On 06/03/2012 18:54, Jonas Smedegaard wrote:
> On 12-03-06 at 06:32pm, Jérémy Lal wrote:
>> On 06/03/2012 18:06, Jonas Smedegaard wrote:
>>> On 12-03-06 at 05:40pm, Jérémy Lal wrote:
>>>> On 06/03/2012 17:14, Felipe Sateler wrote:
>>>>> I might (hopefully...) have time to help, and I will need this. Is
>>>>> there any roadmap on what is missing to get an updated npm
>>>>> package?
>>>>
>>>> There is :
>>>>
>>>> * package dependencies that are in node_modules
>>>> * make sure the license is free (MIT +no-false-attribs)
>>>>
>>>> I update regularly the collab-maint git repository, and npm 1.1.4
>>>> package built from it is usable.
>>>
>>> Related to that: Please don't update in git the copyright_hints file
>>> if changes not reflected in copyirhgt file. The very point of that
>>> hints file is to track changes.
>>
>> In the case of npm, i intended to ignore changes in ./node_modules
>> since what's inside must be either reviewed or excluded from tarball.
>>
>> I usually check all new files, and differences, manually.
>> I am not perfect at this, though.
>
> Ah, come on - this is bigger than nitpicking, so not about perfection:
> Silencing warnings before solving the underlying problem is wrong IMO.
> Makes sense to ignore tracking excluded parts only when truly excluded,
> not when deciding to exlude in a future packaging release.
>
>>> Also, please do not blindly bump format hint in first line of
>>> copyright file. License field is not allowed to contain spaces in
>>> final release of the format.
>>
>> Lintian warnings about copyright format let me naively think the
>> syntax was strictly checked...
>
> Lintian should only ever be considered to help catch rough edges you
> might have otherwise missed, i.e. helps improve quality of packaging but
> not to be trusted to _assure_ high quality packaging.
>
>>> Oh, and that fork or Expat license seems to have a flaw: It is not
>>> clear whether it talks only about the original author or any
>>> subsequent author. I would recommend upstream to not try hack
>>> legalese but instead simply document clearly a friendly _request_ to
>>> do do same as now codified in license.
>>
>> I will forward that remark and recommandation to Isaac.
>> He's hard to convince for now.
>
> Yeah, I got the impression that he holds strong principles. I tried
> elaborate a bit, hope it makes sense also to him. :-)
His reply is attached.
Jérémy.
--- Begin Message ---
I disagree with the argument that it's ambiguous. It's defined in the
first line:
> Copyright 2009, 2010, 2011 Isaac Z. Schlueter (the "Author")
Nevertheless, I clarified it by changing all instances to "Original
Author".
On 06/03/2012 18:06, Jonas Smedegaard wrote:
> I would recommend upstream to not try hack legalese but instead
> simply document clearly a friendly _request_ to do do same as now
> codified in license.
This recommendation does not surprise me. However, it is clearly in
Debian's interest, and not in mine, so it is not reasonable for me to
comply with it.
The goal is to prevent distros from clobbering my software and letting me
handle the fallout. Friendly requests in the past have gone unheeded by
several different groups, some of which asserted that they have the right
to direct bug reports to me, claiming that it's *my* responsibility to make
my software work with their distribution (after they've modified it without
my knowledge!)
The only thing that distros pay any attention to is LICENSE files, so that
what I use here. (Evidenced clearly by the degree of attention that has
been paid to it in this case - would anyone care if it was a plain old
MIT?) If a particular person or distribution would like a special
dispensation to disable or alter features in npm, and to then distribute
their modified copy without changing the name, then they may ask for that
directly, and we can perhaps work something out, whereby they take
ownership of their changes, clearly communicate them to users, and perhaps
even rebrand the software as a downstream fork.
If a distro wishes to alter or disable features of npm, and does *not* wish
to take ownership of their changes, then it would be better for me if they
did not include npm in their distribution. Linux users can already install
npm quite easily from source. Debian users can get it from Chris Lea's
PPA, which does not alter the source code, and thus has no problem
complying with the license. Mac and PC users can get it automatically
along with the node binary installers. Anyone who installs node from
source gets it by default.
In other words, if the terms of this license keep npm out of Debian Stable,
or any particular distro, then that means it's working. The fact that npm
is not in the distro is worse for the distro than it is for npm.
On Tue, Mar 6, 2012 at 09:45, Jérémy Lal <[email protected]> wrote:
> Hi,
> might this be a good argument ? :
>
> On 06/03/2012 18:06, Jonas Smedegaard wrote:
> > I would recommend upstream to not try hack legalese but instead
> > simply document clearly a friendly _request_ to do do same as now
> > codified in license.
>
> Cheers,
> Jérémy.
>
>
>
> On 24/01/2012 21:03, Isaac Schlueter wrote:
> > I don't mind the spam at all :)
> >
> > The three-clause BSD is not appropriate in this case. I'm not trying
> > to prevent unauthorized indications of endorsements, but rather user
> > confusion by causing bugs or removing functionality, and inadvertently
> > attributing the change to me.
> >
> > The word "Author" is defined in the first line. But, of course, I'm
> > open to changing the language of the license, or using a different one
> > entirely, so long as it retains the key features I need from it. I
> > haven't been able to find an existing license that meets npm's needs
> > exactly. If there's one that I've overlooked, or some edits to the
> > language that could make it less problematic to Debian, definitely let
> > me know.
> >
> > Thanks.
> >
> >
> >
> > On Tue, Jan 24, 2012 at 08:00, Jérémy Lal <[email protected]> wrote:
> >> Very sorry for the spam...
> >> This is the start of thread :
> >> http://lists.debian.org/debian-legal/2012/01/msg00104.html
> >>
> >> I won't disturb anymore, unless you ask to :)
> >>
> >> Jérémy.
> >>
> >>
> >>
> >> -------- Original Message --------
> >> Subject: Re: MIT +no-false-attribs
> >> Resent-Date: Tue, 24 Jan 2012 15:50:46 +0000 (UTC)
> >> Resent-From: [email protected]
> >> Date: Tue, 24 Jan 2012 15:49:41 +0000 (UTC)
> >> From: Tanguy Ortolo <[email protected]>
> >> Organization: Individual
> >> To: [email protected]
> >>
> >> Jérémy Lal, 2012-01-24 01:55+0100:
> >>> I will, and concur. But knowing upstream i can tell he'll need
> stronger arguments.
> >>
> >> The 3-clause BSD license would seem to be appropriate considering what
> >> the author apparently wants.
> >>
> >> --
> >> ,--.
> >> : /` ) Tanguy Ortolo <xmpp:[email protected]> <irc://
> irc.oftc.net/Tanguy>
> >> | `-' Debian Developer
> >> \_
> >>
>
>
--- End Message ---
