Bug#835658: RFS: backbone/1.3.3+ds-1

[Jonas Smedegaard]

[dropping most cc'ed people]

Quoting Julien Puydt (2016-08-28 21:05:33)
> On 28/08/2016 19:51, Jonas Smedegaard wrote:
>> It seems you prepared a new package release with radical structural 
>> changes, and requested sponsorship for uploading it.  First I hear of 
>> all that is today, after the fact.  Did I miss some prior 
>> communication somehow?

> () 3 may 2014, Julian Taylor files #746787 asking for a new upstream 
> release (eh, that was for ipython and I'm aiming for jupyter, which is 
> ex-ipython!) ;
> () 1 september 2015, Pirate Praveen files #797698, asking for a new 
> version ;
> () 5 october 2015, David Prévot notes that #797698 affects owncloud ;
> () 16 november 2015, Thomas Goirand comments #797698 that he also 
> needs a higher version and asked if he could do the work -- he didn't 
> get a reply (at least in the bug) ;
> () 1 june 2016, Christophe Troestler complains in #797698 that 
> owncloud in Debian is stuck to a older version (here the situation is 
> cloudy [pun intended], because the same day David Prévot removes the 
> indication that this bug affects owncloud...)

What I first heard of few days ago was _your_ involvement with Backbone 
packaging, Julien.

Seems from your summary that I didn't miss prior communication with you.


>> If not, then please elaborate a bit on the kind of collaboration you 
>> want.

> I mean work together (in fact, that's precisely what the Latin roots 
> of the word mean). And by together, I don't mean always together. Just 
> that several hands cooperate to get things done. If one hand is 
> resting, the other can do whatever is needed alone.

Thanks for clarifying.

Thanks for your offer to collaborate on maintaining Backbone packaging 
for Debian, but I choose to decline: I do not find your style of 
colaboration beneficial to the package, as it evidently involves doing 
radical structural changes to packaging without prior coordination with 
peer package maintainers, some of it arguably weakening the quality of 
the package.

I certainly believe that you did it with the best of intentions, and I 
would very much appreciate your collaboration in another style involving 
coordination _before_ action.


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: signature

Bug#835658: RFS: backbone/1.3.3+ds-1

[Julien Puydt]

Hi,

On 28/08/2016 19:51, Jonas Smedegaard wrote:

Quoting Julien Puydt (2016-08-28 16:12:33)

On 28/08/2016 15:43, Jonas Smedegaard wrote:

Quoting Julien Puydt (2016-08-28 14:35:32)

On 28/08/2016 10:00, Gianfranco Costamagna wrote:

  I am looking for a sponsor for the package "backbone"


Do you intend to help maintain the package collaboratively, or take
over maintenance?


Collaboratively.


Great!

...but perhaps we have different understandings of that term, then:

It seems you prepared a new package release with radical structural
changes, and requested sponsorship for uploading it.  First I hear of
all that is today, after the fact.  Did I miss some prior communication
somehow?

If not, then please elaborate a bit on the kind of collaboration you
want.

[holding back on answering the rest of the mail until above is more
clear]


Answer about the missed prior communication :

() 3 may 2014, Julian Taylor files #746787 asking for a new upstream 
release (eh, that was for ipython and I'm aiming for jupyter, which is 
ex-ipython!) ;
() 1 september 2015, Pirate Praveen files #797698, asking for a new 
version ;

() 5 october 2015, David Prévot notes that #797698 affects owncloud ;
() 16 november 2015, Thomas Goirand comments #797698 that he also needs 
a higher version and asked if he could do the work -- he didn't get a 
reply (at least in the bug) ;
() 1 june 2016, Christophe Troestler complains in #797698 that owncloud 
in Debian is stuck to a older version (here the situation is cloudy [pun 
intended], because the same day David Prévot removes the indication that 
this bug affects owncloud...)


Answer about the meaning of collaboration : I mean work together (in 
fact, that's precisely what the Latin roots of the word mean). And by 
together, I don't mean always together. Just that several hands 
cooperate to get things done. If one hand is resting, the other can do 
whatever is needed alone.


