[gentoo-dev] last ritest for dev-java/saxon-bin

[gentoo-dev] last ritest for dev-java/saxon-bin

[Petteri Räty]

[-- Attachment #1: Type: text/plain, Size: 370 bytes --]

dev-java/saxon-bin is going to be removed from the tree as soon as the
from sources version dev-java/saxon gets marked stable on arches that
have saxon-bin stable. I will add a package move and a revision bump so
that user will have a smooth upgrade. I will also adjust app-text/jing
to work with the from sources version.

Regards,
Petteri Räty (Betelgeuse)


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 252 bytes --]

Re: [gentoo-dev] last ritest for dev-java/saxon-bin

[Alec Warner]

Enrico Weigelt wrote:
> * Petteri R?ty <betelgeuse@gentoo.org> schrieb:
>> dev-java/saxon-bin is going to be removed from the tree as soon as the
>> from sources version dev-java/saxon gets marked stable on arches that
>> have saxon-bin stable. I will add a package move and a revision bump so
>> that user will have a smooth upgrade. I will also adjust app-text/jing
>> to work with the from sources version.
> 
> Are such things also announced separately on some user list ?
> The problem I see: from the view of an common user (who's not 
> reading the dev lists) packages they're using suddenly disappear. 
> They don't have time to handle it properly.
> 
> There should be an additional flagging (ie. scheduled for masking
> or removal), maybe such flags could be put into an separate
> database (ie. some textfile available via web or rsync).
> 

Things scheduled for removal (by treecleaners, and I hope by many) enter
pmask prior to removal; this is really the current system of notification.

> An tool could check if the current system has some of those
> scheduled packages installed.
> 
> 
> cu

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] last ritest for dev-java/saxon-bin

[Enrico Weigelt]

* Petteri R?ty <betelgeuse@gentoo.org> schrieb:
> dev-java/saxon-bin is going to be removed from the tree as soon as the
> from sources version dev-java/saxon gets marked stable on arches that
> have saxon-bin stable. I will add a package move and a revision bump so
> that user will have a smooth upgrade. I will also adjust app-text/jing
> to work with the from sources version.

Are such things also announced separately on some user list ?
The problem I see: from the view of an common user (who's not 
reading the dev lists) packages they're using suddenly disappear. 
They don't have time to handle it properly.

There should be an additional flagging (ie. scheduled for masking
or removal), maybe such flags could be put into an separate
database (ie. some textfile available via web or rsync).

An tool could check if the current system has some of those
scheduled packages installed.


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: last ritest for dev-java/saxon-bin

[Duncan]

Alec Warner <antarus@gentoo.org> posted 44D0F2E7.3060409@gentoo.org,
excerpted below, on  Wed, 02 Aug 2006 18:45:59 +0000:

> Enrico Weigelt wrote:
>> * Petteri R?ty <betelgeuse@gentoo.org> schrieb:
>>> dev-java/saxon-bin is going to be removed from the tree as soon as the
>>> from sources version dev-java/saxon gets marked stable on arches that
>>> have saxon-bin stable. I will add a package move and a revision bump so
>>> that user will have a smooth upgrade. I will also adjust app-text/jing
>>> to work with the from sources version.
>> 
>> Are such things also announced separately on some user list ?
>> The problem I see: from the view of an common user (who's not 
>> reading the dev lists) packages they're using suddenly disappear. 
>> They don't have time to handle it properly.
>> 
>> There should be an additional flagging (ie. scheduled for masking
>> or removal), maybe such flags could be put into an separate
>> database (ie. some textfile available via web or rsync).
>> 
> 
> Things scheduled for removal (by treecleaners, and I hope by many) enter
> pmask prior to removal; this is really the current system of notification.

Just to add... 30 days is the usual package mask.  As AW states, the idea
of the mask is so users will have time after masking to save the masked
ebuild to their overlay and/or file a bug saying they use it, as well as
that it's a wake-up call for any devs, if they want it saved, to get busy
and take over maintenance and start fixing the bugs that having remained
unfixed for so long, got the package masked in the first place.

If a user isn't syncing and updating at least once a month, well... the
ebuilds and related files remain available from viewCVS, which is open to
the public, so they can still be retrieved and stuck in an overlay, if
necessary, and Gentoo devs can of course fix the bugs and return the
ebuild to the tree, if they believe it's worth it.  The files aren't gone
without a trace, forever.

