It is about how long folks remain addicted to donkey-tasks; the
> > inevitable stream, cycle, of repeated types of errors that ensue; and
> > the inevitable stream, cycle, of repeated types of donkey-tasks that
> > ensue of correcting said inevitable errors.
>
> We are waiting for you to fork blfs with a 'better way'.
>
>-- Bruce
>


If a person long-term demonstrates repeatedly that they won't, or can't,
implement properly even the handling of checksum calculations -
whether doing in-house or availing themselves of any of those already
widely implemented and readily available;
then how likely are they to be able &/or willing to avail themselves
satisfactorily of any new such placed in front of them.


It is therefore not enough in such a case, for the person to attempt
to deflect attention from the matter by asininely/ trying
to get others to implement yet another instance,
while the person's past and current behaviours, attitudes, indicate that
the person is too-likely to continue as-ever.


Instead, the person has to at least exhibit some intent,
willingness, some good-faith, that such new materials
will not be wasted on them.


It is advisable and appropriate in such a case that
an approach of 'lead to water', 'teach to fish', be used.


To that end, a reasonable suggested starting sequence
of infos, tasks, for such a person, is:
---
* implement auto-calc of checksum.
  Then add cross-checking.

* then implement auto-writing-out of the info to a file.
  Then add cross-checking.

* then implement auto-writing-out to the required place in
  the required xml/ file(s).
  Then add cross-checking.

* then implement auto-sync/commit - with cross-checking - with svn/
---
Keep a note of any issues encountered; and of any ideas
(look it up, if nec) for how the process might be improved further.


If that goes all-ok, then progress to auto-handling of date/time info;
ref above seq of steps.


Any problems, let the lists know.


It is stated perennially that a central goal, pillar, of b/lfs is education;
is the person per the above, willing &/or able to allow themselves to learn.


Folks, I know that you may be hurting to be queried: if
you at core know and understand the matter -
the issues and resolutions - and will improve,
then that's the main thing,
and relegates the protestations-too-much to where they belong.



akh





--
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[akhiezer]

> Date: Tue, 29 Aug 2017 19:26:32 +0100
> From: Ken Moffat 
>
> On Tue, Aug 29, 2017 at 06:19:42PM +0100, akhiezer wrote:
> > > Date: Tue, 29 Aug 2017 17:44:13 +0100
> > > From: Ken Moffat 
> > >
> > >
> > > I look forward to your fork where you show us how to automate this.
> > > Bonus points for using no more than 4 cores on machines with more
> > > cores (for things using rust and ninja).  More bonus points for not
> > > requiring user input to specify which files to read for
> > > measurements.
> > >
> > 
> > 
> > Here's a clue: break the problem down: start with (for you) low-hanging
> > fruit: how about you start with checksums; and then maybe progress to
> > dates; see if you can manage that first.
> > 
> > 
> > Then you might start to get some vision - taste, abilities, even -
> > for the parts that currently seem to be over-the-horizon difficult for
> > you. But do keep breaking the problem down into manageable pieces.
> > 
>
> Thanks for your expressed lack of confidence in my taste and vision.
> Your comment on my lack of ability is, of course, well-founded - but
> since the project lacks skilled and able contributors, I try to
> contribute what I can.
>
> As far as I am concerned, I _use_ parts of BLFS, so I try to ensure
> they are up to date.  Anything more is a distraction - sometimes the
> distractions are interesting, other times not.
>
> So no, taking time out to figure out how to automate finding the
> md5sums in the book, as a first step towards comparing them, is not
> attractive.
>
> We work with what we have.  If we started today, I very much doubt
> that the book would be structured as it is - a new structure *might*
> support some of your proposed automation, it might not.  So, it is
> up to you to fork at least a proof of concept.  However, I suspect
> that I at least will not share your taste.
>
> Either you really are keen to contribute, or it's just another excuse
> to be passive-aggressive.  I'll be charitable, so "Enjoy forking, and
> let us know when you have a proof of concept."
>


Might such lazy facile attempted 'pssv/aggrsv' labelling be,
rather, applied to your reactions at your poor practices
being at least queried.


Wrong instructions and other wrong infos are being put
into the book - the central repo - that have not even been
tested and verified-ok; letting others hit the problems.


And you try to nay-say when queried about your practices
and urged to change and not just keep repeating the cycle.
Who do you think you are. The issues are not so much just
about abilities per se; it's also attitude.


A person's own repo should be in order before
pushing to central.


Your talk-to-the-hand final position statement is
interestingly attemptedly high-handed.

Demonstrate ability to learn and improve, such that
you might learn from such materials.

To get you started on the first steps towards improvement,
ref the initial getting-you-started checklists in the note
to your similarly-postured colleague.


Folks, I know that you may be hurting to be queried: if
you at core know and understand the matter -
the issues and resolutions - and will improve,
then that's the main thing,
and relegates the protestations-too-much to where they belong.



akh





--
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[Ken Moffat]

On Tue, Aug 29, 2017 at 06:19:42PM +0100, akhiezer wrote:
> > Date: Tue, 29 Aug 2017 17:44:13 +0100
> > From: Ken Moffat 
> >
> >
> > I look forward to your fork where you show us how to automate this.
> > Bonus points for using no more than 4 cores on machines with more
> > cores (for things using rust and ninja).  More bonus points for not
> > requiring user input to specify which files to read for
> > measurements.
> >
> 
> 
> Here's a clue: break the problem down: start with (for you) low-hanging
> fruit: how about you start with checksums; and then maybe progress to
> dates; see if you can manage that first.
> 
> 
> Then you might start to get some vision - taste, abilities, even -
> for the parts that currently seem to be over-the-horizon difficult for
> you. But do keep breaking the problem down into manageable pieces.
> 

Thanks for your expressed lack of confidence in my taste and vision.
Your comment on my lack of ability is, of course, well-founded - but
since the project lacks skilled and able contributors, I try to
contribute what I can.

