public inbox for gentoo-user@lists.gentoo.org
 help / color / mirror / Atom feed
From: gottlieb@nyu.edu
To: gentoo-user@lists.gentoo.org
Subject: Re: [gentoo-user] updating ALL packages
Date: Fri, 25 Jul 2014 10:42:55 -0400	[thread overview]
Message-ID: <87ha25jy8g.fsf@nyu.edu> (raw)
In-Reply-To: <53D1D94C.5020008@gmail.com> (Alan McKinnon's message of "Fri, 25 Jul 2014 06:13:00 +0200")

On Fri, Jul 25 2014, Alan McKinnon wrote:

> On 25/07/2014 03:51, gottlieb@nyu.edu wrote:
>> On Thu, Jul 24 2014, Neil Bothwick wrote:
>> 
>>> On Wed, 23 Jul 2014 09:24:44 -0400, gottlieb@nyu.edu wrote:
>>>
>>>> I had mistakenly thought this would update all packages not at the
>>>> latest version (subject to package.accept_keywords, package.mask, ...).
>>>
>>> It only updates runtime dependencies, you need --with-bdeps=y to update
>>> all dependencies.
>> 
>> Thank you and michael for this point.
>> 
>>> However, the default is no for a good reason, there's no need to
>>> update build time deps once the package is installed.
>> 
>> I see.
>> 
>>>> I now realize that it only does this for the packages in world and then
>>>> follows the dependency tree.  So if package A in world is up to date, A
>>>> depends of B, and a new version of B appears, B will not be updated.
>>>>
>>>> As a result eix-test-obsolete finds that I have packages installed that
>>>> are no longer in the database.
>>>
>>> That shouldn't happen. If an installed package is removed for the tree,
>>> portage should either install the highest version that matches your
>>> settings or print a warning.
>> 
>> I am not sure if you consider the message from eix-test-obsolete
>> as the message from portage.
>> 
>> eix-test-obsolete prints (among other things)
>> 
>> Installed packages with a version not in the database (or masked):
>> [lines omitted]
>> [U] virtual/perl-CPAN-Meta-Requirements (2.125.0@10/29/2013 ->
>> (~)2.125.0-r1): Virtual for CPAN-Meta-Requirements
>> 
>> eix virtual/perl-CPAN-Meta-Requirements   prints
>> 
>> [U] virtual/perl-CPAN-Meta-Requirements
>>      Available versions:  2.122.0-r2 (~)2.125.0-r1
>>      Installed versions:  2.125.0(09:25:35 PM 10/29/2013)
>> 
>> /etc/portage/package.accept_keywords/goingstable contains
>> ~virtual/perl-CPAN-Meta-Requirements-2.125.0
>> 
>> I thought this would be updated to  2.125.0-r1 but
>> my update world (withOUT bdeps=y) says nothing to merge
>> and prints no error or warning
>
> That is correct. The package is needed to build stuff and nothing in the
> current list of packages to be built needs the package to do it.
> Should you sometime update a package that does depend on perl-CPAN-Meta
> to be built, then perl-CPAN-Meta will then be updated
>
>
>
>>>> I could do 
>>>>
>>>> emerge --update the-2-dozen-such-packages
>>>>
>>>> Is that wise?
>>>
>>> No, as it will add them to world (this behaviour of -u appears to vary
>>> depending on portage version, wind direction and sunspot activity). Use
>>> --oneshot.
>> 
>> Understood.  And I remember the discussion on the list about the meaning
>> of -u.
>> 
>> emerge -u -1 virtual/perl-CPAN-Meta-Requirements
>> reveals what is probably my real problem
>> 
>> [ebuild     U ~] virtual/perl-CPAN-Meta-Requirements-2.125.0-r1 [2.125.0] 0 kB
>> [nomerge       ]  perl-core/CPAN-Meta-Requirements-2.125.0 
>> [ebuild     UD ]   virtual/perl-version-0.990.100 [0.990.400] 0 kB
>> [ebuild     UD ]    perl-core/version-0.990.100 [0.990.400] 105 kB
>> 
>> upgrading virtual/perl-CPAN-Meta-Requirements entails downgrading two
>> other perl packages (or bumping their version in goingstable, which I
>> remember you suggest).
>> 
>> I am going away for 2 weeks, but when I return I will look carefully at
>> the (mostly perl) files that eix-test-obsolete complains about.  I am
>> hopeful that armed with emerge -u -1 and/or --with-bdeps=y I can remove
>> the warnings from eix-test-obsolete.
>
> Just do one world update with bdeps=y
>
> Portage will then update the packages that it has been skipping
>
Not quite that simple due to the bothwick
package.accept_keywords/goingstable.  I did a --pretend run and saw
several proposed downgrades (to packages required by the ones mentioned
in eix-test-obsolete).  Neil recommends that in these cases I update
goingstable to permit upgrades instead.

thanks for your interest and help
allan


  reply	other threads:[~2014-07-25 14:43 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-23 13:24 [gentoo-user] updating ALL packages gottlieb
2014-07-23 13:59 ` [gentoo-user] " gottlieb
2014-07-24 22:18   ` Michael Orlitzky
2014-07-24 22:57 ` [gentoo-user] " Neil Bothwick
2014-07-25  1:51   ` gottlieb
2014-07-25  4:13     ` Alan McKinnon
2014-07-25 14:42       ` gottlieb [this message]
2014-07-25  4:07 ` Alan McKinnon
2014-07-25 14:43   ` gottlieb

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87ha25jy8g.fsf@nyu.edu \
    --to=gottlieb@nyu.edu \
    --cc=gentoo-user@lists.gentoo.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox