Please stop useless removals (was: [gentoo-dev] last rites: games-arcade/bitefusion)

Please stop useless removals (was: [gentoo-dev] last rites: games-arcade/bitefusion)

[Vaeth]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2199 bytes --]


>>    # Upstream is dead and gone.
>>    # Masked for removal on 20130302
>
> Erm, so this is the _only_ reason - dead upstream?

++

Please, please, stop removing packages for no reason!
This happens now way too often:

app-dicts/ispell*
app-portage/epm
app-text/ispell
games-arcade/bitefusion
games-arcade/xboing
games-action/trackballs
games-emulation/xmame
...

These are just some of the previous examples which I remember
because I had to put them in my local overlay.

None of these removals alone was so valuable to me that I saw
a reason to step up, but the removals for no reasons accumulate
previously so much that I see the need to say something:

You are destroying the charme of gentoo by systematically
removing all these little tools and toys.  The availability
of a lot of software was once a strength of gentoo, so removing
these things is really bad, especially if it happens for no
real reason.

I was understanding if e.g. someting was removed which needs
the <gtk-2 or <qt-4 framework or something similar and had
a dead upstream. But just needing a small tool like imake (xboing)
or having open feature requestes (epm) or even nothing and
just dead upstream is IMHO really not a reason.

If something really does not compile anymore and nobody cares,
then remove keywords (or, for god's sake, mask it);
if something might theoretically become a security issue (xpdf)
then it should be masked.

But please do not throw things out of the tree unless
really necessary:

It does not hurt anybody to have such package in the tree,
but removing it - especially if upstream is dead - means
that the tarbalös will be removed from the mirrors and thus
nobody is able anymore to install it (even if he would care and
fix some minor issues) unless he had kept a copy on
his local machine (which will mean in the future that he can only
do it if he had used gentoo already many years ago and cared
during the time of the removal).

(If the resources are an argument: I am not speaking about monster
packages taking gigabytes of data - these might need to be
discussed separately - but mainly about reasonably sized packages
which even if summed up do not take much data).

Regards
Martin

Re: Please stop useless removals (was: [gentoo-dev] last rites: games-arcade/bitefusion)

[Alec Warner]

On Thu, Jan 31, 2013 at 11:36 PM, Vaeth
<vaeth@mathematik.uni-wuerzburg.de> wrote:
>
>>>    # Upstream is dead and gone.
>>>    # Masked for removal on 20130302
>>
>>
>> Erm, so this is the _only_ reason - dead upstream?
>
>
> ++
>
> Please, please, stop removing packages for no reason!
> This happens now way too often:
>
> app-dicts/ispell*
> app-portage/epm
> app-text/ispell
> games-arcade/bitefusion
> games-arcade/xboing
> games-action/trackballs
> games-emulation/xmame
> ...
>
> These are just some of the previous examples which I remember
> because I had to put them in my local overlay.
>
> None of these removals alone was so valuable to me that I saw
> a reason to step up, but the removals for no reasons accumulate
> previously so much that I see the need to say something:

If folks do not want to maintain it anymore, then it will be removed.
Feel free to contribute to Gentoo and maintain the packages.

>
> You are destroying the charme of gentoo by systematically
> removing all these little tools and toys.  The availability
> of a lot of software was once a strength of gentoo, so removing
> these things is really bad, especially if it happens for no
> real reason.

Gentoo is not a software archival service.

>
> I was understanding if e.g. someting was removed which needs
> the <gtk-2 or <qt-4 framework or something similar and had
> a dead upstream. But just needing a small tool like imake (xboing)
> or having open feature requestes (epm) or even nothing and
> just dead upstream is IMHO really not a reason.
>
> If something really does not compile anymore and nobody cares,
> then remove keywords (or, for god's sake, mask it);
> if something might theoretically become a security issue (xpdf)
> then it should be masked.
>
> But please do not throw things out of the tree unless
> really necessary:
>
> It does not hurt anybody to have such package in the tree,
> but removing it - especially if upstream is dead - means
> that the tarbalös will be removed from the mirrors and thus
> nobody is able anymore to install it (even if he would care and
> fix some minor issues) unless he had kept a copy on
> his local machine (which will mean in the future that he can only
> do it if he had used gentoo already many years ago and cared
> during the time of the removal).

Again I highly recommend archiving the software yourself; but I don't
think Gentoo should be doing it.

-A

>
> (If the resources are an argument: I am not speaking about monster
> packages taking gigabytes of data - these might need to be
> discussed separately - but mainly about reasonably sized packages
> which even if summed up do not take much data).
>
> Regards
> Martin

[gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Michael Weber]

On 02/01/2013 09:21 AM, Alec Warner wrote:
> On Thu, Jan 31, 2013 at 11:36 PM, Vaeth
> <vaeth@mathematik.uni-wuerzburg.de> wrote:
>>
>>>>    # Upstream is dead and gone.
>>>>    # Masked for removal on 20130302
>>>
>>>
>>> Erm, so this is the _only_ reason - dead upstream?
>>
> If folks do not want to maintain it anymore, then it will be removed.
> Feel free to contribute to Gentoo and maintain the packages.

Hereby done, becoming a dev is a big step for just one package a user
would keep.

Ihmo, what you call "upstream dead" is a kind of positive situation.

If the author has no longer time to contribute (we all have a real life)
then it's ok, no need to wipe his contribution from the face of the world.

If the software is just working as the author intendend, and it has no
major bugs, then there's no need to do further trivial releases just to
keep the disto maintainers busy.

If it's broken, uncompatible and nobody steps up, drop it, agreed.


>> You are destroying the charme of gentoo by systematically
>> removing all these little tools and toys.  The availability
>> of a lot of software was once a strength of gentoo, so removing
>> these things is really bad, especially if it happens for no
>> real reason.