I hope that clears things as requested,

Snark on #debian-js

Bug#835658: RFS: backbone/1.3.3+ds-1

[Gianfranco Costamagna]

(dropping the cc list from my next email, please follow #835658 if you care 
about the thread)


>Don't take for granted that what you find obvious is equally obvious to 
>others.  Especially when raised as an explicit question: Makes you seem 
>like taking me for a fool (which I am sure was not intended).


yes, you are right, I didn't mean to take you as a fool of course :)
(glad you didn't take seriously my bad communication)


In my opinion help in maintenance means
"hey maintainer/team, is it ok to add me as uploader"?

in this case I see more as a:
"hey maintainer/team, can I update it to the latest release because foo and 
bar"?

this seems intended as a one-shot upload (of course the goal release is Stretch
so we have to see it properly migrate to testing), to fix
the reverse-dependencies.

each case is different, pkg-javascript is a big team, and probably with not so 
good
defined rules
(specially because the packaging for most of the packages is so trivial)

>>> If the former, then you should coordinate your work with current 
>>> maintainers (fine that you seek help and guidance from others too, 
>>> but don't go ahead with only input from external parties).
>> 
>> this is something that a good sponsor has to check, and the reason for 
>> me putting *you* in cc :)
>
>Could you please elaborate on what you find an obvious role for a 
>"sponsor" here.  Assume I have never¹ hear that term before.
>
>
>¹ I have, but from the context here I suspect we mean different things 
>by that term.


(still not a native speaker, so apologizes in advance and again if I made
some communication mistakes)


"sponsor" is a word that I guess comes from the RFS term, in my opinion a 
sponsor
is a person that tries to facilitate the fix for a particular bug or package, 
and
eventually uploads the package.
In this case the correct word would have probably been "mentor", because
what I checked was to be sure the Team/Maintainer/Uploaders were agreeing
on the changes.

(BTW to be fair, I also cc'd the people who requested the new version, to give
publicly a thread to say their opinion)

so, even if I used the "sponsor" word, I don't intend of course to upload 
anything,
at least until somebody with an authoritative hat asks me to do it :)

>[remaining mail skipped for now, to keep a focus]


this is nice, thanks
(I hope my good faith is not in discussion here!)

G.

Bug#835658: RFS: backbone/1.3.3+ds-1

[Jonas Smedegaard]

Quoting Gianfranco Costamagna (2016-08-28 17:59:41)
>>Do you intend to help maintain the package collaboratively, or take 
>>over maintenance?
> I see a team upload, so I guess the former of course

Don't take for granted that what you find obvious is equally obvious to 
others.  Especially when raised as an explicit question: Makes you seem 
like taking me for a fool (which I am sure was not intended).


>> If the former, then you should coordinate your work with current 
>> maintainers (fine that you seek help and guidance from others too, 
>> but don't go ahead with only input from external parties).
> 
> this is something that a good sponsor has to check, and the reason for 
> me putting *you* in cc :)

Could you please elaborate on what you find an obvious role for a 
"sponsor" here.  Assume I have never¹ hear that term before.


¹ I have, but from the context here I suspect we mean different things 
by that term.

[remaining mail skipped for now, to keep a focus]

 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: signature

Bug#835658: RFS: backbone/1.3.3+ds-1

[Jonas Smedegaard]

Quoting Julien Puydt (2016-08-28 16:12:33)
> On 28/08/2016 15:43, Jonas Smedegaard wrote:
>> Quoting Julien Puydt (2016-08-28 14:35:32)
>>> On 28/08/2016 10:00, Gianfranco Costamagna wrote:
>   I am looking for a sponsor for the package "backbone"
>>
>> Do you intend to help maintain the package collaboratively, or take 
>> over maintenance?
>
> Collaboratively.

Great!

...but perhaps we have different understandings of that term, then:

It seems you prepared a new package release with radical structural 
changes, and requested sponsorship for uploading it.  First I hear of 
all that is today, after the fact.  Did I miss some prior communication 
somehow?

If not, then please elaborate a bit on the kind of collaboration you 
want.

[holding back on answering the rest of the mail until above is more 
clear]


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: signature

Bug#835658: RFS: backbone/1.3.3+ds-1

[Gianfranco Costamagna]

Hi Jonas!

>Do you intend to help maintain the package collaboratively, or take over 
>maintenance?


