[gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Patrice Clement]

Hi dlan

This commit is breaking the tree:
https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec

I did try to work around the issue you've introduced when but the enlightenment
eclasses are a bit of mystery to me and I eventually gave up.

Could you revert this commit and fix this issue?

Thank you!

----- Forwarded message from mgorny@bonedaddy.net -----

> From: mgorny@bonedaddy.net
> Subject: [gentoo-automated-testing] BROKEN: repository became broken!
> To: gentoo-automated-testing@lists.gentoo.org
> CC: dlan@gentoo.org
> Date: Mon, 01 Feb 2016 03:31:56 +0000
> 
> Looks like someone just broke Gentoo!
> 
> New issues:
> https://qa-reports.gentoo.org/output/gentoo-ci/780f65b/output.html#dev-libs/efl
> 
> 
> Introduced by commits:
> https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec
> 
> 
> Changes since last check:
> https://gitweb.gentoo.org/repo/gentoo.git/log/?qt=range&q=2496d1d..37c79f8
> 
> --
> Gentoo repository CI
> 

----- End forwarded message -----

-- 
Patrice Clement
Gentoo Linux developer
http://www.gentoo.org

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 937 bytes --]

On 01 Feb 2016 19:55, Patrice Clement wrote:
> > New issues:
> > https://qa-reports.gentoo.org/output/gentoo-ci/780f65b/output.html#dev-libs/efl
>
> This commit is breaking the tree:
> https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec
> 
> I did try to work around the issue you've introduced when but the enlightenment
> eclasses are a bit of mystery to me and I eventually gave up.
> 
> Could you revert this commit and fix this issue?

the issue is that efl-1.15.2 is marked stable for alpha/ia64/sparc, and
it depends on app-i18n/ibus, but commit 97a6aec deleted the only ibus
ebuild that was marked stable for those arches.

it can be fixed in a few ways (i'm listing in order of preference):
(1) mark a newer ibus stable
(2) revert that commit to re-add the old stable ebuilds
(3) add USE=ibus to package.use.stable.mask for these arches
(4) degrade all packages for these arches to unstable
-mike

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Andreas K. Huettel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Am Montag, 1. Februar 2016, 21:30:41 schrieb Mike Frysinger:
> On 01 Feb 2016 19:55, Patrice Clement wrote:
> > > New issues:
> > > https://qa-reports.gentoo.org/output/gentoo-ci/780f65b/output.html#dev-> > > libs/efl
> > 
> > This commit is breaking the tree:
> > https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec
> > 
> > I did try to work around the issue you've introduced when but the
> > enlightenment eclasses are a bit of mystery to me and I eventually gave
> > up.
> > 
> > Could you revert this commit and fix this issue?
> 
> the issue is that efl-1.15.2 is marked stable for alpha/ia64/sparc, and
> it depends on app-i18n/ibus, but commit 97a6aec deleted the only ibus
> ebuild that was marked stable for those arches.
> 
> it can be fixed in a few ways (i'm listing in order of preference):
> (1) mark a newer ibus stable
> (2) revert that commit to re-add the old stable ebuilds
> (3) add USE=ibus to package.use.stable.mask for these arches
> (4) degrade all packages for these arches to unstable
> -mike

I took the liberty of doing (2) and reverted the commit. Not sure why this 
needs so much discussion; after all a broken tree is always suboptimal.

Now there's lots of time to come up with a better solution.


- -- 

Andreas K. Huettel
Gentoo Linux developer 
dilfridge@gentoo.org
http://www.akhuettel.de/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.1
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=S77D
-----END PGP SIGNATURE-----

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1669 bytes --]

On 01 Feb 2016 22:16, Andreas K. Huettel wrote:
> Am Montag, 1. Februar 2016, 21:30:41 schrieb Mike Frysinger:
> > On 01 Feb 2016 19:55, Patrice Clement wrote:
> > > > New issues:
> > > > https://qa-reports.gentoo.org/output/gentoo-ci/780f65b/output.html#dev-libs/efl
> > > 
> > > This commit is breaking the tree:
> > > https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec
> > > 
> > > I did try to work around the issue you've introduced when but the
> > > enlightenment eclasses are a bit of mystery to me and I eventually gave
> > > up.
> > > 
> > > Could you revert this commit and fix this issue?
> > 
> > the issue is that efl-1.15.2 is marked stable for alpha/ia64/sparc, and
> > it depends on app-i18n/ibus, but commit 97a6aec deleted the only ibus
> > ebuild that was marked stable for those arches.
> > 
> > it can be fixed in a few ways (i'm listing in order of preference):
> > (1) mark a newer ibus stable
> > (2) revert that commit to re-add the old stable ebuilds
> > (3) add USE=ibus to package.use.stable.mask for these arches
> > (4) degrade all packages for these arches to unstable
> 
> I took the liberty of doing (2) and reverted the commit. Not sure why this 
> needs so much discussion; after all a broken tree is always suboptimal.

unless things are on fire (which i don't think this was), i don't
generally clamor for 0-day fixes.  if we can find a better fix in
a day or so, then i'm happy for that.  i dislike repos with history
that is just a constant stream of land, revert, land, revert, land.

not that i'm saying your revert was wrong ... just airing my
general personal preferences.
-mike

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Yixun Lan]

[-- Attachment #1: Type: text/plain, Size: 1533 bytes --]