We need to maintain a certain quality. Sheer mass does has no charm, if
nothing works. But I'd rather like to see gentoo as a broad selection of
tools, that build. maybe some really cool stuff nobody else has.

> Gentoo is not a software archival service.
>> I was understanding if e.g. someting was removed which needs
>> the <gtk-2 or <qt-4 framework or something similar and had
>> a dead upstream. But just needing a small tool like imake (xboing)
>> or having open feature requestes (epm) or even nothing and
>> just dead upstream is IMHO really not a reason.
>>
>> If something really does not compile anymore and nobody cares,
>> then remove keywords (or, for god's sake, mask it);
>> if something might theoretically become a security issue (xpdf)
>> then it should be masked.
>>
>> But please do not throw things out of the tree unless
>> really necessary:
>>
>> It does not hurt anybody to have such package in the tree,
>> but removing it - especially if upstream is dead - means
>> that the tarbalös will be removed from the mirrors and thus
>> nobody is able anymore to install it (even if he would care and
>> fix some minor issues) unless he had kept a copy on
>> his local machine (which will mean in the future that he can only
>> do it if he had used gentoo already many years ago and cared
>> during the time of the removal).
> 
> Again I highly recommend archiving the software yourself; but I don't
> think Gentoo should be doing it.

It costs resources:
 - distfiles and all their mirrors accumulate
 - emerge dependency calculation

If it's out-waged by increasing disc capacity and processor power is up
to discussion.

Last but not least, we have gattered some extra info besides the
tarballs, our precious ebuild scripts. Which is why I started my
involvement with Gentoo (maybe somebody should have told me about BSDs
tree before that).

As Martin said, tarballs get lost. I steal them from debian mirror on a
regular basis, maybe we should contribute ourselves.

PROPOSAL

  Let's create an overlay "frozen stuff" which contains all the
  software no longer developed with following features:

    Users showed interest in having them

    Web-presence to be picked up on Google search.
       (viewvc.cgi show dead is kinda hidden [1])

    Separate distfile mirror
       no need to stress our mirror peers
       make it a sepearate repo,
          feed by upstream and mirror://gentoo
       I can contribute the space/bandwith.

    Feedback/Bugs/Voting can be handled inside b.g.o
       no need for extra login,
       frozen-bugs can be auto-generated,
       whitelist [frozen]
       just like the sunrise tracker bugs.

BENEFIT

   User can choose whether or not layman -a frozen.

   Non-trivial ebuilds are preserved.

   Tarballs are preserved.

   Nobody gets hurt.

Comments?


[1] http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/

-- 
Michael Weber
Gentoo Developer
web: https://xmw.de/
mailto: Michael Weber <xmw@gentoo.org>

Re: [gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Sergey Popov]

[-- Attachment #1: Type: text/plain, Size: 491 bytes --]

01.02.2013 12:53, Michael Weber wrote:
> BENEFIT
> 
>    User can choose whether or not layman -a frozen.
> 
>    Non-trivial ebuilds are preserved.
> 
>    Tarballs are preserved.
> 
>    Nobody gets hurt.

Well, we can move such software to sunrise, can't we? But proposition of
splitted mirrors makes sense, cause quite often dead upstream means dead
links to original tarballs too.

-- 
Best regards, Sergey Popov
Gentoo Linux Developer
Desktop-effects project lead


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 554 bytes --]

Re: [gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Dennis Lan (dlan)]

[-- Attachment #1: Type: text/plain, Size: 5034 bytes --]

HI Michael:
   I can think of it's almost kind of a staging area, some package may be
partial broken(or partial functional),
but still useful for user.
   Generally speaking, It should be a good idea! The end users will benefit
a lot.

   Also if user show his interests, then he can report bug, send patch,
or step in to active maintain the package. Leave a opportunity to him...


Dennis


On Fri, Feb 1, 2013 at 4:53 PM, Michael Weber <xmw@gentoo.org> wrote:

> On 02/01/2013 09:21 AM, Alec Warner wrote:
> > On Thu, Jan 31, 2013 at 11:36 PM, Vaeth
> > <vaeth@mathematik.uni-wuerzburg.de> wrote:
> >>
> >>>>    # Upstream is dead and gone.
> >>>>    # Masked for removal on 20130302
> >>>
> >>>
> >>> Erm, so this is the _only_ reason - dead upstream?
> >>
> > If folks do not want to maintain it anymore, then it will be removed.
> > Feel free to contribute to Gentoo and maintain the packages.
>
> Hereby done, becoming a dev is a big step for just one package a user
> would keep.
>
> Ihmo, what you call "upstream dead" is a kind of positive situation.
>
> If the author has no longer time to contribute (we all have a real life)
> then it's ok, no need to wipe his contribution from the face of the world.
>
> If the software is just working as the author intendend, and it has no
> major bugs, then there's no need to do further trivial releases just to
> keep the disto maintainers busy.
>
> If it's broken, uncompatible and nobody steps up, drop it, agreed.
>
>
> >> You are destroying the charme of gentoo by systematically
> >> removing all these little tools and toys.  The availability
> >> of a lot of software was once a strength of gentoo, so removing
> >> these things is really bad, especially if it happens for no
> >> real reason.
>
> We need to maintain a certain quality. Sheer mass does has no charm, if
> nothing works. But I'd rather like to see gentoo as a broad selection of
> tools, that build. maybe some really cool stuff nobody else has.
>
> > Gentoo is not a software archival service.
> >> I was understanding if e.g. someting was removed which needs
> >> the <gtk-2 or <qt-4 framework or something similar and had
> >> a dead upstream. But just needing a small tool like imake (xboing)
> >> or having open feature requestes (epm) or even nothing and
> >> just dead upstream is IMHO really not a reason.
> >>
> >> If something really does not compile anymore and nobody cares,
> >> then remove keywords (or, for god's sake, mask it);
> >> if something might theoretically become a security issue (xpdf)
> >> then it should be masked.
> >>
> >> But please do not throw things out of the tree unless
> >> really necessary:
> >>
> >> It does not hurt anybody to have such package in the tree,
> >> but removing it - especially if upstream is dead - means
> >> that the tarbalös will be removed from the mirrors and thus
> >> nobody is able anymore to install it (even if he would care and
> >> fix some minor issues) unless he had kept a copy on
> >> his local machine (which will mean in the future that he can only
> >> do it if he had used gentoo already many years ago and cared
> >> during the time of the removal).
> >
> > Again I highly recommend archiving the software yourself; but I don't
> > think Gentoo should be doing it.
>
> It costs resources:
>  - distfiles and all their mirrors accumulate
>  - emerge dependency calculation
>
> If it's out-waged by increasing disc capacity and processor power is up
> to discussion.
>
> Last but not least, we have gattered some extra info besides the
> tarballs, our precious ebuild scripts. Which is why I started my
> involvement with Gentoo (maybe somebody should have told me about BSDs
> tree before that).
>
> As Martin said, tarballs get lost. I steal them from debian mirror on a
> regular basis, maybe we should contribute ourselves.
>
> PROPOSAL
>
>   Let's create an overlay "frozen stuff" which contains all the
>   software no longer developed with following features:
>
>     Users showed interest in having them
>
>     Web-presence to be picked up on Google search.
>        (viewvc.cgi show dead is kinda hidden [1])
>
>     Separate distfile mirror
>        no need to stress our mirror peers
>        make it a sepearate repo,
>           feed by upstream and mirror://gentoo
>        I can contribute the space/bandwith.
>
>     Feedback/Bugs/Voting can be handled inside b.g.o
>        no need for extra login,
>        frozen-bugs can be auto-generated,
>        whitelist [frozen]
>        just like the sunrise tracker bugs.
>
> BENEFIT
>
>    User can choose whether or not layman -a frozen.
>
>    Non-trivial ebuilds are preserved.
>
>    Tarballs are preserved.
>
>    Nobody gets hurt.



