Lwt: Change license to MIT

[aantron]

Lwt is has an odd license situation, described below. To reduce confusion and make Lwt easier on users, I propose we change to the MIT license, or another such permissive license.

Please give permission

I have ccd all the contributors to Lwt further down. If you allow Lwt to change the license, please leave a comment on this issue. If you object, definitely please leave a comment :)

Anyone else is free to comment as well, of course.

I don't think we need literally every contributor to comment here. We will make public announcements, leave this open for a long while, and if all seems well, make the switch. But the more contributors comment, the clearer and cleaner the process.

Status

• The "main" body of the license is the LGPL 2.1.
• However, there is special text for use with OpenSSL: https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L1-L3
• ...and another block of extraordinary text: https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L5-L17
• In addition, Lwt_react is licensed under the 3-clause BSD license, which is what I believe this last bit of extra text refers to: https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L19-L20

cc @acieroid @agarwal @andrewray @avsm @bobbypriambodo @c-cube @cedlemo @chambart @copy @damiendoligez @dannywillems @dbuenzli @diml @dmbaturin @domsj @Drup @dsheets @edwintorok @emillon @fdopen @gabelevi @gasche @gdsfh @glondu @haesbaert @hannesm @hcarty @hhugo @hnrgrgr @iitalics @jasone @jjd27 @jpdeplaix @jsthomas @kandu @kennetpostigo @kerneis @Leonidas-from-XIV @leowzukw @malo-denielou @mfp @mor1 @nojb @olasd @orbifx @persianturtle @pqwy @pveber @raphael-proust @rgrinberg

[aantron]

Second comment because of cc limit.

cc @Richard-Degenne @rneswold @ryanslade @seliopou @simonjbeaumont @smithjessk @smondet @stijn-devriendt @talex5 @timbertson @toolslive @vbmithr @vouillon @whitequark @yallop @ygrek @ZackC @zoggy

[yallop]

I'm fine with this change.

[domsj]

ok for me :+1:

[persianturtle]

No objection here. 👍

[c-cube]

fine by me! :+1:

[Leonidas-from-XIV]

I'm ok with this change :+1:

[copy]

I'm ok with this change.

[Drup]

I'm pretty sure the license is in fact standard "LGPL + linking exception" which is common in the OCaml community. I'm very confused why Lwt_react is under a difference license, though.

That being said, I don't mind the change.

[smondet]

MIT license works for me :+1:

[ghost]

I'd prefer if the licence remained copyleft, such as LGPL.

[kit-ty-kate]

Fine by me

[jasone]

Fine by me.

[dmbaturin]

I don't see anything strange with the current situation. OpenSSL exception will fortunately become redundant soon as they are switching to a BSD license. The "extraordinary block of text" is there to permit free distribution of binaries statically linked with the library.

However, if you think MIT or equivalent will be better for the project, I have no objections.

I feel a bit uneasy about the third clause of the 3-clause BSD license though — while it was judged to be GPL-compatible, I feel like it may be discouraging innocous and even beneficial activities such as hanging a "powered by Lwt" banner on the project website. If we are at it, can we also get rid of the third clause?

[acieroid]

Fine by me 👍

[timbertson]

Fine by me :+1:

[seliopou]

I'm ok with this

[edwintorok]

I'm ok with this

[hcarty]

I'm happy with the license change 👍

[dbuenzli]

Ok for me.

[kerneis]

I'm ok with the change as far as my code is concerned (I'm not even sure
any of it has survived until today).

I don't think we need literally every contributor to comment here. We
will make public announcements, leave this open for a long while, and
if all seems well, make the switch. But the more contributors comment,
the clearer and cleaner the process.
But this is risky I think (note that i'm not a lawyer). See
https://www.videolan.org/press/lgpl-libvlc.html for an example of
relicensing; they reached out to all contributors, and rewrote the
code of the one they couldn't get agreement from.

[bobbypriambodo]

No problem for me. 👍

[whitequark]

I'l fine with this

[rgrinberg]

Relicense my code as you see fit. +1

On Wed, Mar 7, 2018, 7:42 AM whitequark [e-mail] wrote:

I'l fine with this

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ocsigen/lwt/issues/560#issuecomment-370981289, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAIe-7yZ57R9bioNvh8VjDHbI7ZMAA2Fks5tbyzjgaJpZM4Sffjo
.