I see a team upload, so I guess the former of course
>If the former, then you should coordinate your work with current 
>maintainers (fine that you seek help and guidance from others too, but 
>don't go ahead with only input from external parties).


this is something that a good sponsor has to check, and the reason for 

me putting *you* in cc :)

I hope the workflow was good enough for you, we didn't upload anything
and we are here to reach a common view, that might be beneficial
from people needing the new release
(I personally sponsored ~10 packages for the new ipython, and this one
seems needing an update too)

>If the latter, then be cautious that you either have consent from 
>current maintainers to do so, or ensure that you are following the 
>procedures in Debian for takeover (search list archives for "hijacking" 
>for more info).


never been the case here, but nice to put it clear, hijacking
is *never* something good, except for MIA maintainers.
Since you put the package under team maintenance, I guess a Team
Upload is something we can handle, right?
(of course, I wanted you to be eplicitly aware of this RFS, in my cc
and the main goal is to agree on something all together)


>Bugs requesting new upstream releases are wishlist bugs, and they are 
>targeted the maintainer of the package.
>
>Wishlist bugs being old is not a good reason to bypass the maintainer.


not this case, of course :)
He did the work, but "bypassing" the maintainer needs somebody to upload it :)
in case you want to upload by yourself, you might find some work already done,
and cherry-pick it ;)

>Each team collaborate differntly - don't assume it is ok to change a 
>package without coordinating with current maintainers (i.e. those listed 
>as "Uploaders" for team-maintained packages).


not sure if you mean the archive or the packaging repo...
but yeah, probably I would have done a git push on an external branch/repo
and pushed only after the current upload.
You are right, every team has a different workflow, and remembering all of them
and all of the different maintainers is something difficult...

But with git it is trivial to revert/cherry-pick/branch/delete stuff :)


>> That was the first time I met a cdbs package... isn't it deprecated?
>
>No.  Never was.
>
>Some people disliking CDBS wants everyone to believe that's the case.


I personally don't like CDBS, but I never had a war against it. I don't like
because probably I had a first approach without it, but I have NMUed packages
with it, and as I said, I preferred to be explicit about that point.

CDBS is used, probably not the first one by popcon, but it has reasons
to be there, and a Team Upload should probably respect that.
(BTW, it seems true that for a package that needs a bunch of files copied
over two directories, a ~100 lines rules file seems a little bit complicated
to read, do we really need all of this file?
(I'm interested in this part, I might find that CDBS has some nice features
and make me switch or partially switch to it)

>Please consult current maintainers of a package before making large 
>changes, like restructuring to a different core packaging helper tool.


this is true, but we cal always revert if it isn't ok for you I hope

>Please consult current maintainers of a package before making large 

>changes, like "starting from scratch".

true,

but at the end, would you consider sponsoring an updated package or not?
ok the change back to cdbs, ok the testing, but what are the points you
would like to see addressed (if eventually you want somebody else sponsoring
this one)

thanks for the quick reply, I know you are busy, and this is appreciated
(actually I remember me taking over of a package or two, some years ago,
they were RC-buggy and you gave them to me so nicely, I hope we will "fix" the
current situation in the same way, to help reverse-dependencies move
forward)

thanks,

G.

Bug#835658: RFS: backbone/1.3.3+ds-1

[Julien Puydt]

Hi,

On 28/08/2016 15:43, Jonas Smedegaard wrote:

Quoting Julien Puydt (2016-08-28 14:35:32)

On 28/08/2016 10:00, Gianfranco Costamagna wrote:

  I am looking for a sponsor for the package "backbone"


Do you intend to help maintain the package collaboratively, or take over
maintenance?


Collaboratively.


If the former, then you should coordinate your work with current
maintainers (fine that you seek help and guidance from others too, but
don't go ahead with only input from external parties).


Ok.


If the latter, then be cautious that you either have consent from
current maintainers to do so, or ensure that you are following the
procedures in Debian for takeover (search list archives for "hijacking"
for more info).


See above : team maintenance.


3) Did you get in touch with the maintainer for this upload?


Contrary to what I usually do, no : I saw two "please package new
upstream" bugs, both old (may 2014 and september 2015), none closed as
a duplicate of the other -- so I considered the "getting in touch"
step was done and pushed forward with the "team maintenance" step.


Bugs requesting new upstream releases are wishlist bugs, and they are
targeted the maintainer of the package.

>

Wishlist bugs being old is not a good reason to bypass the maintainer.

Each team collaborate differntly - don't assume it is ok to change a
package without coordinating with current maintainers (i.e. those listed
as "Uploaders" for team-maintained packages).


Ok.


5) I'm not sure Jonas will like to move away from cdbs...


That was the first time I met a cdbs package... isn't it deprecated?


No.  Never was.

Some people disliking CDBS wants everyone to believe that's the case.


I didn't dislike it since I thought it was deprecated. But I have to 
admit seeing what d/rules looked like for that simple a package doesn't 
make me want to learn more about it.



9)

"This source package uses CDBS"
I would change that :)


git rm README.source... but I seem to have forgotten a git commit -m
somewhere so it doesn't appear in history... drat.


