Hackage suggestion: Gather the list of the licenses of all dependencies of a package

newer
Can cabal be turned into a package...

Petr P

13 Dec 2012 13 Dec '12

5:41 a.m.

Dear Haskellers, following up the recent discussion about copyleft licenses, I'd suggest a (hopefully minor) improvement of Hackage: For each package, gather the list of the licenses of everything it depends on. I think this would help considerably people who don't want or can't use software licensed under a particular license (most often (L)GPL). In particular, we can have a BSD package that depends on a LGPL package, and this is fine for FOSS developers. But for a commercial developer, this can be a serious issue that is not apparent until one examines *every* transitive dependency. This idea is a bit vague, because a dependency is actually a range of packages, which in theory could have different licenses. But I suppose this will rarely happen in practice, so it'd be safe just to take the last package in the range (or maybe take all licences of the packages in the range). Best regards, Petr

Attachments:

attachment.html (text/html — 1018 bytes)

Show replies by date

Henk-Jan van Tuyl

13 Dec 13 Dec

7:49 a.m.

On Thu, 13 Dec 2012 11:41:14 +0100, Petr P wrote:

...

For each package, gather the list of the licenses of everything it depends on. I think this would help considerably people who don't want or can't use software licensed under a particular license (most often (L)GPL). In particular, we can have a BSD package that depends on a LGPL package, and this is fine for FOSS developers. But for a commercial developer, this can be a serious issue that is not apparent until one examines *every* transitive dependency.

This idea is a bit vague, because a dependency is actually a range of packages, which in theory could have different licenses. But I suppose this will rarely happen in practice, so it'd be safe just to take the last package in the range (or maybe take all licences of the packages in the range).

cab[0] can do that, for installed packages: cab deps -i -r -a vector will generate a list of licenses for the packages that vector depends upon, like this: base 4.3.1.0 BSD3 "" ghc-prim 0.2.0.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" integer-gmp 0.2.0.3 BSD3 "" ghc-prim 0.2.0.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" primitive 0.4.0.1 BSD3 "Roman Leshchinskiy " base 4.3.1.0 BSD3 "" ghc-prim 0.2.0.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" integer-gmp 0.2.0.3 BSD3 "" ghc-prim 0.2.0.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" ghc-prim 0.2.0.0 BSD3 "" rts 1.0 BSD3 "" ffi 1.0 BSD3 "" Regards, Henk-Jan van Tuyl [0] http://hackage.haskell.org/package/cab -- http://Van.Tuyl.eu/ http://members.chello.nl/hjgtuyl/tourdemonad.html Haskell programming --

Michael Snoyman

7:51 a.m.

I think that's a great idea. I just implemented this on PackDeps: http://packdeps.haskellers.com/licenses As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future. Michael On Thu, Dec 13, 2012 at 12:41 PM, Petr P wrote:

...

Dear Haskellers,

following up the recent discussion about copyleft licenses, I'd suggest a (hopefully minor) improvement of Hackage: For each package, gather the list of the licenses of everything it depends on. I think this would help considerably people who don't want or can't use software licensed under a particular license (most often (L)GPL). In particular, we can have a BSD package that depends on a LGPL package, and this is fine for FOSS developers. But for a commercial developer, this can be a serious issue that is not apparent until one examines *every* transitive dependency.

This idea is a bit vague, because a dependency is actually a range of packages, which in theory could have different licenses. But I suppose this will rarely happen in practice, so it'd be safe just to take the last package in the range (or maybe take all licences of the packages in the range).

Best regards, Petr

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Vincent Hanquez

8:53 a.m.

On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...

I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future.

awesome Michael ! However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes". -- Vincent

Michael Snoyman

9:03 a.m.

On Thu, Dec 13, 2012 at 3:53 PM, Vincent Hanquez wrote:

...

On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...
I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.**com/licenses http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future.

awesome Michael !

However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes".

-- Vincent

Hmm, that's a good point. I'll admit I hadn't really thought this through, but I can actually see an argument going both ways on this: * Viral licenses won't actually affect you if they're just used for test suites. * But company lawyers will probably be nervous about it anyway. Nonetheless, I think you have the right of it. Unless people say otherwise, I'm going to implement Vincent's change. Michael

Felipe Almeida Lessa

9:12 a.m.

While you're at it, maybe whitelisting cpphs would be nice as well =). On Thu, Dec 13, 2012 at 12:03 PM, Michael Snoyman wrote:

...