All that said, IMO these qualify for "announcements" as much as the GLSAs
normally inhabiting the announce list do.  That list isn't "glsa", it's
"announce" and these are announcements that could affect a number of
users, so IMO they belong there as well as here.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: last ritest for dev-java/saxon-bin

[Alec Warner]

Duncan wrote:
> Alec Warner <antarus@gentoo.org> posted 44D0F2E7.3060409@gentoo.org,
> excerpted below, on  Wed, 02 Aug 2006 18:45:59 +0000:
> 
> 
> All that said, IMO these qualify for "announcements" as much as the GLSAs
> normally inhabiting the announce list do.  That list isn't "glsa", it's
> "announce" and these are announcements that could affect a number of
> users, so IMO they belong there as well as here.
> 

Too bad posted stuff to announce is a huge pita :)
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: last ritest for dev-java/saxon-bin

[Michael Weyershäuser]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Duncan wrote:
> If a user isn't syncing and updating at least once a month, well... the
> ebuilds and related files remain available from viewCVS, which is open to
> the public, so they can still be retrieved and stuck in an overlay, if
> necessary, and Gentoo devs can of course fix the bugs and return the
> ebuild to the tree, if they believe it's worth it.  The files aren't gone
> without a trace, forever.

Not to forget that there is a copy of each ebuild installed on your
system in /var/db/pkg/, so even if you sleep for several months and
miss the removal you can still grab it from there for your overlay.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE0bh66q4f+IV6B/wRAsrXAJ45pvOVbbmiGAoF7N9/uUsy9LhpuwCfcL8r
NqFx5NQb9V0KbIOl+B95T+U=
=iFGd
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Masking practics

[Alec Warner]

Enrico Weigelt wrote:
> * Duncan <1i5t5.duncan@cox.net> schrieb:
> 
> <snip>
> 
>> How can it be too late?  You sync during that 30 days it's masked, do an
>> emerge --pretend --update world, and see that it's going to downgrade. 
> 
> How do I see it (without actually having the version numbers in mind?)
> 
> Why I need is simply an command line, which just tells me, which 
> packages are now re-masked and which ones are scheduled for removal.
> 
> 
> cu

I can't really imagine a way to gather that data, aside from us
documenting it somewhere, and that won't happen unless we magically gain
more developers.
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Masking practics

[Enrico Weigelt]

* Duncan <1i5t5.duncan@cox.net> schrieb:

<snip>

> Just to add... 30 days is the usual package mask.  As AW states, 
> the idea of the mask is so users will have time after masking to 
> save the masked ebuild to their overlay and/or file a bug saying 

but then it may be too late. 
I just did a test, moved by package.mask away and called 
emerge --pretend --update world .
It would downgrade lots of packages, ie. bugzilla to an older version 
not yet supporting postgresql and so immediately make many services
(ie. all my bugzilla installations) unusable.

Re-masking must be handled very carefully. At least portage has to
know how some package was masked at installation time and shout out
an big-fat warning, if an mask was added later.


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Masking practics

[Duncan]

Enrico Weigelt <weigelt@metux.de> posted
20060803160326.GC19696@nibiru.local, excerpted below, on  Thu, 03 Aug 2006
18:03:26 +0200:

> * Duncan <1i5t5.duncan@cox.net> schrieb:
> 
> <snip>
> 
>> Just to add... 30 days is the usual package mask.  As AW states, 
>> the idea of the mask is so users will have time after masking to 
>> save the masked ebuild to their overlay and/or file a bug saying 
> 
> but then it may be too late. 
> I just did a test, moved by package.mask away and called 
> emerge --pretend --update world .
> It would downgrade lots of packages, ie. bugzilla to an older version 
> not yet supporting postgresql and so immediately make many services
> (ie. all my bugzilla installations) unusable.
> 
> Re-masking must be handled very carefully. At least portage has to
> know how some package was masked at installation time and shout out
> an big-fat warning, if an mask was added later.

How can it be too late?  You sync during that 30 days it's masked, do an
emerge --pretend --update world, and see that it's going to downgrade. 
You check and the package is still there, so it's obviously a package.mask
thing.  You check package.mask and sure enough, there's a note in there
saying that it's being removed.  You decide you want to keep it, and move
the still intact ebuild to your overlay and add an appropriate entry to
your package.unmask.

As others have pointed out, even if it's not in the tree any more, it's in
/var/db/pkg.  Even if for some reason insanity strikes and you don't do an
emerge --pretend before you do the real emerge --update, and it's gone out
of both the tree and /var/db/pkg, if you are running FEATURES=buildpkg (a
/great/ way to avoid remerging, and have customized binary backups
available to reinstall if necessary), you can get it from there (the
ebuild is tacked onto the end of the tarball, thus the warning about
additional data after the end of the tarball if you open it using tar or
whatever, I've rescued or compared install ebuilds to current tree ebuilds
from there a couple times).  Even if all local copies, including backups,
are wiped, the ebuilds and related files are in the viewCVS attic, as long
as the raw sources are still available somewhere.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Masking practics

[Enrico Weigelt]

* Duncan <1i5t5.duncan@cox.net> schrieb:

<snip>

> How can it be too late?  You sync during that 30 days it's masked, do an
> emerge --pretend --update world, and see that it's going to downgrade. 