[smithjessk]

Sounds good.

[nojb]

Ok for me.

[ryanslade]

Fine with me :+1:

[ygrek]

ok

[pveber]

I'm fine with the change, thanks!

[toolslive]

ok.

[gasche]

For me the main question is: the people who created Lwt (and decided on the license) and, before you, developed most of it (I would think of @vouillon and @diml, do they agree with the proposed license change? (I would expect that this was discussed with them beforehand.)

On my own contribution, I am of course fine with using the MIT license.

[emillon]

I'm fine with relicensing to MIT 👍

[raphael-proust]

Ok

[talex5]

OK

[jeremiedimino]

fine by me

[haesbaert]

Fine by me.

Just a note, usually you only ask permission to change the license from
copyright holders, so people who actually have their name in the header of
a file.

People like me who just sent one or two diffs, do not hold copyrights and
do not need to gove consent.

As an example, in the past the usual process for relicensing bsd3 to bsd2
was done on a file by file basis, getting approval from each copyright
holder.

My point is, i have no claims on the license since I dont hold any
copyrights, but at any rate if I did, I’d be fine.

[mor1]

I concur with @haesbaert on this.
But in any case: no objection to relicensing as MIT.
(I'd forgotten I'd even made a contribution!)

[fdopen]

✓

[jsthomas]

Fine by me. 👍

[dmbaturin]

@haesbaert While I may agree with your opinion _philosophically_ (though I have reservations — if I wanted to do what @aantron is doing with Lwt license, I would also ask every contributor), _legally_, it's simply not the case. In the countries that ratified the Berne convention (which is almost all countries in the world by now), you automatically become a copyright owner when you create something, whether you want it or not.
A copyright notice in the file header or any other form is not required. Simply publishing it without your name on it doesn't wave copyright protection.
In fact, in most countries there is absolutely no way to give a public domain status to your own work other than by using a license that allows everything — this is why absolutely permissive licenses such as WTFPL or CC-0 exist.

This is why re-licensing is so tricky and many projects avoid it. Some organizations require that you give them exclusive rights to your code to avoid the issues, but that's open to abuse, and some of them are known to have used it to change the license to proprietary and lock the contributors out of their own code. I'm glad most don't.

[malo-denielou]

Ok for me to go to MIT license.

On Wed, Mar 7, 2018, 07:06 Daniil Baturin [e-mail] wrote:

@haesbaert https://github.com/haesbaert While I may agree with your
opinion philosophically (though I have reservations — if I wanted to do
what @aantron https://github.com/aantron is doing with Lwt license, I
would also ask every contributor), legally, it's simply not the case.
In the countries that ratified the Berne convention (which is almost all
countries in the world by now), you automatically become a copyright owner
when you create something, whether you want it or not.
A copyright notice in the file header or any other form is not required.
Simply publishing it without your name on it doesn't wave copyright
protection.
In fact, in most countries there is absolutely no way to give a public
domain status to your own work other than by using a license that allows
everything — this is why absolutely permissive licenses such as WTFPL or
CC-0 exist.

This is why re-licensing is so tricky and many projects avoid it. Some
organizations require that you give them exclusive rights to your code to
avoid the issues, but that's open to abuse, and some of them are known to
have used it to change the license to proprietary and lock the contributors
out of their own code. I'm glad most don't.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ocsigen/lwt/issues/560#issuecomment-371167537, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AA75gzHN5SlPyyAIfH7yEO53leTiDLxDks5tb_dSgaJpZM4Sffjo
.

[cedlemo]

ok

[gabelevi]

👍

[simonjbeaumont]

👍

[vbmithr]

Le 06/03/2018 à 22:50, Anton Bachin a écrit :

Second comment because of cc limit.

Ok for me.

--
Vincent

[iitalics]

ok

[mfp]

Fine by me

--
Mauricio Fernández

[hannesm]

fine with me, highly appreciated! <3

[avsm]

Fine by me, and also highly appreciated!

[aantron]

@dmbaturin I'd like to change Lwt_react to MIT as well, so we won't have the BSD license at all anymore.