On Thu, Dec 13, 2012 at 3:53 PM, Vincent Hanquez wrote:

...
On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...
I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future.

awesome Michael !

However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes".

-- Vincent

Hmm, that's a good point. I'll admit I hadn't really thought this through, but I can actually see an argument going both ways on this:

* Viral licenses won't actually affect you if they're just used for test suites. * But company lawyers will probably be nervous about it anyway.

Nonetheless, I think you have the right of it. Unless people say otherwise, I'm going to implement Vincent's change.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

-- Felipe.

Michael Snoyman

10:08 a.m.

Are you referring to: http://code.haskell.org/cpphs/LICENCE-commercial If the package is dual-licensed BSD3 and LGPL, maybe Malcolm could change the cabal file to mention the BSD3 so that its package description is less intimidating? On Thu, Dec 13, 2012 at 4:12 PM, Felipe Almeida Lessa < felipe.lessa@gmail.com> wrote:

...

While you're at it, maybe whitelisting cpphs would be nice as well =).

On Thu, Dec 13, 2012 at 12:03 PM, Michael Snoyman wrote:

...
On Thu, Dec 13, 2012 at 3:53 PM, Vincent Hanquez wrote:

...
On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...
I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as

...
...
...
as Hackage incorporates the feature in the future.

awesome Michael !

However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes".

-- Vincent

Hmm, that's a good point. I'll admit I hadn't really thought this

soon through,

...
but I can actually see an argument going both ways on this:

* Viral licenses won't actually affect you if they're just used for test suites. * But company lawyers will probably be nervous about it anyway.

Nonetheless, I think you have the right of it. Unless people say otherwise, I'm going to implement Vincent's change.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

-- Felipe.

Felipe Almeida Lessa

11:37 a.m.

...

From [1] I gather that its license really is LGPL/GPL. However, when used as a preprocessor its license doesn't really matter. Many packages on that list have a LGPL "taint" because one of its deps use cpphs. So the whitelist of cpphs would be stating that nobody is using cpphs as a library (which may be false, but is mostly true ;).

[1] http://code.haskell.org/cpphs/README On Thu, Dec 13, 2012 at 1:08 PM, Michael Snoyman wrote:

...

Are you referring to:

http://code.haskell.org/cpphs/LICENCE-commercial

If the package is dual-licensed BSD3 and LGPL, maybe Malcolm could change the cabal file to mention the BSD3 so that its package description is less intimidating?

On Thu, Dec 13, 2012 at 4:12 PM, Felipe Almeida Lessa wrote:

...
While you're at it, maybe whitelisting cpphs would be nice as well =).

On Thu, Dec 13, 2012 at 12:03 PM, Michael Snoyman wrote:

...
On Thu, Dec 13, 2012 at 3:53 PM, Vincent Hanquez wrote:

...
On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...
I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future.

awesome Michael !

However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes".

-- Vincent

Hmm, that's a good point. I'll admit I hadn't really thought this through, but I can actually see an argument going both ways on this:

* Viral licenses won't actually affect you if they're just used for test suites. * But company lawyers will probably be nervous about it anyway.

Nonetheless, I think you have the right of it. Unless people say otherwise, I'm going to implement Vincent's change.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

-- Felipe.

-- Felipe.

Michael Snoyman

1:40 p.m.

I'm not quite certain what to make of: If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial). It seems like that's saying "if you really want to, use the BSD license instead." But I'm not sure what the legal meaning of "If you have a commercial use" is. Malcolm: could you clarify what the meaning is? On Thu, Dec 13, 2012 at 6:37 PM, Felipe Almeida Lessa < felipe.lessa@gmail.com> wrote:

...

From [1] I gather that its license really is LGPL/GPL. However, when used as a preprocessor its license doesn't really matter. Many packages on that list have a LGPL "taint" because one of its deps use cpphs. So the whitelist of cpphs would be stating that nobody is using cpphs as a library (which may be false, but is mostly true ;).

[1] http://code.haskell.org/cpphs/README

On Thu, Dec 13, 2012 at 1:08 PM, Michael Snoyman wrote:

...
Are you referring to:

http://code.haskell.org/cpphs/LICENCE-commercial

If the package is dual-licensed BSD3 and LGPL, maybe Malcolm could change the cabal file to mention the BSD3 so that its package description is less intimidating?

On Thu, Dec 13, 2012 at 4:12 PM, Felipe Almeida Lessa wrote:

...
While you're at it, maybe whitelisting cpphs would be nice as well =).