How do I see it (without actually having the version numbers in mind?)

Why I need is simply an command line, which just tells me, which 
packages are now re-masked and which ones are scheduled for removal.


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Masking practics

[Jakub Moc]

[-- Attachment #1: Type: text/plain, Size: 672 bytes --]

Alec Warner wrote:
> Enrico Weigelt wrote:
>> * Duncan <1i5t5.duncan@cox.net> schrieb:
>>
>> <snip>
>>
>>> How can it be too late?  You sync during that 30 days it's masked, do an
>>> emerge --pretend --update world, and see that it's going to downgrade. 
>> How do I see it (without actually having the version numbers in mind?)
>>
>> Why I need is simply an command line, which just tells me, which 
>> packages are now re-masked and which ones are scheduled for removal.

Erm, am I missing something?


# emerge -uDpv world

!!! Ebuilds for the following packages are either all
!!! masked or don't exist:
media-libs/tunepimp



-- 

jakub


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-dev] Re: Re: Masking practics

[Duncan]

Jakub Moc <jakub@gentoo.org> posted 44D25D79.2050704@gentoo.org, excerpted
below, on  Thu, 03 Aug 2006 22:32:57 +0200:

> Alec Warner wrote:
>> Enrico Weigelt wrote:
>>> * Duncan schrieb:
>>>
>>>> How can it be too late?  You sync during that 30 days it's masked, do
>>>> an emerge --pretend --update world, and see that it's going to
>>>> downgrade.
>>> How do I see it (without actually having the version numbers in mind?)
>>>
>>> Why I need is simply an command line, which just tells me, which
>>> packages are now re-masked and which ones are scheduled for removal.
> 
> Erm, am I missing something?
> 
> # emerge -uDpv world
> 
> !!! Ebuilds for the following packages are either all !!! masked or don't
> exist:
> media-libs/tunepimp

Exactly, altho I did make the slight mistake of saying downgrade, when we
are talking removing from the tree so it would be (as above) all masked,
not just a downgrade.  With all versions masked, the warning is more
verbose than simply a downgrade, including a mention of why (in this case
package.mask, with the package.mask entry and associated comment, with
portage 2.1.1-preX anyway).

<quote>
!!! All ebuilds that could satisfy "tunepimp" have been masked.
!!! One of the following masked packages is required to complete your request:
- media-libs/tunepimp-0.4.2 (masked by: package.mask)
# Diego Pettenò <flameeyes@gentoo.org> (24 Jul 2006)
# TunePimp has security problems (#140184) and isn't API stable yet
# scheduled for removal.

- media-libs/tunepimp-0.3.0 (masked by: package.mask)
- media-libs/tunepimp-0.4.0 (masked by: package.mask)
- media-libs/tunepimp-0.3.0-r1 (masked by: package.mask)

For more information, see MASKED PACKAGES section in the emerge man page or
refer to the Gentoo Handbook.
</quote>

So it tells you exactly which packages are available but masked, that
they're masked and by what (package.mask), and what the comment masking
them is -- scheduled for removal due to security bugs, with a bug number.

That seems like all the necessary info to me, including a bug to look up
for more information, if desired.  The "scheduled for removal" certainly
suggests to me that if I want to keep it, I better put it in my overlay
/now/.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Alec Warner]

Enrico Weigelt wrote:
> <big_snip />
> 
> My problem still seems unsolved (or did I miss something) ?
> 
> Lets say, if I've, installed foo-1.1, and it gets masked due
> some bug(s), but 1.0 isn't, I want to get informed with an big
> fat warning, *before* anything actually done, ie. 
> 
> [...]
> # WARNING: installed package foo-1.1 has been masked and would
> # be downgraded:
> # <masking comment ...>
> [...]
> 
> An fully-automatic downgrade should *never* downgrade anything. 
> This is too dangerous, because essential features can get lost.
> Again, my bugzilla example: assuming 2.22 will be unmasked some
> day and I installed it w/ postgres support. Now there are some 
> bugs found, but not fixed fast enough, so it gets masked. 
> I run an update w/o knowing that it downgrades, and my whole
> bugzilla hosting is suddenly broken.
> 
> Do you consider this as stability, seriously ?!
> 

I would call you a horrible administrator since this:
"I run an update w/o knowing that it downgrades"
should NEVER happen.

emerge -pv foo
[ebuild UD] cat/foo-currentversion [downgraded-version] <stuff>

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Enrico Weigelt]

<big_snip />

My problem still seems unsolved (or did I miss something) ?