Comments?
>
>
> [1] http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/
>
> --
> Michael Weber
> Gentoo Developer
> web: https://xmw.de/
> mailto: Michael Weber <xmw@gentoo.org>
>
>

[-- Attachment #2: Type: text/html, Size: 6500 bytes --]

Re: [gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Michael Weber]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256


On 02/01/2013 10:35 AM, Dennis Lan (dlan) wrote:> HI Michael:
> I can think of it's almost kind of a staging area, some package
> may be partial broken(or partial functional), but still useful for
> user.

Please see [1] for the proposal of betagarden overlay, which might
grab attention by posting a project page, @sping *plz*

> Generally speaking, It should be a good idea! The end users will 
> benefit a lot.
thanks.

On 02/01/2013 10:30 AM, Sergey Popov wrote:
> Well, we can move such software to sunrise, can't we? But
> proposition of splitted mirrors makes sense, cause quite often dead
> upstream means dead links to original tarballs too.
Maybe betagarden/sunrise would benefit from mirror-coverage,
hosting situation is a recurring question on #-sunrise. Votes?

Sunrise commit access is limited to sunrise devs. And I see the _rise_
in context of software and devs.
I don't say sundown, cause for mentioned arguments, I just wanna have
functioning/maybe superseeded software around, regardless of it's
commit-frequency, author involvement or century of creation.

Again: We need to proceed as a contemporary distribution ("Does not
build with latest ~** gcc/" argument), but we can preserve our trail
for those who like.

The line between removed packages and obsoleted slots has to be drawn.

I'm in a tension between overlay scatter and providing an automated
time capsule (that certainly will mess up any of the aforementioned
repos).

[1]
http://archives.gentoo.org/gentoo-dev/msg_384ad55a02bf02154397f29d10a0f68e.xml

- -- 
Michael Weber
Gentoo Developer
web: https://xmw.de/
mailto: Michael Weber <xmw@gentoo.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iF4EAREIAAYFAlELnq4ACgkQknrdDGLu8JAY3gD/TOifKZZNqVb6VJkfp/VLGaGT
MZzWVOYVsPPAQi0B+voA/3D8afTh5TjxeWJvAKIZwIG6O/rwVrVBAI4YHgC4T59x
=bnDb
-----END PGP SIGNATURE-----

Re: Please stop useless removals (was: [gentoo-dev] last rites: games-arcade/bitefusion)

[Rich Freeman]

On Fri, Feb 1, 2013 at 3:21 AM, Alec Warner <antarus@gentoo.org> wrote:
> On Thu, Jan 31, 2013 at 11:36 PM, Vaeth
> <vaeth@mathematik.uni-wuerzburg.de> wrote:
>> Please, please, stop removing packages for no reason!
>> This happens now way too often:
>>
>
> If folks do not want to maintain it anymore, then it will be removed.
> Feel free to contribute to Gentoo and maintain the packages.
>

I think this makes sense, IFF there is something fundamentally wrong
with the package.

Being unmaintained in and of itself is not something fundamentally
wrong with the package.

Having a few open bugs is not either.

Having security problems or being unusable is.  I'd throw in things
like serious file collisions and similar serious quality problems as
well.

I'd even throw in a missing distfile, but only if no user of the
software is willing to proxy maintain that aspect of the package.

The one thing missing from this discussion is that users CAN
proxy-maintain things.  Oh, and ANYBODY can run an overlay (it just
won't necessarily be listed in layman - but that is how every distro
does it).

I do think it is a loss for Gentoo if we start removing packages
simply because they don't change (which is all a dead upstream means -
it isn't always a bad thing).

Rich

[gentoo-dev] Re: Please stop useless removals

[Diego Elio Pettenò]

On 01/02/2013 12:11, Rich Freeman wrote:
> I do think it is a loss for Gentoo if we start removing packages
> simply because they don't change (which is all a dead upstream means -
> it isn't always a bad thing).

The problem is that a package that doesn't change _will_ bitrot. Full stop.

Trying to pretend that the problem does not exist, that an unmaintained
package is just as fine as a maintained one is stupid and shortsighted,
and explains why I have 1600 bugs open...

-- 
Diego Elio Pettenò — Flameeyes
flameeyes@flameeyes.eu — http://blog.flameeyes.eu/

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 6:20 AM, Diego Elio Pettenò
<flameeyes@flameeyes.eu> wrote:
> On 01/02/2013 12:11, Rich Freeman wrote:
>> I do think it is a loss for Gentoo if we start removing packages
>> simply because they don't change (which is all a dead upstream means -
>> it isn't always a bad thing).
>
> The problem is that a package that doesn't change _will_ bitrot. Full stop.
>

Then remove it when it does.  Full stop.

Rich

Re: [gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 857 bytes --]

On Fri, 01 Feb 2013 13:30:04 +0400
Sergey Popov <pinkbyte@gentoo.org> wrote:

> 01.02.2013 12:53, Michael Weber wrote:
> > BENEFIT
> > 
> >    User can choose whether or not layman -a frozen.
> > 
> >    Non-trivial ebuilds are preserved.
> > 
> >    Tarballs are preserved.
> > 
> >    Nobody gets hurt.
> 
> Well, we can move such software to sunrise, can't we? But proposition of
> splitted mirrors makes sense, cause quite often dead upstream means dead
> links to original tarballs too.

No, Sunrise project has rather specific goals [1] and is certainly
not supposed to be a junkyard for packages removed from Gentoo. I'd
even say that packages are put in Sunrise with some hope that they will
be moved to gx86 at some point, not the other way.

[1]:http://www.gentoo.org/proj/en/sunrise/

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-dev] Re: Please stop useless removals

[Michael Weber]

On 02/01/2013 12:20 PM, Diego Elio Pettenò wrote:
> On 01/02/2013 12:11, Rich Freeman wrote:
>> I do think it is a loss for Gentoo if we start removing packages
>> simply because they don't change (which is all a dead upstream means -
>> it isn't always a bad thing).
> 
> The problem is that a package that doesn't change _will_ bitrot. Full stop.
> 
> Trying to pretend that the problem does not exist, that an unmaintained
> package is just as fine as a maintained one is stupid and shortsighted,
> and explains why I have 1600 bugs open...

Making up new situations up like cross-dev, Gentoo/Prefix, or jet
another cluttered C compiler should not doom working software.

I agree on your testing effort and practice, but compliance with the
weirdest of all setups shouldn't be ultimate reason.

-- 
Michael Weber
Gentoo Developer
web: https://xmw.de/
mailto: Michael Weber <xmw@gentoo.org>

Re: [gentoo-dev] Re: Please stop useless removals

[Diego Elio Pettenò]

On 01/02/2013 13:07, Michael Weber wrote:
> Making up new situations up like cross-dev, Gentoo/Prefix, or jet
> another cluttered C compiler should not doom working software.

Which would be all fine and dandy ....

> I agree on your testing effort and practice, but compliance with the
> weirdest of all setups shouldn't be ultimate reason.

... if you had a clue on what you were saying.

The tinderbox _by design_ is not testing "weirdest of all setups", it's
testing baseline. And if nobody's interested in getting (example)
media-video/w3cam working (#247917 — last activity on the bug by me on
2010; last activity by someone else in 2008!), I don't see why it should
be kept in tree.

Bloody hell, I wonder how many people complaining about removing
packages are actually using said packages, rather that complaining on
principles!

-- 
Diego Elio Pettenò — Flameeyes
flameeyes@flameeyes.eu — http://blog.flameeyes.eu/

Re: [gentoo-dev] Re: Please stop useless removals

[Michael Weber]

On 02/01/2013 01:22 PM, Diego Elio Pettenò wrote:
> On 01/02/2013 13:07, Michael Weber wrote:
>> Making up new situations up like cross-dev, Gentoo/Prefix, or jet
>> another cluttered C compiler should not doom working software.
> 
> Which would be all fine and dandy ....
> 
>> I agree on your testing effort and practice, but compliance with the
>> weirdest of all setups shouldn't be ultimate reason.
> 
> ... if you had a clue on what you were saying.
> 
> The tinderbox _by design_ is not testing "weirdest of all setups", it's
> testing baseline. 
Yeah, but test for /usr/share/doc/${PF} (random to irrelevant),
$CFLAGS/$LDFLAGS/$AR (enable these miraculous setup), automake-1.12 (at
what point in future do you see that as oldest in-tree) last are no
statement regarding a packages functionality on a plain system.


>And if nobody's interested in getting (example)
> media-video/w3cam working (#247917 — last activity on the bug by me on
> 2010; last activity by someone else in 2008!), I don't see why it should
> be kept in tree.
*insert random example here*
I did not argue to keep these in tree, or to label them a+++.
Martin and I did not argue that there are no circumstances an software
should be left alone. We both said, that not working with qt3/... may be
a strong argument.

> Bloody hell, I wonder how many people complaining about removing
> packages are actually using said packages, rather that complaining on
> principles!
Keep on the ground.
I rather prefer a combined discussion on "principles" or workflow, than
bringing up this discussion for every single package.
This is a general Gentoo list, so the mails might get some kind of
"general".

-- 
Michael Weber
Gentoo Developer
web: https://xmw.de/
mailto: Michael Weber <xmw@gentoo.org>

Re: [gentoo-dev] Re: Please stop useless removals

[Diego Elio Pettenò]

On 01/02/2013 13:36, Michael Weber wrote:
> Yeah, but test for /usr/share/doc/${PF} (random to irrelevant),

Which I don't open bugs about any longer.

> $CFLAGS/$LDFLAGS/$AR (enable these miraculous setup),

WTF does "enable these miraculous setup" mean? Seriously.