On 22:16 Mon 01 Feb     , Andreas K. Huettel wrote:
> Am Montag, 1. Februar 2016, 21:30:41 schrieb Mike Frysinger:
> > On 01 Feb 2016 19:55, Patrice Clement wrote:
> > > > New issues:
> > > > https://qa-reports.gentoo.org/output/gentoo-ci/780f65b/output.html#dev-> > > libs/efl
> > > 
> > > This commit is breaking the tree:
> > > https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=97a6aec
> > > 
> > > I did try to work around the issue you've introduced when but the
> > > enlightenment eclasses are a bit of mystery to me and I eventually gave
> > > up.
> > > 
> > > Could you revert this commit and fix this issue?
> > 
> > the issue is that efl-1.15.2 is marked stable for alpha/ia64/sparc, and
> > it depends on app-i18n/ibus, but commit 97a6aec deleted the only ibus
> > ebuild that was marked stable for those arches.
> > 
> > it can be fixed in a few ways (i'm listing in order of preference):
> > (1) mark a newer ibus stable
> > (2) revert that commit to re-add the old stable ebuilds
> > (3) add USE=ibus to package.use.stable.mask for these arches
> > (4) degrade all packages for these arches to unstable
> > -mike
> 
> I took the liberty of doing (2) and reverted the commit. Not sure why this 
> needs so much discussion; after all a broken tree is always suboptimal.
> 
> Now there's lots of time to come up with a better solution.
> 
thanks for doing this

yes, 2) is the easiest and quickest way to fix.. 


-- 
Yixun Lan (dlan)
Gentoo Linux Developer
GPG Key ID AABEFD55

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 951 bytes --]

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Andreas K. Huettel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Am Dienstag, 2. Februar 2016, 02:33:30 schrieb Mike Frysinger:
>
> > I took the liberty of doing (2) and reverted the commit. Not sure why
> > this needs so much discussion; after all a broken tree is always
> > suboptimal.
> 
> unless things are on fire (which i don't think this was), i don't
> generally clamor for 0-day fixes.  if we can find a better fix in
> a day or so, then i'm happy for that.  i dislike repos with history
> that is just a constant stream of land, revert, land, revert, land.
> 
> not that i'm saying your revert was wrong ... just airing my
> general personal preferences.

You're right of course... but there's one thing we have to keep in mind.

We're not running a project were releases are made from the vcs. The vcs *is* 
the release... and whatever is out there gets pushed to users.

This is why my personal preference is more to revert if I'm not sure that the 
fix will happen soon.

- -- 

Andreas K. Huettel
Gentoo Linux developer 
dilfridge@gentoo.org
http://www.akhuettel.de/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.1
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=O//d
-----END PGP SIGNATURE-----

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1216 bytes --]

On 03 Feb 2016 22:35, Andreas K. Huettel wrote:
> Am Dienstag, 2. Februar 2016, 02:33:30 schrieb Mike Frysinger:
> > > I took the liberty of doing (2) and reverted the commit. Not sure why
> > > this needs so much discussion; after all a broken tree is always
> > > suboptimal.
> > 
> > unless things are on fire (which i don't think this was), i don't
> > generally clamor for 0-day fixes.  if we can find a better fix in
> > a day or so, then i'm happy for that.  i dislike repos with history
> > that is just a constant stream of land, revert, land, revert, land.
> > 
> > not that i'm saying your revert was wrong ... just airing my
> > general personal preferences.
> 
> You're right of course... but there's one thing we have to keep in mind.
> 
> We're not running a project were releases are made from the vcs. The vcs *is* 
> the release... and whatever is out there gets pushed to users.
> 
> This is why my personal preference is more to revert if I'm not sure that the 
> fix will happen soon.

which is why you weigh the impact on users.  how many people are actually
affected and for how long ?  in this case, fairly sure no actual user saw
the failure on their system.
-mike

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [gentoo-dev] [Fwd: [gentoo-automated-testing] BROKEN: repository became broken!]

[Raymond Jennings]

[-- Attachment #1: Type: text/plain, Size: 1492 bytes --]

I think best of all would be the good discipline not to break the tree in
the first place.

Is this something that Repoman could have caught?  If no, should it in the
future?

On Tue, Feb 9, 2016 at 11:30 AM, Mike Frysinger <vapier@gentoo.org> wrote:

> On 03 Feb 2016 22:35, Andreas K. Huettel wrote:
> > Am Dienstag, 2. Februar 2016, 02:33:30 schrieb Mike Frysinger:
> > > > I took the liberty of doing (2) and reverted the commit. Not sure why
> > > > this needs so much discussion; after all a broken tree is always
> > > > suboptimal.
> > >
> > > unless things are on fire (which i don't think this was), i don't
> > > generally clamor for 0-day fixes.  if we can find a better fix in
> > > a day or so, then i'm happy for that.  i dislike repos with history
> > > that is just a constant stream of land, revert, land, revert, land.
> > >
> > > not that i'm saying your revert was wrong ... just airing my
> > > general personal preferences.
> >
> > You're right of course... but there's one thing we have to keep in mind.
> >
> > We're not running a project were releases are made from the vcs. The vcs
> *is*
> > the release... and whatever is out there gets pushed to users.
> >
> > This is why my personal preference is more to revert if I'm not sure
> that the
> > fix will happen soon.
>
> which is why you weigh the impact on users.  how many people are actually
> affected and for how long ?  in this case, fairly sure no actual user saw
> the failure on their system.
> -mike
>

[-- Attachment #2: Type: text/html, Size: 2096 bytes --]