Lets say, if I've, installed foo-1.1, and it gets masked due
some bug(s), but 1.0 isn't, I want to get informed with an big
fat warning, *before* anything actually done, ie. 

[...]
# WARNING: installed package foo-1.1 has been masked and would
# be downgraded:
# <masking comment ...>
[...]

An fully-automatic downgrade should *never* downgrade anything. 
This is too dangerous, because essential features can get lost.
Again, my bugzilla example: assuming 2.22 will be unmasked some
day and I installed it w/ postgres support. Now there are some 
bugs found, but not fixed fast enough, so it gets masked. 
I run an update w/o knowing that it downgrades, and my whole
bugzilla hosting is suddenly broken.

Do you consider this as stability, seriously ?!


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Edward Catmur]

On Mon, 2006-08-07 at 15:01 +0200, Enrico Weigelt wrote:
> Lets say, if I've, installed foo-1.1, and it gets masked due
> some bug(s), but 1.0 isn't, I want to get informed with an big
> fat warning, *before* anything actually done, ie. 
> 
> [...]
> # WARNING: installed package foo-1.1 has been masked and would
> # be downgraded:
> # <masking comment ...>
> [...]
> 
> An fully-automatic downgrade should *never* downgrade anything. 
> This is too dangerous, because essential features can get lost.
> Again, my bugzilla example: assuming 2.22 will be unmasked some
> day and I installed it w/ postgres support. Now there are some 
> bugs found, but not fixed fast enough, so it gets masked. 
> I run an update w/o knowing that it downgrades, and my whole
> bugzilla hosting is suddenly broken.

That would not happen. bugzilla is a webapp and as such is fully
SLOTted. Upgrades (and downgrades) are manual.

Ed

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Simon Stelling]

Enrico Weigelt wrote:
> <big_snip />
> 
> My problem still seems unsolved (or did I miss something) ?
> 
> Lets say, if I've, installed foo-1.1, and it gets masked due
> some bug(s), but 1.0 isn't, I want to get informed with an big
> fat warning, *before* anything actually done, ie. 
> 
> [...]
> # WARNING: installed package foo-1.1 has been masked and would
> # be downgraded:
> # <masking comment ...>
> [...]
> 
> An fully-automatic downgrade should *never* downgrade anything. 
> This is too dangerous, because essential features can get lost.
> Again, my bugzilla example: assuming 2.22 will be unmasked some
> day and I installed it w/ postgres support. Now there are some 
> bugs found, but not fixed fast enough, so it gets masked. 
> I run an update w/o knowing that it downgrades, and my whole
> bugzilla hosting is suddenly broken.
> 
> Do you consider this as stability, seriously ?!

If your bugzilla hosting breaks with lower versions, then the ebuild contains a 
RDEPEND="postgres? ( >=dev-db/postgresql-2.22 )". Now if >=postgresql-2.22 gets 
masked, portage will bail out with an error because it can't find a valid 
dependency tree. This will cause the comment above the masking line in p.mask to 
be shown. You can then decide whether the breakage affects you or not and 
depending on that unmask it locally or remove your bugzilla installation.

If there is a bugzilla-ebuild which works with <postgresql-2.22, it will be 
downgraded too, leaving you with a working bugzilla.

I can't quite see the massive problem.

-- 
Kind Regards,

Simon Stelling
Gentoo/AMD64 Developer
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Re: Re: Masking practics

[Duncan]

Enrico Weigelt <weigelt@metux.de> posted
20060807130156.GE25236@nibiru.local, excerpted below, on  Mon, 07 Aug 2006
15:01:57 +0200:

> My problem still seems unsolved (or did I miss something) ?

You missed something.

> Lets say, if I've, installed foo-1.1, and it gets masked due some bug(s),
> but 1.0 isn't, I want to get informed with an big fat warning, *before*
> anything actually done, ie.
> 
> [...]
> # WARNING: installed package foo-1.1 has been masked and would # be
> downgraded:
> # <masking comment ...>
> [...]

That's precisely what emerge --pretend --verbose covers.  Or, if you want
the display with a question to continue or not, use --ask instead of
--verbose.

A good Gentoo user (read that as a good system administrator, because
that's exactly what a Gentoo distribution user is in this context) will
never run a straight "fully automatic" upgrade/downgrade, without knowing
exactly what's going to be done, because that's foolhardy, as you
correctly point out.  They will always know what to expect, because they
will have either used --pretend first, or will use --ask as a matter of
course.