Also, no I don't test or bother opening bugs for either $AR or $CC. I do
test for and open bugs for $CFLAGS/$LDFLAGS handling because _that is
what Gentoo is about_ and among other things they work as a good sanity
check.

> automake-1.12 (at
> what point in future do you see that as oldest in-tree)

Are you dense? If automake-1.12 is installed, the majority of the tree
_will_ use it. The fact that I test for it is to avoid you getting the
bugs from users who really want to use your package.

> last are no
> statement regarding a packages functionality on a plain system.

If the package is TFU, and nobody cares enough to fix it, the
functionality "on a plain system" is screwed up anyway.

If you can't be bothered to make your package comply with at least the
minimum style of the rest of the tree, I'd honestly prefer you gave up
tree access.

> Keep on the ground.
> I rather prefer a combined discussion on "principles" or workflow, than
> bringing up this discussion for every single package.
> This is a general Gentoo list, so the mails might get some kind of
> "general".

The problem here is that it's not general. It's fantasy.

I'm not saying that we should remove a package because it has one
trivial bug not fixed in three months. But when upstream is dead, and
nobody in Gentoo is caring for it, has half a dozen open bug (trivial or
not), unsolved or unsolvable for over an year... punt the crap from the
tree and reduce the overload.

Also, since you are a dev, instead of complaining at how team $x removes
their packages, you can step in and save the package. As Alec said
"Gentoo is not a software archival service." so arguing on the principle
that we should never delete any package from our tree is simply
preposterous.

-- 
Diego Elio Pettenò — Flameeyes
flameeyes@flameeyes.eu — http://blog.flameeyes.eu/

Re: [gentoo-dev] "frozen" overlay Re: Please stop useless removals

[Ian Stakenvicius]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 01/02/13 05:53 AM, Michael Weber wrote:
> Sunrise commit access is limited to sunrise devs. And I see the
> _rise_ in context of software and devs. I don't say sundown,

..there once was a "sunset" overlay, wasn't there?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iF4EAREIAAYFAlELwb4ACgkQ2ugaI38ACPC9xgD+NYja02p1q1tQTrTkjBBpyoop
xLFmGsGcw6sUT6e4bY4A/RiMbkAM0b6nmYkhA/zfJLFqMUudTAWd8VaLB7aD9nBe
=WePv
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Ian Stakenvicius]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 01/02/13 06:20 AM, Diego Elio Pettenò wrote:
> On 01/02/2013 12:11, Rich Freeman wrote:
>> I do think it is a loss for Gentoo if we start removing packages 
>> simply because they don't change (which is all a dead upstream
>> means - it isn't always a bad thing).
> 
> The problem is that a package that doesn't change _will_ bitrot.
> Full stop.
> 
> Trying to pretend that the problem does not exist, that an
> unmaintained package is just as fine as a maintained one is stupid
> and shortsighted, and explains why I have 1600 bugs open...
> 

True -- but then, the reason for that package's removal is one or many
of those bugs, not because upstream is dead and the package is old and
might at some point in the future have bugs due to bitrot.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iF4EAREIAAYFAlELwj4ACgkQ2ugaI38ACPCa1QEAggm0vXETySkPrLJD3Lquvc4Q
Kkt7ft0dBamMGH86bE4BAL1S1X7T9dZZS88on2GhAZKy81iY8G8VWch8GUXw3Q5k
=6TbE
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 7:53 AM, Diego Elio Pettenò
<flameeyes@flameeyes.eu> wrote:
> I'm not saying that we should remove a package because it has one
> trivial bug not fixed in three months. But when upstream is dead, and
> nobody in Gentoo is caring for it, has half a dozen open bug (trivial or
> not), unsolved or unsolvable for over an year... punt the crap from the
> tree and reduce the overload.

Open trivial bugs don't create any overload, except for those who go
looking at them and worrying about them at night.

As long as it builds on 80%+ of systems and has no serious issues
(security in particular) there is no reason to remove a package.  Yes,
quality issues might cause it to have issues on 80% of systems in the
future, and when that happens prune it.

I have no idea how many open bugs Gentoo has.  The reason for this is
that I search for bugs that I care about, and the only thing that has
to worry about the rest is the database server.  If we had a trillion
open bugs I'd start worrying about that more, though simply closing
them wouldn't help in that case.

Remove things when they cause problems, not before.

Rich

Re: [gentoo-dev] Re: Please stop useless removals

[Wulf C. Krueger]

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

On 01.02.2013 14:26, Rich Freeman wrote:
> As long as it builds on 80%+ of systems and has no serious issues 
> (security in particular) there is no reason to remove a package.

And how will you get to know about current or future security issues if
nobody (in Gentoo) cares about the package?

> Remove things when they cause problems, not before.

You mean, not before your users' systems have been compromised and they
complain loudly about it?

Best regards, Wulf
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlELxNgACgkQnuVXRcSi+5qP8wCghvWTuQvcFfJojX9HS8Jln6O/
144AnipUMY1NU8DbrtzesEbvpSHeYkPt
=awFq
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 8:36 AM, Wulf C. Krueger <wk@mailstation.de> wrote:
>
> And how will you get to know about current or future security issues if
> nobody (in Gentoo) cares about the package?

The same way that you know about security issues in Firefox or
Chromium - somebody reports them.  Security bugs still go to the
security team, and they're welcome to treeclean with a vengence.

I guarantee that you have unreported security bugs in whatever browser
and email client you're using right now.  Until somebody tells
upstream about them you're going to be vulnerable.

That said, I'm fine with having some kind of overlay for stuff like
this (we need to reduce the stigma on overlays), and I think that
having some kind of quality tagging system also makes sense for
communicating just how clean packages are.  Give the users a choice.
Overlays seem to be largely used to do just this - the overlay itself
has some connotation of level-of-quality.

Rich

[gentoo-dev] Re: Please stop useless removals

[Michael Palimaka]

On 2/02/2013 00:36, Wulf C. Krueger wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 01.02.2013 14:26, Rich Freeman wrote:
>> As long as it builds on 80%+ of systems and has no serious issues
>> (security in particular) there is no reason to remove a package.
>
> And how will you get to know about current or future security issues if
> nobody (in Gentoo) cares about the package?
The security team routinely monitors various information sources to 
ensure that issues are tracked regardless of maintainer.

>> Remove things when they cause problems, not before.
>
> You mean, not before your users' systems have been compromised and they
> complain loudly about it?
>
> Best regards, Wulf
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.19 (GNU/Linux)
> Comment: Using GnuPG with undefined - http://www.enigmail.net/
>
> iEYEARECAAYFAlELxNgACgkQnuVXRcSi+5qP8wCghvWTuQvcFfJojX9HS8Jln6O/
> 144AnipUMY1NU8DbrtzesEbvpSHeYkPt
> =awFq
> -----END PGP SIGNATURE-----
>
>

Re: [gentoo-dev] Re: Please stop useless removals

[Wulf C. Krueger]

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

On 01.02.2013 14:47, Rich Freeman wrote:
>> And how will you get to know about current or future security 
>> issues if nobody (in Gentoo) cares about the package?
> The same way that you know about security issues in Firefox or 
> Chromium [...] Until somebody tells upstream about them you're
> going to be vulnerable.

Indeed. In contrast to many of the packages that were mentioned in this
thread, Firefox and Chromium have an active upstream, though.

What do you think will happen to projects with a dead upstream? I
think the answer is pretty simple: Nothing.

Thus, your users' systems will remain vulnerable and you won't even
know about it.

Best regards, Wulf
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlELyXkACgkQnuVXRcSi+5q6UgCfQLgmYQkShYNu2bwokxzP32Fv
FBEAoNz/qw2QRArkSUugGXgL3bII6zn9
=aboK
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Ian Stakenvicius]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 01/02/13 08:56 AM, Wulf C. Krueger wrote:
> On 01.02.2013 14:47, Rich Freeman wrote:
>>> And how will you get to know about current or future security 
>>> issues if nobody (in Gentoo) cares about the package?
>> The same way that you know about security issues in Firefox or 
>> Chromium [...] Until somebody tells upstream about them you're 
>> going to be vulnerable.
> 
> Indeed. In contrast to many of the packages that were mentioned in
> this thread, Firefox and Chromium have an active upstream, though.
> 
> What do you think will happen to projects with a dead upstream? I 
> think the answer is pretty simple: Nothing.

Not really, no.  A dead upstream means that there isn't an upstream to
push a fix or release a new version.  That's all.

If security bugs occur then there's two options -- fix, or remove.  So
if the gentoo dev in question doesn't have time/ability/desire to fix,
they or security remove it at that point.

This isn't "nothing" to me; I must be missing something from your
response?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iF4EAREIAAYFAlELyo8ACgkQ2ugaI38ACPC1FAD/fxM93LFEKtl8t87qc6QSIkTL
HkQtk2t4xFQxoBAZNIUBALrMJxstxw4pBwOytiQfJq9CLxf3dOnUIQCdRDwIxA6Y
=j28W
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Wulf C. Krueger]

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

Sorry for quoting a lot this time but it's important for understanding
the issue.

On 01.02.2013 15:00, Ian Stakenvicius wrote:
> On 01/02/13 08:56 AM, Wulf C. Krueger wrote:
>> On 01.02.2013 14:47, Rich Freeman wrote:
>>>> And how will you get to know about current or future
>>>> security issues if nobody (in Gentoo) cares about the
>>>> package?
>>> The same way that you know about security issues in Firefox or 
>>> Chromium [...] Until somebody tells upstream about them you're 
>>> going to be vulnerable.
>> Indeed. In contrast to many of the packages that were mentioned
>> in this thread, Firefox and Chromium have an active upstream,
>> though. What do you think will happen to projects with a dead
>> upstream? I think the answer is pretty simple: Nothing.
> Not really, no.  A dead upstream means that there isn't an upstream
> to push a fix or release a new version.  That's all. If security
> bugs occur then there's two options -- fix, or remove.  So if the
> gentoo dev in question doesn't have time/ability/desire to fix, 
> they or security remove it at that point. This isn't "nothing" to
> me; I must be missing something from your response?

Yes, the topmost two lines in my quote:

>>>> And how will you get to know about current or future
>>>> security issues if nobody (in Gentoo) cares about the
>>>> package?

In the "dead upstream" case it's unlikely anyone is checking the
package for security issues in the first place. So neither the Gentoo
security people will get notice via the usual sources nor will any
upstream be informed.

If there's a *known* bug, you're right. Case closed.

If the package in question is just bit-rotting and nobody cares, you
most likely won't ever know about any security issues, though - until
something nasty happens. This is one of the problems with "dead
upstream" packages.

Best regards, Wulf
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlELzGEACgkQnuVXRcSi+5rJAwCfYGcHAJzmxwD+2L0WZlajnfP4
TzsAn1NN88QQDG3Q9br73nM1KcFT9rDW
=5aeo
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 9:08 AM, Wulf C. Krueger <wk@mailstation.de> wrote:
>
> In the "dead upstream" case it's unlikely anyone is checking the
> package for security issues in the first place. So neither the Gentoo
> security people will get notice via the usual sources nor will any
> upstream be informed.

That seems rather speculative.  I'm sure that people look for
vulnerabilities in unmaintained software - if they didn't then nobody
would be able to exploit them in the first place (you have to find a
vulnerability to exploit it).  I imagine most vulnerabilities are
found by people outside of projects in the first place.

We don't know how many vulnerabilities there are in maintained
packages, let alone unmaintained ones, so a comparison is a bit
difficult.