As far as I am concerned, I _use_ parts of BLFS, so I try to ensure
they are up to date.  Anything more is a distraction - sometimes the
distractions are interesting, other times not.

So no, taking time out to figure out how to automate finding the
md5sums in the book, as a first step towards comparing them, is not
attractive.

We work with what we have.  If we started today, I very much doubt
that the book would be structured as it is - a new structure *might*
support some of your proposed automation, it might not.  So, it is
up to you to fork at least a proof of concept.  However, I suspect
that I at least will not share your taste.

Either you really are keen to contribute, or it's just another excuse
to be passive-aggressive.  I'll be charitable, so "Enjoy forking, and
let us know when you have a proof of concept."

ĸen
-- 
Truth, in front of her huge walk-in wardrobe, selected black leather
boots with stiletto heels for such a barefaced truth.
 - Unseen Academicals
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[Bruce Dubbs]

akhiezer wrote:


The matter is clearly, obviously, not about how long a single error
persists.


It is about how long folks remain addicted to donkey-tasks; the
inevitable stream, cycle, of repeated types of errors that ensue; and
the inevitable stream, cycle, of repeated types of donkey-tasks that
ensue of correcting said inevitable errors.


We are waiting for you to fork blfs with a 'better way'.

  -- Bruce


--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[akhiezer]

> From: Bruce Dubbs 
> Date: Tue, 29 Aug 2017 10:48:07 -0500
>
> akhiezer wrote:
> >> From: Wayne Blaszczyk 
> >> Date: Tue, 29 Aug 2017 19:16:41 +1000
> >>
> >> Again, it seemed not to change from the previous version.
> >>
> >
> >
> > The central issue is that much of that 'header' info
> > (dates/checksums/sizes/build-times/...) get input manually, instead of
> > done automatically. That the former persists for so long is technically
> > and administratively incompetent; and it's further worsened that it
> > gets clung onto.
>
> For 'so long' is less than 24 hours.  I check new commits every morning on 
> the day after the commit.
>
> Note that there are sometimes problems with stealth updates and 
> differences between mirrors.
>


As often, you (willfully?, cos "nobody's _that_ stupid, surely") happily,
brightly, address a different point.


The matter is clearly, obviously, not about how long a single error
persists.


It is about how long folks remain addicted to donkey-tasks; the
inevitable stream, cycle, of repeated types of errors that ensue; and
the inevitable stream, cycle, of repeated types of donkey-tasks that
ensue of correcting said inevitable errors.



akh





--
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[Ken Moffat]

On Tue, Aug 29, 2017 at 12:42:35PM +0100, akhiezer wrote:
> > From: Wayne Blaszczyk 
> > Date: Tue, 29 Aug 2017 19:16:41 +1000
> >
> > Again, it seemed not to change from the previous version.
> >
> 
> 
> The central issue is that much of that 'header' info
> (dates/checksums/sizes/build-times/...) get input manually, instead of
> done automatically. That the former persists for so long is technically
> and administratively incompetent; and it's further worsened that it
> gets clung onto.
> 
> 
> 
> akh
> 

I'll begin by mentioning a few things about what has to be done when
editing - I'm sure you know all this, but I'd hate for people
inspired by automation to overlook the pesky details.

For build times, and sizes, the figures have to be measured and
recorded, and these figures then need to be picked up by your
process which puts them into the book.

In particular, for BLFS an SBU itself has to be recalculated when the
box has been booted for the first time with the new LFS system
(perhaps 10% longer than the SBU for the previous release).

For configure or equivalent, the switches need to be checked : did
what we use, or the defaults, change ?  Do we need a new switch to
get something ?

There is also the question of dependency changes, but I'm sure that
will always require manual investigation and judgement.

For installing a package we will do a DESTDIR or equivalent if at
all possible (with whatever variations to the invocation are needed.
See in particular QT packages, but also biber, some python modules
for variations - and a few need root even for DESTDIR).

In some packages, repeat with testsuite and observe if the book's
instructions are still correct.

I look forward to your fork where you show us how to automate this.
Bonus points for using no more than 4 cores on machines with more
cores (for things using rust and ninja).  More bonus points for not
requiring user input to specify which files to read for
measurements.

ĸen
-- 
Truth, in front of her huge walk-in wardrobe, selected black leather
boots with stiletto heels for such a barefaced truth.
 - Unseen Academicals
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[Bruce Dubbs]

akhiezer wrote:

From: Wayne Blaszczyk 
Date: Tue, 29 Aug 2017 19:16:41 +1000

Again, it seemed not to change from the previous version.




The central issue is that much of that 'header' info
(dates/checksums/sizes/build-times/...) get input manually, instead of
done automatically. That the former persists for so long is technically
and administratively incompetent; and it's further worsened that it
gets clung onto.


For 'so long' is less than 24 hours.  I check new commits every morning on 
the day after the commit.


Note that there are sometimes problems with stealth updates and 
differences between mirrors.


  -- Bruce


--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[akhiezer]

> From: Wayne Blaszczyk 
> Date: Tue, 29 Aug 2017 19:16:41 +1000
>
> Again, it seemed not to change from the previous version.
>


The central issue is that much of that 'header' info
(dates/checksums/sizes/build-times/...) get input manually, instead of
done automatically. That the former persists for so long is technically
and administratively incompetent; and it's further worsened that it
gets clung onto.



akh





--
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-dev] ImageMagick-7.0.6-10 md5sum is not correct

[Wayne Blaszczyk]

Again, it seemed not to change from the previous version.

Regards,
Wayne.
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page