Are you actually suggesting that you run emerge --update /blind/,
/without/ having viewed the --pretend (or --ask) output to see what it's
going to actually do, first?  No /wonder/ you are having problems if so! 
That's why the --pretend and --ask switches are /there/!  Use them for
what they are intended for, and you'll no longer be troubled by downgrades
without warning.  Only /after/ you are comfortable that the command will
do what you expect/want, do you run it without the --pretend, or say yes to
the --ask.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Joshua Nichols]

Duncan wrote:
> That's precisely what emerge --pretend --verbose covers.  Or, if you want
> the display with a question to continue or not, use --ask instead of
> --verbose.
>   
I'm pretty sure you mean to use --ask instead of --pretend, not --verbose.

-- 
Joshua Nichols
Gentoo/Java - Project Lead

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Enrico Weigelt]

* Alec Warner <antarus@gentoo.org> schrieb:

<snip>

> I would call you a horrible administrator since this:
> "I run an update w/o knowing that it downgrades"
> should NEVER happen.
> 
> emerge -pv foo
> [ebuild UD] cat/foo-currentversion [downgraded-version] <stuff>

Great. I have to explicitly compare the versions on each package.

Not actually an great help.


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Enrico Weigelt]

* Duncan <1i5t5.duncan@cox.net> schrieb:

<snip>

> > # WARNING: installed package foo-1.1 has been masked and would # be
> > downgraded:
> > # <masking comment ...>
> > [...]
> 
> That's precisely what emerge --pretend --verbose covers.  Or, if you want
> the display with a question to continue or not, use --ask instead of
> --verbose.

Ah, emerge actually tells you that it's going to *downgrade*
(not just printing out two version numbers and letting the user 
compare each packege's version number for its own) ?

> A good Gentoo user (read that as a good system administrator, because
> that's exactly what a Gentoo distribution user is in this context) will
> never run a straight "fully automatic" upgrade/downgrade, without knowing
> exactly what's going to be done, because that's foolhardy, as you
> correctly point out.  They will always know what to expect, because they
> will have either used --pretend first, or will use --ask as a matter of
> course.

Of course I always use -p, but having to look at each package's 
version numbers if it someday may want to downgrade. For such number
comparison, humans are not the best processor, the computer is 
far much better for it.

Why can't emerge just print out an fat warning if its going to
downgrade ? Would save people from much, much trouble.


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Steve Dibb]

Enrico Weigelt wrote:
> Why can't emerge just print out an fat warning if its going to
> downgrade ? Would save people from much, much trouble.
>   
# emerge -pv =coreutils-5.2.1-r7

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

Calculating dependencies... done!
[ebuild     UD] sys-apps/coreutils-5.2.1-r7 [5.94-r1] USE="-acl -build 
-nls -static" 0 kB

Total size of downloads: 0 kB

See the little "D"?

It's not big, it's not fat, but it's warning you. :)

Steve
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Enrico Weigelt]

* Steve Dibb <beandog@gentoo.org> schrieb:

<snip>

> [ebuild     UD] sys-apps/coreutils-5.2.1-r7 [5.94-r1] USE="-acl -build 
> -nls -static" 0 kB
> 
> Total size of downloads: 0 kB
> 
> See the little "D"?
> 
> It's not big, it's not fat, but it's warning you. :)

Not actually an eye-catching.

To be fair, do *you* actually look through *all* the emerge 
output if there's any "D" flag, without the risk of overlooking
it someday ?

Why can't emerge shout out some more verbose text ? Something
really eye-catching ?


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Steve Dibb]

Enrico Weigelt wrote:
> * Steve Dibb <beandog@gentoo.org> schrieb:
>
> <snip>
>
>   
>> [ebuild     UD] sys-apps/coreutils-5.2.1-r7 [5.94-r1] USE="-acl -build 
>> -nls -static" 0 kB
>>
>> Total size of downloads: 0 kB
>>
>> See the little "D"?
>>
>> It's not big, it's not fat, but it's warning you. :)
>>     
>
> Not actually an eye-catching.
>   
No, but it's going to be on every ebuild that's downgrading.  They're in 
*color* too. Whee! :D
> To be fair, do *you* actually look through *all* the emerge 
> output if there's any "D" flag, without the risk of overlooking
> it someday ?
>   
Yes.  I do. :)

emerge -puvN world | less
> Why can't emerge shout out some more verbose text ? Something
> really eye-catching ?
>
>   
Well the devs could expand on FEATURES="candy" or whatever it's called.  
Feel free to submit a patch.  I don't know how to code it either, or I 
would.

Steve
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Steev Klimaszewski]

Enrico Weigelt wrote:

> Not actually an eye-catching.