Popularity is probably a better indicator of whether something will
have vulnerabilities reported than whether it has an upstream.  The
two are of course loosely connected.

Rich

Re: [gentoo-dev] Re: Please stop useless removals

[Richard Yao]

[-- Attachment #1: Type: text/plain, Size: 1106 bytes --]

On 02/01/2013 07:07 AM, Michael Weber wrote:
> On 02/01/2013 12:20 PM, Diego Elio Pettenò wrote:
>> On 01/02/2013 12:11, Rich Freeman wrote:
>>> I do think it is a loss for Gentoo if we start removing packages
>>> simply because they don't change (which is all a dead upstream means -
>>> it isn't always a bad thing).
>>
>> The problem is that a package that doesn't change _will_ bitrot. Full stop.
>>
>> Trying to pretend that the problem does not exist, that an unmaintained
>> package is just as fine as a maintained one is stupid and shortsighted,
>> and explains why I have 1600 bugs open...
> 
> Making up new situations up like cross-dev, Gentoo/Prefix, or jet
> another cluttered C compiler should not doom working software.
> 
> I agree on your testing effort and practice, but compliance with the
> weirdest of all setups shouldn't be ultimate reason.
> 

Being broken on one architecture should not prevent a package from being
available to others where it works. You just do not keyword things on
architectures where they are broken. This is why we have keywording.


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 900 bytes --]

[gentoo-dev] Re: Please stop useless removals

[Richard Yao]

[-- Attachment #1: Type: text/plain, Size: 2880 bytes --]

On 02/01/2013 02:36 AM, Vaeth wrote:
> 
>>>    # Upstream is dead and gone.
>>>    # Masked for removal on 20130302
>>
>> Erm, so this is the _only_ reason - dead upstream?
> 
> ++
> 
> Please, please, stop removing packages for no reason!
> This happens now way too often:
> 
> app-dicts/ispell*
> app-portage/epm
> app-text/ispell
> games-arcade/bitefusion
> games-arcade/xboing
> games-action/trackballs
> games-emulation/xmame
> ...
> 
> These are just some of the previous examples which I remember
> because I had to put them in my local overlay.
> 
> None of these removals alone was so valuable to me that I saw
> a reason to step up, but the removals for no reasons accumulate
> previously so much that I see the need to say something:
> 
> You are destroying the charme of gentoo by systematically
> removing all these little tools and toys.  The availability
> of a lot of software was once a strength of gentoo, so removing
> these things is really bad, especially if it happens for no
> real reason.
> 
> I was understanding if e.g. someting was removed which needs
> the <gtk-2 or <qt-4 framework or something similar and had
> a dead upstream. But just needing a small tool like imake (xboing)
> or having open feature requestes (epm) or even nothing and
> just dead upstream is IMHO really not a reason.
> 
> If something really does not compile anymore and nobody cares,
> then remove keywords (or, for god's sake, mask it);
> if something might theoretically become a security issue (xpdf)
> then it should be masked.
> 
> But please do not throw things out of the tree unless
> really necessary:
> 
> It does not hurt anybody to have such package in the tree,
> but removing it - especially if upstream is dead - means
> that the tarbalös will be removed from the mirrors and thus
> nobody is able anymore to install it (even if he would care and
> fix some minor issues) unless he had kept a copy on
> his local machine (which will mean in the future that he can only
> do it if he had used gentoo already many years ago and cared
> during the time of the removal).
> 
> (If the resources are an argument: I am not speaking about monster
> packages taking gigabytes of data - these might need to be
> discussed separately - but mainly about reasonably sized packages
> which even if summed up do not take much data).
> 
> Regards
> Martin

I suspect that the removal message is inaccurate. The actual reason for
removal is the following:

https://bugs.gentoo.org/show_bug.cgi?id=425298

If you were to make a webpage for it and host the tarball for people, it
should be possible to resolve that bug. That should be sufficient to
have the removal mask removed. I suspect that the Anapnea network will
be more than happy to provide you with hosting for this:

http://www.anapnea.net/


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 900 bytes --]

Re: [gentoo-dev] Re: Please stop useless removals

[Christopher Head]

On Fri, 1 Feb 2013 09:45:07 -0500
Rich Freeman <rich0@gentoo.org> wrote:


> That seems rather speculative.  I'm sure that people look for
> vulnerabilities in unmaintained software - if they didn't then nobody
> would be able to exploit them in the first place (you have to find a
> vulnerability to exploit it).  I imagine most vulnerabilities are
> found by people outside of projects in the first place.
> 
> We don't know how many vulnerabilities there are in maintained
> packages, let alone unmaintained ones, so a comparison is a bit
> difficult.

Also, there are plenty of packages that can't really *have* interesting
security vulnerabilities in the first place. I don't know the specifics
of the games that were removed, but games in general, if they are
purely single-player and only ever read and write files in the player's
home directory, don't really have an attack surface to start with. You
can't remotely exploit a program that never creates a socket, and you
can't locally exploit a program that never tries to access files other
than those in its invoker's home directory and root-writable
directories like /usr/share, and does so with the invoker's usual
privileges. Do you treeclean those because "they might have security
holes"?

Chris

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 10:51 AM, Richard Yao <ryao@gentoo.org> wrote:
> I suspect that the removal message is inaccurate. The actual reason for
> removal is the following:
>
> https://bugs.gentoo.org/show_bug.cgi?id=425298
>
> If you were to make a webpage for it and host the tarball for people, it
> should be possible to resolve that bug. That should be sufficient to
> have the removal mask removed.

Yes, after sending out my email I took a closer look and came to the
same conclusion.

I'm perfectly fine with masking/removing packages that do not have
valid SRC_URIs, and if somebody wants to host the tarball somewhere
and submit a patch to fix it we shouldn't have a problem with a dev
committing that patch and prolonging the package a bit longer (though
ideally a proxy maintainer would be helpful).