On Thu, Dec 13, 2012 at 12:03 PM, Michael Snoyman wrote:

...
On Thu, Dec 13, 2012 at 3:53 PM, Vincent Hanquez

wrote:

...
...
...
On 12/13/2012 12:51 PM, Michael Snoyman wrote:

...
I think that's a great idea. I just implemented this on PackDeps:

http://packdeps.haskellers.com/licenses

As with all features on that site, I'll be happy to deprecate it as soon as Hackage incorporates the feature in the future.

awesome Michael !

However i think ithis shouldn't take dependencies from tests and benchmarks. This doesn't make differences for the "overall" license that the library "exposes".

-- Vincent

Hmm, that's a good point. I'll admit I hadn't really thought this through, but I can actually see an argument going both ways on this:

* Viral licenses won't actually affect you if they're just used for test suites. * But company lawyers will probably be nervous about it anyway.

Nonetheless, I think you have the right of it. Unless people say otherwise, I'm going to implement Vincent's change.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

-- Felipe.

-- Felipe.

Daniel Trstenjak

2:51 p.m.

On Thu, Dec 13, 2012 at 08:40:09PM +0200, Michael Snoyman wrote:

...

If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial).

I think that depedencies to binaries, like cpphs, should be treated differently than depedencies to libraries, because using a (L)GPL-ed binary mostly hasn't any implications for a "commercial" user and also for the output of a (L)GPL-ed binary usually the (L)GPL doesn't apply.

Michael Snoyman

2:56 p.m.

On Thu, Dec 13, 2012 at 9:51 PM, Daniel Trstenjak < daniel.trstenjak@gmail.com> wrote:

...

On Thu, Dec 13, 2012 at 08:40:09PM +0200, Michael Snoyman wrote:

...
If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial).

I think that depedencies to binaries, like cpphs, should be treated differently than depedencies to libraries, because using a (L)GPL-ed binary mostly hasn't any implications for a "commercial" user and also for the output of a (L)GPL-ed binary usually the (L)GPL doesn't apply.

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

In the case of cpphs, there's no way to determine that we're using it as a library or an executable, since it's just listed in the build-depends. Michael

Petr P

15 Dec 15 Dec

2:13 a.m.

This is strange, I thought that cpphs should be specified in "build-tools:", not in "build-depends:". < http://www.haskell.org/cabal/users-guide/developing-packages.html#build-info...

...

Best regards, Petr 2012/12/13 Michael Snoyman

...

On Thu, Dec 13, 2012 at 9:51 PM, Daniel Trstenjak < daniel.trstenjak@gmail.com> wrote:

...
On Thu, Dec 13, 2012 at 08:40:09PM +0200, Michael Snoyman wrote:

...
If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial).

I think that depedencies to binaries, like cpphs, should be treated differently than depedencies to libraries, because using a (L)GPL-ed binary mostly hasn't any implications for a "commercial" user and also for the output of a (L)GPL-ed binary usually the (L)GPL doesn't apply.

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

In the case of cpphs, there's no way to determine that we're using it as a library or an executable, since it's just listed in the build-depends.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Brent Yorgey

8:41 a.m.

On Sat, Dec 15, 2012 at 08:13:44AM +0100, Petr P wrote:

...

This is strange, I thought that cpphs should be specified in "build-tools:", not in "build-depends:". < http://www.haskell.org/cabal/users-guide/developing-packages.html#build-info...

...
Best regards, Petr

Presumably the reason to list it in build-depends instead of build-tools is that in the latter case cabal will not automatically install it as a dependency. But you are right that this is a strange situation, since if it is being used only as a preprocessor, semantically it ought to be listed in build-tools. -Brent

...

2012/12/13 Michael Snoyman

...
On Thu, Dec 13, 2012 at 9:51 PM, Daniel Trstenjak < daniel.trstenjak@gmail.com> wrote:

...
On Thu, Dec 13, 2012 at 08:40:09PM +0200, Michael Snoyman wrote:

...
If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial).

I think that depedencies to binaries, like cpphs, should be treated differently than depedencies to libraries, because using a (L)GPL-ed binary mostly hasn't any implications for a "commercial" user and also for the output of a (L)GPL-ed binary usually the (L)GPL doesn't apply.

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

In the case of cpphs, there's no way to determine that we're using it as a library or an executable, since it's just listed in the build-depends.

Michael

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

...