Ummm, D, as opposed to U... Yeah, that catches my eye.  I am weird like
that though.

> 
> To be fair, do *you* actually look through *all* the emerge 
> output if there's any "D" flag, without the risk of overlooking
> it someday ?

Yes, honestly, I DO look at it - I wouldn't be much of a system
administrator if I didn't.  I *always* *always* run emerge -uDpv world
when I do an update world - and not only that - but I *always* *always*
run emerge -av packagename/world - I never ever blindly run an emerge -
even if it is the same one I just ran a -p on.  I dunno, guess I am just
careful like that.  Obviously, not everyone wants to be vigilant - but
if you aren't - then you don't get to whine about it when it breaks.....
well actually I guess you do...

> 
> Why can't emerge shout out some more verbose text ? Something
> really eye-catching ?
> 

Well, why not look into color.map - and as has been mentioned in this or
other flags, read the finely written documentation - in fact, I know of
other distros who use OUR documentation when pointing out how to do
things to others on THEIR distros.  Documentation isn't exactly written
because the doc writers were bored.  It was written so things aren't
constantly repeated as have been over and over and over...infinity.
Seriously, stop trying to be lazy and be a responsible system admin.

> cu

No, cu!
Steev
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Richard Fish]

On 8/7/06, Enrico Weigelt <weigelt@metux.de> wrote:
> To be fair, do *you* actually look through *all* the emerge
> output if there's any "D" flag, without the risk of overlooking
> it someday ?

Um, sorry, but users *should* be looking at the output of --pretend to
get an idea of what portage wants to do to their systems.  And not
just for the 'D' either.

> Why can't emerge shout out some more verbose text ? Something
> really eye-catching ?

Please don't.  Portage already has enough messages that try to be "eye
catching".  I've only got two eyes after all...

-Richard
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Enrico Weigelt]

* Steev Klimaszewski <steev@gentoo.org> schrieb:

<snip>

> Obviously, not everyone wants to be vigilant - but
> if you aren't - then you don't get to whine about it when it breaks.....
> well actually I guess you do...

So, you don't have any intention to help people who don't have 
such an eagle-eye on single chars, even it could be easy ?

<snip> 

> > Why can't emerge shout out some more verbose text ? Something
> > really eye-catching ?
> > 
> 
> Well, why not look into color.map - and as has been mentioned in this or

Would be a first start. But doesnt help much in emails or monochrome 
terminals. Even piping to less kills colors. 

So I'll probably have no other chance than writing a frontend 
to emerge, parsing its output - hoping the output syntax remains
the same for an sufficiant time :(


cu
-- 
---------------------------------------------------------------------
 Enrico Weigelt    ==   metux IT service - http://www.metux.de/
---------------------------------------------------------------------
 Please visit the OpenSource QM Taskforce:
 	http://wiki.metux.de/public/OpenSource_QM_Taskforce
 Patches / Fixes for a lot dozens of packages in dozens of versions:
	http://patches.metux.de/
---------------------------------------------------------------------
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Michael Weyershäuser]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Enrico Weigelt wrote:
> To be fair, do *you* actually look through *all* the emerge 
> output if there's any "D" flag, without the risk of overlooking
> it someday ?

Yes.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE16tG6q4f+IV6B/wRApCbAJoC6r1YXAsp1SEcrVz0ODCQ1ngqVACfcrvg
X0acTdrIGo9FN2l8aD7YmDo=
=bsrK
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Thomas Cort]

[-- Attachment #1: Type: text/plain, Size: 620 bytes --]

On Mon, 7 Aug 2006 23:01:45 +0200
Enrico Weigelt <weigelt@metux.de> wrote:

> > Obviously, not everyone wants to be vigilant - but
> > if you aren't - then you don't get to whine about it when it breaks.....
> > well actually I guess you do...
> 
> So, you don't have any intention to help people who don't have 
> such an eagle-eye on single chars, even it could be easy ?

Users can help themselves. The D is in the exact same place (right
before the ']') and in a different color (dark blue). You can
use grep to find the downgrades: 

   emerge -uD world -vp | grep D\]

or

   grep D\] email.txt

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Benno Schulenberg]

Enrico Weigelt wrote:
> * Steve Dibb <beandog@gentoo.org> schrieb:
> > See the little "D"?
> >
> > It's not big, it's not fat, but it's warning you. :)
>
> Not actually an eye-catching.
>
> To be fair, do *you* actually look through *all* the emerge
> output if there's any "D" flag, without the risk of overlooking
> it someday ?

You don't have to look at it.  Write your own little emerge wrapper 
script, let it amongst other things grep for "UD", and let it howl 
when it finds it.

Benno
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Masking practics

[Alexandre Buisse]

[-- Attachment #1: Type: text/plain, Size: 558 bytes --]

On Mon, Aug  7, 2006 at 22:18:35 +0200, Enrico Weigelt wrote:

> * Alec Warner <antarus@gentoo.org> schrieb:
> 
> <snip>
> 
> > I would call you a horrible administrator since this:
> > "I run an update w/o knowing that it downgrades"
> > should NEVER happen.
> > 
> > emerge -pv foo
> > [ebuild UD] cat/foo-currentversion [downgraded-version] <stuff>
> 
> Great. I have to explicitly compare the versions on each package.

And what about the big blue 'D'?

/Alexandre
-- 
Hi, I'm a .signature virus! Please copy me in your ~/.signature.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Curtis Napier]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: MD5

Enrico Weigelt wrote:
> * Steve Dibb <beandog@gentoo.org> schrieb:
> 
> <snip>
> 
>> [ebuild     UD] sys-apps/coreutils-5.2.1-r7 [5.94-r1] USE="-acl -build 
>> -nls -static" 0 kB
>>
>> Total size of downloads: 0 kB
>>
>> See the little "D"?
>>
>> It's not big, it's not fat, but it's warning you. :)
> 
> Not actually an eye-catching.
> 
> To be fair, do *you* actually look through *all* the emerge 
> output if there's any "D" flag, without the risk of overlooking
> it someday ?

Yes. I look at the entire output of emerge -pv *EVERY* single time. Not
doing so would be insane and I would get what I deserve... a broken system.

> 
> Why can't emerge shout out some more verbose text ? Something
> really eye-catching ?
> 

How much more eye-catching does it need to be? It's color coded in a
column along the left hand side of the screen in a nice format that is
easily parse-able and grep-able. You don't even have to scan left to
right when reading it!

It's obvious to me that you haven't been a long time Gentoo user. All of
the issues you have brought up in the several different threads are all
mistakes that n00bs make when changing their Linux Admin paradigm to the
Gentoo Way(tm).

We see a thousand questions a year just like yours in the forums from
new Gentoo users. Every single one of those users comes back in a few
months and recants once they get a good understanding of what they are
doing and how powerful Gentoo actually is.

Once you use it for a while you will find that the current system (USE,
SLOTs, etc...) is *extremely* powerful and flexible from just about any
POV. No binary distro can beat it.

- --Curtis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iQEVAwUBRNguEUb8Q0uRCeTQAQFmyAgAqUxtDU2HP0UPunU9yvvxL4ab3LqYsOF5
zHqn8gqZyU0zMoY57JkwulyjQfRVMg9VuZ/D9+Rel6eqsJ1GXVRbyyOx9mK80hnP
mYbT7q1kuH3pwSXzmmNn8MN5gKvTVJUVSfmdYn/ZhiXmnMTQOjCoLfoGSS4vfdpK
R82WQWJ/ozw4um6kVyci1M3JVdQSmaPxFDgUrU6Y9+LzbnzBcalW8ns1oVIRL6Za
0XGfZcDRtVIlGzUKcimhRxzucaGu3Ua8ySIC2R+y0WnSBUAUI6qiVQ0A5TUYiynZ
PFB9Zhwr6LDk4Itkxh1UfL70igLWv61GQxpML+Bysx3eMo2EY9JYlQ==
=5OM1
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Patrick McLean]

Enrico Weigelt wrote:
> So I'll probably have no other chance than writing a frontend 
> to emerge, parsing its output - hoping the output syntax remains
> the same for an sufficiant time :(
> 

[ebuild   R   ] sys-fs/cryptsetup-luks-1.0.3-r2
[ebuild     U ] x11-terms/rxvt-unicode-7.9 [7.8-r1]
[ebuild     U ] sci-chemistry/gromacs-3.3.1 [3.3]
[ebuild     UD] app-foo/bar-1.0.2 [1.1.0]
[ebuild     U ] app-text/evince-0.5.5 [0.5.4]

Why do you think emerge has columnar output like this, notice how the D
is in a different column than anything else, it makes it pretty easy to
spot, if you are too lazy to at least scan the output of emerge -p
before actually running the emerge, don't complain when you break your
system.

We generally assume that our users have some intelligence, we don't put
big fat warnings on everything that might break. We give the users all
the information and let them decide if its going to break. Gentoo never
was and never will be idiot-proof. We do not, and never will try to
protect people from their own stupidity.
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[FieldySnuts]

On Tuesday 08 August 2006 10:41, Patrick McLean wrote:
> We generally assume that our users have some intelligence, we don't put
> big fat warnings on everything that might break. We give the users all
> the information and let them decide if its going to break. Gentoo never
> was and never will be idiot-proof. We do not, and never will try to
> protect people from their own stupidity.