Bottom line is that we shouldn't drop packages simply because they're
unmaintained or lack an upstream.  Missing SRC_URIs on unmaintained
packages are fair game, however, as are other serious issues.  I have
no desire to make the mirror maintainers sort through log noise on
something like this.

For those who are doing the treecleaning, please do yourself a favor
and point out the actual show-stoppers so that you don't have a war on
your hand every time you mask something.  :)

Rich

Re: [gentoo-dev] Re: Please stop useless removals

[Diego Elio Pettenò]

On 01/02/2013 23:52, Rich Freeman wrote:
> 
> For those who are doing the treecleaning, please do yourself a favor
> and point out the actual show-stoppers so that you don't have a war on
> your hand every time you mask something.  :)

Or maybe, you know, stop starting idiotic flamewars on principles
assuming that all of QA is out to ruin your life, which seems to happen
pretty often to you.

-- 
Diego Elio Pettenò — Flameeyes
flameeyes@flameeyes.eu — http://blog.flameeyes.eu/

Re: [gentoo-dev] Re: Please stop useless removals

[Rich Freeman]

On Fri, Feb 1, 2013 at 5:54 PM, Diego Elio Pettenò
<flameeyes@flameeyes.eu> wrote:
> On 01/02/2013 23:52, Rich Freeman wrote:
>>
>> For those who are doing the treecleaning, please do yourself a favor
>> and point out the actual show-stoppers so that you don't have a war on
>> your hand every time you mask something.  :)
>
> Or maybe, you know, stop starting idiotic flamewars on principles
> assuming that all of QA is out to ruin your life, which seems to happen
> pretty often to you.

The argument was made that unmaintained packages that have dead
upstreams should be removed.  I explained why this was bad policy.
This is not a flamewar.

It turns out that this wasn't actually why these packages were
removed, but it doesn't really change the validity of anything I said.
 In the end the error wasn't in the removal of the packages, but in
the justification for doing so.

It really isn't meant personally, and I certainly don't take it as such.

Rich

Re: [gentoo-dev] Re: Please stop useless removals

[hasufell]

On 02/02/2013 12:17 AM, Rich Freeman wrote:
> On Fri, Feb 1, 2013 at 5:54 PM, Diego Elio Pettenò
> <flameeyes@flameeyes.eu> wrote:
>> On 01/02/2013 23:52, Rich Freeman wrote:
>>>
>>> For those who are doing the treecleaning, please do yourself a favor
>>> and point out the actual show-stoppers so that you don't have a war on
>>> your hand every time you mask something.  :)
>>
>> Or maybe, you know, stop starting idiotic flamewars on principles
>> assuming that all of QA is out to ruin your life, which seems to happen
>> pretty often to you.
> 
> The argument was made that unmaintained packages that have dead
> upstreams should be removed.  I explained why this was bad policy.
> This is not a flamewar.
> 

+1

Dead upstream is no reason alone to treeclean any package. A reason
would be a severe runtime or buildtime bug, that needs a non-trivial
fix, but no upstream to take care of that.

Re: [gentoo-dev] Re: Please stop useless removals

[Philip Webb]

130201 Rich Freeman wrote:
> On Fri, Feb 1, 2013 at 10:51 AM, Richard Yao <ryao@gentoo.org> wrote:
>> The actual reason for removal is the following:
>>   https://bugs.gentoo.org/show_bug.cgi?id=425298
> I'm perfectly fine with masking/removing packages
> that do not have valid SRC_URIs
> and if somebody wants to host the tarball somewhere
> and submit a patch to fix it we shouldn't have a problem
> with a dev committing that patch and prolonging the package a bit longer.
> Bottom line is that we shouldn't drop packages
> simply because they're unmaintained or lack an upstream.

+1

> Missing SRC_URIs on unmaintained packages are fair game, however,
> as are other serious issues.  I have no desire
> to make the mirror maintainers sort thro log noise on something like this.

If a mere user may comment (smile),
I use  >= 1  pkg which hasn't been updated for a long time, Apwal,
but is in fact an excellent little app which deserves wider knowledge.
It's one of those apps which needs no further development.

There are also pkgs like Nethack, which is hard-masked
because there's a serious security bug on multi-user systems,
but which offers no problems on a single-user desktop.

-- 
========================,,============================================
SUPPORT     ___________//___,   Philip Webb
ELECTRIC   /] [] [] [] [] []|   Cities Centre, University of Toronto
TRANSIT    `-O----------O---'   purslowatchassdotutorontodotca

[gentoo-dev] Re: Please stop useless removals

[James Cloos]

>>>>> "AW" == Alec Warner <antarus@gentoo.org> writes:

AW> If folks do not want to maintain it anymore, then it will be removed.

That is about as harmful an attitude as possible.

If you don't personally care about a package just leave it alone!

And if you want more maintainers, then drop the schoolkid nonsense to
join the club.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: [gentoo-dev] Re: Please stop useless removals

[Alec Warner]

On Sat, Feb 2, 2013 at 10:14 AM, James Cloos <cloos@jhcloos.com> wrote:
>>>>>> "AW" == Alec Warner <antarus@gentoo.org> writes:
>
> AW> If folks do not want to maintain it anymore, then it will be removed.
>
> That is about as harmful an attitude as possible.
>
> If you don't personally care about a package just leave it alone!

The point of treecleaners is to clean stuff that is broken. They
shouldn't be removing packages that do not work. That being said, if
it doesn't work and no one is willing to maintain it, it doesn't
belong in the tree.

>
> And if you want more maintainers, then drop the schoolkid nonsense to
> join the club.

If this is some opaque comment about how it should easier to
contribute to Gentoo (including being a proxy maintainer) then I
couldn't agree more. That being said after looking at what it takes to
become a debian developer or ubuntu developer...I don't think our
process is any more onerous than those.

>
> -JimC
> --
> James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6
>