@haesbaert Indeed, I usually wouldn't expect to be asked about most contributions I've made to projects. But on the other hand, some people have made considerable contributions to Lwt, without making or being made aware of any statement or agreement about copyright, and/or forgetting (both contributor and reviewer) to change or create the license header, etc. So there seems to be a considerable gray area. It seems reasonable to check thoroughly for objections and other opinions, as suggested by @kerneis and @dmbaturin, even if we expect that most contributors generously don't believe themselves to have rights over their contributions.

[agarwal]

I'm okay with the change.

[kandu]

fine by me

[andrewray]

Fine by me.

On 6 Mar 2018 21:49, "Anton Bachin" [e-mail] wrote:

Lwt is has an odd license situation, described below. To reduce confusion
and make Lwt easier on users, I propose we change to the MIT license, or
another such permissive license.

Please give permission

I have ccd all the contributors to Lwt further down. If you allow Lwt to
change the license, please leave a comment on this issue. If you object,
definitely please leave a comment :)

Anyone else is free to comment as well, of course.

I don't think we need literally every contributor to comment here. We will
make public announcements, leave this open for a long while, and if all
seems well, make the switch. But the more contributors comment, the clearer
and cleaner the process.

Status

• The "main" body of the license
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING
  is the LGPL 2.1.
• However, there is special text for use with OpenSSL:
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a32
  1256a1cfeb/doc/COPYING#L1-L3
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L1-L3
• ...and another block of extraordinary text:
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a32
  1256a1cfeb/doc/COPYING#L5-L17
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L5-L17
• In addition, Lwt_react
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/src/react/lwt_react.mli
  is licensed under the 3-clause BSD license, which is what I believe this
  last bit of extra text refers to: https://github.com/ocsigen/lwt/blob/
  c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L19-L20
  https://github.com/ocsigen/lwt/blob/c3cdfb303663cb1a1733244da53a321256a1cfeb/doc/COPYING#L19-L20

cc @acieroid https://github.com/acieroid @agarwal
https://github.com/agarwal @andrewray https://github.com/andrewray
@avsm https://github.com/avsm @bobbypriambodo
https://github.com/bobbypriambodo @c-cube https://github.com/c-cube
@cedlemo https://github.com/cedlemo @chambart
https://github.com/chambart @copy https://github.com/copy
@damiendoligez https://github.com/damiendoligez @dannywillems
https://github.com/dannywillems @dbuenzli https://github.com/dbuenzli
@diml https://github.com/diml @dmbaturin https://github.com/dmbaturin
@domsj https://github.com/domsj @Drup https://github.com/drup @dsheets
https://github.com/dsheets @edwintorok https://github.com/edwintorok
@emillon https://github.com/emillon @fdopen https://github.com/fdopen
@gabelevi https://github.com/gabelevi @gasche
https://github.com/gasche @gdsfh https://github.com/gdsfh @glondu
https://github.com/glondu @haesbaert https://github.com/haesbaert
@hannesm https://github.com/hannesm @hcarty https://github.com/hcarty
@hhugo https://github.com/hhugo @hnrgrgr https://github.com/hnrgrgr
@iitalics https://github.com/iitalics @jasone
https://github.com/jasone @jjd27 https://github.com/jjd27 @jpdeplaix
https://github.com/jpdeplaix @jsthomas https://github.com/jsthomas
@kandu https://github.com/kandu @kennetpostigo
https://github.com/kennetpostigo @kerneis https://github.com/kerneis
@Leonidas-from-XIV https://github.com/leonidas-from-xiv @leowzukw
https://github.com/leowzukw @malo-denielou
https://github.com/malo-denielou @mfp https://github.com/mfp @mor1
https://github.com/mor1 @nojb https://github.com/nojb @olasd
https://github.com/olasd @orbifx https://github.com/orbifx
@persianturtle https://github.com/persianturtle @pqwy
https://github.com/pqwy @pveber https://github.com/pveber
@raphael-proust https://github.com/raphael-proust @rgrinberg
https://github.com/rgrinberg

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ocsigen/lwt/issues/560, or mute the thread
https://github.com/notifications/unsubscribe-auth/AFqzIxsFief1Bu9CAf8l9JTvGwjJbNvoks5tbwSDgaJpZM4Sffjo
.