As an end user and also an administrator, I am very pleased to see this laid 
out so clearly. I mean, I knew it, but it seems like it needs to be yelled 
once in a while...

Thanks Patrick.
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Joshua Nichols]

Patrick McLean wrote:
> Enrico Weigelt wrote:
> [ebuild   R   ] sys-fs/cryptsetup-luks-1.0.3-r2
> [ebuild     U ] x11-terms/rxvt-unicode-7.9 [7.8-r1]
> [ebuild     U ] sci-chemistry/gromacs-3.3.1 [3.3]
> [ebuild     UD] app-foo/bar-1.0.2 [1.1.0]
> [ebuild     U ] app-text/evince-0.5.5 [0.5.4]
>
> Why do you think emerge has columnar output like this, notice how the D
> is in a different column than anything else, it makes it pretty easy to
> spot, if you are too lazy to at least scan the output of emerge -p
> before actually running the emerge, don't complain when you break your
> system.
>   
While it is columnar, the D is in a dark blue font. If you happen to be
using a dark background, there is extremely little contrast. Perhaps it
could be a different color that would stick out in both light and dark
backgrounds?

Also something that has always bugged me... isn't the U supposed to be
for upgrade and the D for downgrade? In this case, it would make sense
to only show the D when downgrades will occur, and not both, wouldn't it?

Not that I have ever had a problem with downgrades (that I remember),
but I think these two tidbits would probably be an overall improvement.

-- 
Joshua Nichols
Gentoo/Java - Project Lead

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Simon Stelling]

Joshua Nichols wrote:
> While it is columnar, the D is in a dark blue font. If you happen to be
> using a dark background, there is extremely little contrast. Perhaps it
> could be a different color that would stick out in both light and dark
> backgrounds?

There is color-mapping support in portage 2.1, i guess it could be done with that.

> Also something that has always bugged me... isn't the U supposed to be
> for upgrade and the D for downgrade? In this case, it would make sense
> to only show the D when downgrades will occur, and not both, wouldn't it?

I always interpreted it as 'update' instead of 'upgrade'.

-- 
Kind Regards,

Simon Stelling
Gentoo/AMD64 Developer
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Paul de Vrieze]

[-- Attachment #1: Type: text/plain, Size: 839 bytes --]

On Tuesday 08 August 2006 22:36, Enrico Weigelt wrote:
> > As an end user and also an administrator, I am very pleased to see
> > this laid out so clearly. I mean, I knew it, but it seems like it
> > needs to be yelled once in a while...
>
> hmm, now that I know of these flags, I can take a look at them
> (although its not very comfortable having to look at those details).
>
> Yes, I could have read the whole docs to learn about this, but I never
> expected such things. IMHO many people may get into such pitfalls.

You're probably one of the new users of the installer. I knew that we 
shouldn't do that. The whole point of gentoo having (a reputation of having) 
good documentation is because people need it.

Paul

-- 
Paul de Vrieze
Gentoo Developer
Mail: pauldv@gentoo.org
Homepage: http://www.devrieze.net

[-- Attachment #2: Type: application/pgp-signature, Size: 200 bytes --]

Re: [gentoo-dev] Re: Re: Re: Masking practics

[Chris Gianelloni]

[-- Attachment #1: Type: text/plain, Size: 1338 bytes --]

On Thu, 2006-08-10 at 14:24 +0200, Paul de Vrieze wrote:
> On Tuesday 08 August 2006 22:36, Enrico Weigelt wrote:
> > > As an end user and also an administrator, I am very pleased to see
> > > this laid out so clearly. I mean, I knew it, but it seems like it
> > > needs to be yelled once in a while...
> >
> > hmm, now that I know of these flags, I can take a look at them
> > (although its not very comfortable having to look at those details).
> >
> > Yes, I could have read the whole docs to learn about this, but I never
> > expected such things. IMHO many people may get into such pitfalls.
> 
> You're probably one of the new users of the installer. I knew that we 
> shouldn't do that. The whole point of gentoo having (a reputation of having) 
> good documentation is because people need it.

Alright, you seriously need to quit this sort of FUD.  The Installer
really doesn't make it that much easier to install, it simply makes it
quicker.  If you don't believe me, try following gli-bugs on bugzilla or
hanging out in #gentoo-installer for a while.  Trying to blame one
user's refusal to read documentation on a project that you seem to know
little about does nothing to serve Gentoo.

-- 
Chris Gianelloni
Release Engineering - Strategic Lead
x86 Architecture Team
Games - Developer
Gentoo Linux

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]