Re: [gentoo-user] Libsld, what gives?

2022-11-17 Thread Peter Humphrey
On Thursday, 17 November 2022 12:52:33 GMT Dr Rainer Woitok wrote:
> Paul,
> 
> On Thursday, 2022-11-17 17:52:17 +1100, you wrote:
> > On Wednesday, November 16, 2022 6:11:18 P.M. AEDT Alan Grimes wrote:
> > > ...
> > > 
> > >Usually I should wait six months just to save myself
> > > 
> > > the aggrivation...
> > 
> > No, waiting 6 months between updates *causes* aggrivation.
> 
> The last word you're using here was originally coined by Allan.  Since I
> didn't find it in my dictionary, I assume it doesn't exist.   So I would
> suggest using either "aggravation" or "aggrievation", another word I did
> not find in my dictionary, but which seems to fit even better ... :-)

If you consult a dictionary, you'll find that 'aggravate' means 'make worse'; 
thus, a medical condition may be aggravated by neglect.

The word you want is 'annoyance' - or just 'trouble'.

-- 
Regards,
Peter.






Re: [gentoo-user] Libsld, what gives?

2022-11-17 Thread Dr Rainer Woitok
Paul,

On Thursday, 2022-11-17 17:52:17 +1100, you wrote:

> On Wednesday, November 16, 2022 6:11:18 P.M. AEDT Alan Grimes wrote:
> > ...
> >Usually I should wait six months just to save myself
> > the aggrivation...
> 
> No, waiting 6 months between updates *causes* aggrivation.

The last word you're using here was originally coined by Allan.  Since I
didn't find it in my dictionary, I assume it doesn't exist.   So I would
suggest using either "aggravation" or "aggrievation", another word I did
not find in my dictionary, but which seems to fit even better ... :-)

Sincerely,
  Rainer



Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Paul Colquhoun
On Wednesday, November 16, 2022 6:11:18 P.M. AEDT Alan Grimes wrote:
> Even though only 45 days have passed since my last update, I felt like
> doing one tonight. Usually I should wait six months just to save myself
> the aggrivation...

No, waiting 6 months between updates *causes* aggrivation.

Try updateing on a regular schedule, at 1 or 2 week intervals, and see if 
your experience improves.


-- 
Reverend Paul Colquhoun, ULC. http://andor.dropbear.id.au/
  Asking for technical help in newsgroups?  Read this first:
 http://catb.org/~esr/faqs/smart-questions.html#intro



Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Michael Cook

On 11/16/22 19:18, Laurence Perkins wrote:



tortoise ~ # eclean-dist
  * Building file list for distfiles cleaning...
  * ERROR: games-strategy/boswars-::poly-c failed (depend phase):
  *   EAPI 6 unsupported.

Possibly also need to upgrade portage and/or gentoolkit first.  That's usually 
a good idea regardless.


Hmm, poly-c is looking like it's a common factor here.



Overlays can go stale, even if they were previously necessary to fix something. 
 For long-term use pulling just the ebuilds you need into your own overlay 
often works out better.

LMP


Personally I just mask the overlay and unmask what I need. If something 
breaks, I look into fixing it. Upgrading rather frequently seems to 
limit this, just maintain the system. Waiting months and then having to 
deal with tons of changes and an overlay going dead (I don't think 
poly-c exists anymore?) is way worse.





RE: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Laurence Perkins


>tortoise ~ # eclean-dist
>  * Building file list for distfiles cleaning...
>  * ERROR: games-strategy/boswars-::poly-c failed (depend phase):
>  *   EAPI 6 unsupported.

Possibly also need to upgrade portage and/or gentoolkit first.  That's usually 
a good idea regardless.

>Hmm, poly-c is looking like it's a common factor here.
>
>

Overlays can go stale, even if they were previously necessary to fix something. 
 For long-term use pulling just the ebuilds you need into your own overlay 
often works out better.

LMP


Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Alan Grimes

Laurence Perkins wrote:

If you're going to try to dig all the way to the bottom first then 
--ignore-world and --ignore-built-slot-operator-deps can be helpful for forcing 
it to build what it needs to break a dependency loop.  But do be aware that 
things may cease to function during the intermediate stages.

If those are not sufficiently strong, then you can climb into the package repo 
and start issuing ebuild commands yourself and it will do what it's told 
without bothering about checking dependencies at all.  Make sure you know what 
you're doing...  Obviously...

But seriously, try disabling any overlays first if at all possible.  It's quite 
common for those to lag behind the main repo and turn things into a tangled 
mess.


At this point the jackhammers have gone silent and I'm left with the 
issue I had before:



tortoise ~ # emerge --update --newuse --deep --with-bdeps=y world 
--verbose --backtrack=40 --changed-deps --verbose-conflicts --pretend 
--ignore-built-slot-operator-deps=y


These are the packages that would be merged, in order:

Calculating dependencies... done!

emerge: there are no ebuilds to satisfy 
"media-libs/libsdl2[abi_x86_32?,alsa?,custom-cflags?,joystick?,nas?,opengl?,oss?,pulseaudio?,sound?,static-libs?,video?,X?]".