Please consult current maintainers of a package before making large
changes, like restructuring to a different core packaging helper tool.


Ok.



we should be mostly complete as a preliminary review. I have to
admit, the cdbs rules file was a little bit complicate for a 3
installed js files library :)


I admit I didn't even try to understand what it did : just saw how big
it was, removed the file and started from scratch.


Please consult current maintainers of a package before making large
changes, like "starting from scratch".


Ok.

Jonas, do you want me to revert all my changes on backbone? Do you 
intend to do something about backbone? I notice that you're also 
(co-)maintainer for underscore, which is also on my way to jupyter in 
Debian. For that one there is an important bug which looks pretty 
trivial and a wishlist bug which is a one-liner to d/control [and it 
lacks a wishlist for a newer version]. What are your plans for 
underscore too?


Sorry for the inconvenience,

Snark on #debian-js

Bug#835658: RFS: backbone/1.3.3+ds-1

[Jonas Smedegaard]

Quoting Julien Puydt (2016-08-28 14:35:32)
> On 28/08/2016 10:00, Gianfranco Costamagna wrote:
>>>   I am looking for a sponsor for the package "backbone"

Do you intend to help maintain the package collaboratively, or take over 
maintenance?

If the former, then you should coordinate your work with current 
maintainers (fine that you seek help and guidance from others too, but 
don't go ahead with only input from external parties).

If the latter, then be cautious that you either have consent from 
current maintainers to do so, or ensure that you are following the 
procedures in Debian for takeover (search list archives for "hijacking" 
for more info).


>> 3) Did you get in touch with the maintainer for this upload?
>
> Contrary to what I usually do, no : I saw two "please package new 
> upstream" bugs, both old (may 2014 and september 2015), none closed as 
> a duplicate of the other -- so I considered the "getting in touch" 
> step was done and pushed forward with the "team maintenance" step.

Bugs requesting new upstream releases are wishlist bugs, and they are 
targeted the maintainer of the package.

Wishlist bugs being old is not a good reason to bypass the maintainer.

Each team collaborate differntly - don't assume it is ok to change a 
package without coordinating with current maintainers (i.e. those listed 
as "Uploaders" for team-maintained packages).


> > 5) I'm not sure Jonas will like to move away from cdbs...
> 
> That was the first time I met a cdbs package... isn't it deprecated?

No.  Never was.

Some people disliking CDBS wants everyone to believe that's the case.



> > 9)
> >
> > "This source package uses CDBS"
> > I would change that :)
> 
> git rm README.source... but I seem to have forgotten a git commit -m 
> somewhere so it doesn't appear in history... drat.

Please consult current maintainers of a package before making large 
changes, like restructuring to a different core packaging helper tool.


> > we should be mostly complete as a preliminary review. I have to 
> > admit, the cdbs rules file was a little bit complicate for a 3 
> > installed js files library :)
> 
> I admit I didn't even try to understand what it did : just saw how big 
> it was, removed the file and started from scratch.

Please consult current maintainers of a package before making large 
changes, like "starting from scratch".


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: signature

Bug#835658: RFS: backbone/1.3.3+ds-1

[Julien Puydt]

Hi,

On 28/08/2016 10:00, Gianfranco Costamagna wrote:

control: owner -1 !
control: tags -1 moreinfo

(ccing people who expressed interest in this update)

  I am looking for a sponsor for the package "backbone"


here we are, but I have some questions/issues:

1) why did you drop so much build-dependencies?


I don't know why there was that much - pbuilder is happy with what I put.


2) missing copyrights/licenses:
 * @license RequireJS 2.1.9 Copyright (c) 2010-2012, The Dojo Foundation All 
Rights Reserved.
 * Available via the MIT or new BSD license.

(c) 2009-2015 Jeremy Ashkenas, DocumentCloud and Investigative Reporters & 
Editors
and maybe more


I don't have the time to look into that just now.


3) Did you get in touch with the maintainer for this upload?


Contrary to what I usually do, no : I saw two "please package new 
upstream" bugs, both old (may 2014 and september 2015), none closed as a 
duplicate of the other -- so I considered the "getting in touch" step 
was done and pushed forward with the "team maintenance" step.



4) please tag the two bugs as pending, to avoid double work from other people
(e.g. zigo who expressed)


Done.


5) I'm not sure Jonas will like to move away from cdbs...


That was the first time I met a cdbs package... isn't it deprecated?