[stijn-devriendt]

Approved by our legal team.
Otherwise, fine by me too.

[damiendoligez]

OK for me for the license change.

@dmbaturin I was once asked to sign a Contributor Licence Agreement for a contribution that consisted entirely of removing one character from a source file. Do you think such a contribution is (or should be) covered by copyright? I don't.

[dmbaturin]

@damiendoligez People like SCO are constantly trying to push it as far as they can, and in the civil law it's the defendant who has to prove the claim wrong — I can see why people may want to stay completely on the safe side. Anyway, my own position is more of "be conservative in what you do and liberal in what you accept". I'm fine with my own contributions to Lwt being relicensed whether they are big or small, but as a maintainer I wouldn't assume that everyone else will be.

Still, it's an interesting philosophical question, how physical size of a contribution relates to its importance. A person who fixed a remotely exploitable vulnerability by removing one character might have contributed more to the project than someone who wrote thousands lines of useful but non-essential code — whether the change itself can be copyrighted or not, I would be inclined to ask their opinion before making license changes.

[rneswold]

Fine by me.

[pqwy]

I don't think any of my code is in Lwt, so I also don't believe I have rights over it.

I'm fine with this change either way.

[Richard-Degenne]

Sorry I haven't checked out Lwt in a while.

It's all right with me! :)

[vouillon]

I'm okay with the change.

[olasd]

Hi,

I'm fine with the change.

Thanks!
Nicolas

[dsheets]

This is great news. I'm happy with this change.

[chambart]

I'm ok, no objections.

[aantron]

Thank you all, the license is now MIT :)

The final status was:

• Almost all contributors and organizations involved in Lwt development agreed.
• Some of the agreements were communicated by email or in Gitter.
• There were no objections.

• A few (~5) contributors did not respond, but in each of those cases, several of the following hold:

  • The contributions are of such a nature, that we think most contributors would not consider themselves to have rights over them. Precedent seems to agree at least in the United States.
  • The contributions have since been largely overwritten by other contributions.
  • The contributions were to parts of Lwt that have since been factored out into other repos, whose licenses we aren't changing (lwt_camlp4, lwt_ssl, etc).

Since we didn't get explicit assent from a few contributors, however, I think it's best not to apply MIT retroactively to past Lwt releases. In some of them, the contributions we don't have assent for were still current. So, the MIT license applies to the next release of Lwt, and to master starting from the commit that closed this issue.

[ZackC]

I’m fine with it if you need consent.

I didn’t think my contributions were enough where I should have any influence on the situation.

Zack Coker

On Apr 23, 2018, at 7:25 PM, Anton Bachin [e-mail] wrote:

Thank you all, the license is now MIT :)

The final status was:

Almost all contributors and organizations involved in Lwt development agreed.

Some of the agreements were communicated by email or in Gitter.

There were no objections.

A few (~5) contributors did not respond, but in each of those cases, several of the following hold:

The contributions are of such a nature, that we think most contributors would not consider themselves to have rights over them. Precedent seems to agree at least in the United States.
The contributions have since been largely overwritten by other contributions.
The contributions were to parts of Lwt that have since been factored out into other repos, whose licenses we aren't changing (lwt_camlp4, lwt_ssl, etc).
Since we didn't get explicit assent from a few contributors, however, I think it's best not to apply MIT retroactively to past Lwt releases. In some of them, the contributions we don't have assent for were still current. So, the MIT license applies to the next release of Lwt, and to master starting from the commit that closed this issue.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub https://github.com/ocsigen/lwt/issues/560#issuecomment-383753759, or mute the thread https://github.com/notifications/unsubscribe-auth/ABJTj_5h11QWf3R8bIJpNcpGV8HxIqq5ks5trmLzgaJpZM4Sffjo.

[aantron]

Thank you nonetheless :)

[tychota]

The readme badge still display lgpl, i think.

[tychota]

582 - > update the readme badge too.

[AkihiroSuda]

Could you release 4.0.2 with the new license?

[aantron]

@AkihiroSuda Yes, will do so in the coming days. I guess it will be 4.1.0, because we added a few things.

[aantron]

@AkihiroSuda Lwt 4.1.0 is now out. I'll make a broader announcement shortly as well.