(dependency required by "media-libs/libsdl-1.2.60::poly-c" [ebuild])
(dependency required by "dev-libs/zziplib-0.13.72-r2::gentoo[sdl]" [ebuild])
(dependency required by "app-text/texlive-core-2021-r4::gentoo" [installed])
(dependency required by "dev-texlive/texlive-formatsextra-2021::gentoo" 
[installed])
(dependency required by 
"app-text/docbook-sgml-utils-0.6.14-r5::gentoo[jadetex]" [installed])
(dependency required by "gnome-extra/gnome-color-manager-3.36.0::gentoo" 
[installed])
(dependency required by "gnome-base/gnome-control-center-43.1::gentoo" 
[installed])
(dependency required by "gnome-base/gnome-shell-42.5-r2::gentoo" 
[installed])
(dependency required by 
"gnome-extra/gnome-browser-connector-42.1::gentoo" [ebuild])

tortoise ~ #


I then checked my overlays, and it turns out I had an outrageously 
bloated number of layers installed:


tortoise ~ # layman -l

 * poly-c    [Rsync ] 
(rsync://rsync.gentoofan.org/poly-c )
 * wichtounet    [Git   ] 
(https://github.com/wichtounet/wichtounet-overlay.git )


tortoise ~ #
#

I think one of those was because mainline had stopped maintaining 
seamnokey and I needed it maintained. No idea what I was smoking when I 
applied the other or even which is which. In all likelyhood I had a 
problem and applying that overlay solved it and because I didn't have a 
problem after solving said problem, I forgot about it and can't really 
say what or even when it was I had that problem. All I have now is this 
lingering sense that touching those overlays will result in misery and 
suffering so it's best not to even look up how to remove either of them.


eclan still shows a large number of outdated packages on the system so 
there is definitely lots of updating to work if I can get emerge out of 
its current hissy-fit.


tortoise ~ # eclean-dist
 * Building file list for distfiles cleaning...
 * ERROR: games-strategy/boswars-::poly-c failed (depend phase):
 *   EAPI 6 unsupported.
 *
 * Call stack:
 * ebuild.sh, line 614:  Called source 
'/var/lib/layman/poly-c/games-strategy/boswars/boswars-.ebuild'
 *   boswars-.ebuild, line   7:  Called inherit 'desktop' 
'scons-utils' 'subversion'
 * ebuild.sh, line 294:  Called __qa_source 
'/usr/portage/eclass/scons-utils.eclass'
 * ebuild.sh, line 109:  Called source 
'/usr/portage/eclass/scons-utils.eclass'

 *    scons-utils.eclass, line  89:  Called die
 * The specific snippet of code:
 *  *) die "EAPI ${EAPI} unsupported."
 *
 * If you need support, post the output of `emerge --info 
'=games-strategy/boswars-::poly-c'`,
 * the complete build log and the output of `emerge -pqv 
'=games-strategy/boswars-::poly-c'`.

 * Working directory: '/usr/lib/python3.10/site-packages'
 * S: '/var/tmp/portage/games-strategy/boswars-/work/boswars-'
 * Cleaning distfiles...
 [    1.7 M ] frameworkintegration-5.98.0.tar.xz
 [   86.9 K ] kaccounts-integration-22.08.1.tar.xz
 [   63.1 K ] kaccounts-providers-22.08.1.tar.xz
 [   71.8 K ] kactivities-5.98.0.tar.xz
 [   88.9 K ] kauth-5.98.0.tar.xz
 [  127.0 K ] kbookmarks-5.98.0.tar.xz
 [  185.0 K ] kcmutils-5.98.0.tar.xz
 [  121.6 K ] kcompletion-5.98.0.tar.xz
 [   27.7 K ] kcrash-5.98.0.tar.xz
 [  202.3 K ] kdeclarative-5.98.0.tar.xz
 [   38.6 K ] kded-5.98.0.tar.xz
 [    3.5 M ] kdelibs4support-5.98.0.tar.xz
 [  430.4 K ] kdoctools-5.98.0.tar.xz
 [    1.6 M ] kemoticons-5.98.0.tar.xz
 [   94.9 K ] kglobalaccel-5.98.0.tar.xz
 [   72.1 K ] kguiaddons-5.98.0.tar.xz
 [    1.9 M ] khtml-5.98.0.tar.xz
 [  116.7 K ] kinit-5.98.0.tar.xz
 [  331.3 K ] kjs-5.98.0.tar.xz
 [   85.4 K ] knotifyconfig-5.98.0.tar.xz
 [  192.2 