6)
-Author: Jonas Smedegaard 
+Author: Julien Puydt

I would use both people as authors :)


Ah, that is because I rewrote the patch by hand (didn't apply... perhaps 
just because they changed the order of the lines), then added mly usual 
header. Added back.



7) depends/buil-depends.
typo


Fixed.


8)
Bump standards-version.
To which version?


Added.


9)

"This source package uses CDBS"
I would change that :)


git rm README.source... but I seem to have forgotten a git commit -m 
somewhere so it doesn't appear in history... drat.



10)
why is this a ds repack? it seems to be not written in README.source
or whatever


Well, that is documented in debian/copyright, as usual:
Files-Excluded: *-min.js
and it's automatic, with debian/watch having the repacksuffix=+ds options.


11) compat level is still 8 :)


Fixed.


we should be mostly complete as a preliminary review.
I have to admit, the cdbs rules file was a little bit complicate for a 3 
installed
js files library :)


I admit I didn't even try to understand what it did : just saw how big 
it was, removed the file and started from scratch.



but I think we might go for experimental, wait some more testing for the two+ 
people needing
it and then go for unstable.


Done.


Does it sounds good as a plan


Yes.

I'll have to find the time to take care of d/copyright.

Thanks,

Snark on #debian-js

Bug#835658: RFS: backbone/1.3.3+ds-1

[Gianfranco Costamagna]

control: owner -1 !
control: tags -1 moreinfo


(ccing people who expressed interest in this update)
>   I am looking for a sponsor for the package "backbone"


here we are, but I have some questions/issues:

1) why did you drop so much build-dependencies?
2) missing copyrights/licenses:
 * @license RequireJS 2.1.9 Copyright (c) 2010-2012, The Dojo Foundation All 
Rights Reserved.
 * Available via the MIT or new BSD license.

(c) 2009-2015 Jeremy Ashkenas, DocumentCloud and Investigative Reporters & 
Editors
and maybe more
3) Did you get in touch with the maintainer for this upload?
4) please tag the two bugs as pending, to avoid double work from other people
(e.g. zigo who expressed)
5) I'm not sure Jonas will like to move away from cdbs...
6) 
-Author: Jonas Smedegaard 
+Author: Julien Puydt

I would use both people as authors :)

7) depends/buil-depends.
typo

8) 
Bump standards-version.
To which version?

9)

"This source package uses CDBS"
I would change that :)

10)
why is this a ds repack? it seems to be not written in README.source
or whatever

11) compat level is still 8 :)

we should be mostly complete as a preliminary review.
I have to admit, the cdbs rules file was a little bit complicate for a 3 
installed
js files library :)
but I think we might go for experimental, wait some more testing for the two+ 
people needing
it and then go for unstable.
Does it sounds good as a plan

Bug#835658: RFS: backbone/1.3.3+ds-1

[Julien Puydt]

Package: sponsorship-requests
Severity: normal

  Dear mentors,

  I am looking for a sponsor for the package "backbone"

 * Package name: backbone
   Version : 1.3.3+ds-1
   Upstream Author : Jeremy Ashkenas
 * URL : https://github.com/documentcloud/backbone/
 * License : Expat
   Section : web

  It builds those binary packages:

libjs-backbone - some Backbone for JavaScript applications - 
browser library

 node-backbone - some Backbone for JavaScript applications - Node module

  To access further information about this package, please visit the 
following URL:


  https://mentors.debian.net/package/backbone

  Alternatively, one can download the package with dget using this command:

dget -x 
https://mentors.debian.net/debian/pool/main/b/backbone/backbone_1.3.3+ds-1.dsc


  It is packaged in the Debian Javascript Maintainers team's repositories:
Vcs-Git: https://anonscm.debian.org/git/pkg-javascript/backbone
Vcs-Browser: https://anonscm.debian.org/cgit/pkg-javascript/backbone.git

I'm not the original packager and the package was two years old (still 
using cdbs!), so it took a bit of work to get something going ; 
hopefully I got it back on its feet.


The new changelog entry is:
backbone (1.3.3+ds-1) unstable; urgency=medium
 * Team upload.
 * New upstream release (Closes: #746787, #797698).
 * Bump standards-version.
 * Use https in Vcs-* fields.
 * Drop obsolete patch to support newer underscore.
 * Rewrite the patch to use Debian-packaged libraries.
 * Move away from cdbs.
 * Update d/copyright.
 * Rewrite depends/buil-depends.
 * Don't ship minimized files.

-- Julien Puydt   Sat, 27 Aug 2016 21:02:31 +0200


Cheers,

Snark on #debian-js