RE: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Laurence Perkins

>-Original Message-
>From: Alan Grimes  
>Sent: Wednesday, November 16, 2022 8:18 AM
>To: gentoo-user@lists.gentoo.org; Andreas Fink 
>Subject: Re: [gentoo-user] Libsld, what gives?
>
>Andreas Fink wrote:
>> On Wed, 16 Nov 2022 02:11:18 -0500
>> Alan Grimes  wrote:
>>
>>> I'm jackhammering the system now and I'm getting about 50% error spew
>> I would suggest a different tool than a jackhammer to fix the problems.
>
>Basic assumption: Problems are caused by outdated packages.
>Underlying problem: In order to break a log-jam it is necessary to purge 
>outdated packages.
>Facts on the ground: The only way to make ANY progress in updating outdated 
>packages is to jackhammer the hell out of all packages in hopes that some of 
>them might sucessfully update which will, hopefully either:
>
>A. allow other packages to sucessfully update B. Expose something that can be 
>fixed.
>
>Once the logjam is broken, the system is then --emptytree world'ed and 
>declared healthy... Expected failure rate is on the order of 0.3% of hopefully 
>unimportant packages.
>
>It is known that the KDE group of packages has incomplete dependency graph 
>because it is hopelessly convoluted, and it always causes problems and the 
>only way to break through it is the jackhammer approach.
>
>LLVM and friends also fails to update when updating within a single slot and 
>this is VERY annoying...
>
>What causes me to post to the list is when I get failures that prevent me from 
>even beginning to jackhammer the system.
>
>--
>Beware of Zombies. =O
>#EggCrisis  #BlackWinter
>White is the new Kulak.
>Powers are not rights.
>

If you're going to try to dig all the way to the bottom first then 
--ignore-world and --ignore-built-slot-operator-deps can be helpful for forcing 
it to build what it needs to break a dependency loop.  But do be aware that 
things may cease to function during the intermediate stages.

If those are not sufficiently strong, then you can climb into the package repo 
and start issuing ebuild commands yourself and it will do what it's told 
without bothering about checking dependencies at all.  Make sure you know what 
you're doing...  Obviously...

But seriously, try disabling any overlays first if at all possible.  It's quite 
common for those to lag behind the main repo and turn things into a tangled 
mess.  

An expected failure rate "on the order of" 0.3% means you expect it to be no 
less than 0.03% and no more than 30%...  So I think you're probably already 
within that envelope.  :D

LMP


Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Alan Grimes

Andreas Fink wrote:

On Wed, 16 Nov 2022 02:11:18 -0500
Alan Grimes  wrote:


I'm jackhammering the system now and I'm getting about 50% error spew

I would suggest a different tool than a jackhammer to fix the problems.


Basic assumption: Problems are caused by outdated packages.
Underlying problem: In order to break a log-jam it is necessary to purge 
outdated packages.
Facts on the ground: The only way to make ANY progress in updating 
outdated packages is to jackhammer the hell out of all packages in hopes 
that some of them might sucessfully update which will, hopefully either:


A. allow other packages to sucessfully update
B. Expose something that can be fixed.

Once the logjam is broken, the system is then --emptytree world'ed and 
declared healthy... Expected failure rate is on the order of 0.3% of 
hopefully unimportant packages.


It is known that the KDE group of packages has incomplete dependency 
graph because it is hopelessly convoluted, and it always causes problems 
and the only way to break through it is the jackhammer approach.


LLVM and friends also fails to update when updating within a single slot 
and this is VERY annoying...


What causes me to post to the list is when I get failures that prevent 
me from even beginning to jackhammer the system.


--
Beware of Zombies. =O
#EggCrisis  #BlackWinter
White is the new Kulak.
Powers are not rights.




RE: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Laurence Perkins



>-Original Message-
>From: Andreas Fink  
>Sent: Wednesday, November 16, 2022 6:59 AM
>To: gentoo-user@lists.gentoo.org
>Subject: Re: [gentoo-user] Libsld, what gives?
>
>On Wed, 16 Nov 2022 02:11:18 -0500
>Alan Grimes  wrote:
>
>> I'm jackhammering the system now and I'm getting about 50% error spew
>I would suggest a different tool than a jackhammer to fix the problems.
>
>
A machete is a much better choice usually.  Chop out all the weeds like 
poorly-maintained overlays, programs you don't use any more, and (in extremity) 
the virtuals and perl.

Trying to go deep first and fix the core with all the brush and rubbish still 
cluttering up the top is usually an exercise in frustration and a good way to 
have some errant twig poke you in the eye repeatedly.

LMP



Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Julien Roy
Alan Grimes  writes:

> (dependency required by "media-libs/libsdl-1.2.60::poly-c" [ebuild])

This ebuild is from a different overlay, it's probably not helping. The
first thing I'd try is disabling the overlay to see if this is the cause
of the problem.

-- 
Julien


signature.asc
Description: PGP signature


Re: [gentoo-user] Libsld, what gives?

2022-11-16 Thread Andreas Fink
On Wed, 16 Nov 2022 02:11:18 -0500
Alan Grimes  wrote:

> I'm jackhammering the system now and I'm getting about 50% error spew
I would suggest a different tool than a jackhammer to fix the problems.