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Petr P

9:01 a.m.

2012/12/15 Brent Yorgey

...

On Sat, Dec 15, 2012 at 08:13:44AM +0100, Petr P wrote:

...
This is strange, I thought that cpphs should be specified in "build-tools:", not in "build-depends:". <

http://www.haskell.org/cabal/users-guide/developing-packages.html#build-info...

...
...
Best regards, Petr

Presumably the reason to list it in build-depends instead of build-tools is that in the latter case cabal will not automatically install it as a dependency. But you are right that this is a strange situation, since if it is being used only as a preprocessor, semantically it ought to be listed in build-tools.

So if I put cpphs into build-tools and I don't have it installed, the build will fail? Is this a desired behavior, or a bug? Best regards, Petr

Brandon Allbery

10:14 a.m.

On Sat, Dec 15, 2012 at 9:01 AM, Petr P wrote:

...

So if I put cpphs into build-tools and I don't have it installed, the build will fail? Is this a desired behavior, or a bug?

Shortcoming of cabal; it only "knows about" libraries because it is really just a front-end for ghc-pkg, so can't really figure dependencies involving only executables and not libraries. The same thing comes up with happy, alex, and gtk2hsc2hs. -- brandon s allbery kf8nh sine nomine associates allbery.b@gmail.com ballbery@sinenomine.net unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

Henk-Jan van Tuyl

4:38 p.m.

On Sat, 15 Dec 2012 16:14:59 +0100, Brandon Allbery wrote:

...

On Sat, Dec 15, 2012 at 9:01 AM, Petr P wrote:

...
So if I put cpphs into build-tools and I don't have it installed, the build will fail? Is this a desired behavior, or a bug?

Shortcoming of cabal; it only "knows about" libraries because it is really just a front-end for ghc-pkg, so can't really figure dependencies involving only executables and not libraries. The same thing comes up with happy, alex, and gtk2hsc2hs.

A work-around for this would be to add a dummy library to program-only packages. Regards, Henk-Jan van Tuyl -- http://Van.Tuyl.eu/ http://members.chello.nl/hjgtuyl/tourdemonad.html Haskell programming --

Brandon Allbery

4:56 p.m.

On Sat, Dec 15, 2012 at 4:38 PM, Henk-Jan van Tuyl wrote:

...

On Sat, 15 Dec 2012 16:14:59 +0100, Brandon Allbery wrote:

...
On Sat, Dec 15, 2012 at 9:01 AM, Petr P wrote:

...
So if I put cpphs into build-tools and I don't have it installed, the build will fail? Is this a desired behavior, or a bug?

Shortcoming of cabal; it only "knows about" libraries because it is really

A work-around for this would be to add a dummy library to program-only packages.

But then what do you do about the current case, where the library imports a potentially dubious license dependency? -- brandon s allbery kf8nh sine nomine associates allbery.b@gmail.com ballbery@sinenomine.net unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

Malcolm Wallace

9:15 a.m.

On 13 Dec 2012, at 18:40, Michael Snoyman wrote:

...

I'm not quite certain what to make of:

If you have a commercial use for cpphs, and feel the terms of the (L)GPL are too onerous, you have the option of distributing unmodified binaries (only, not sources) under the terms of a different licence (see LICENCE-commercial).

It seems like that's saying "if you really want to, use the BSD license instead." But I'm not sure what the legal meaning of "If you have a commercial use" is. Malcolm: could you clarify what the meaning is?

No, the LICENCE-commercial is not BSD. Read it more carefully. :-) So, I dual-licensed cpphs (which was originally only LGPL as a library, GPL as a binary), in response to a request from a developer (working for a company) who wished to use it as a library linked into their own software (rather than a standalone executable), but who was unable to convince his boss that LGPL would be acceptable. IIRC, the software was going to end up in some gadget to be sold (and therefore the code was being distributed, indirectly). The commercial licence I provided for him was intended to uphold the spirit of the LGPL, without going as far as BSD in laxity. So, if you simply want to use cpphs in a distributed product (but not modify it), it is very easy. The moment you want to distribute a modified version, you must abide by the LGPL, which to me essentially means that you contribute back your changes to the community. Regards, Malcolm

Malcolm Wallace

9:25 a.m.

On 13 Dec 2012, at 10:41, Petr P wrote:

...

In particular, we can have a BSD package that depends on a LGPL package, and this is fine for FOSS developers. But for a commercial developer, this can be a serious issue that is not apparent until one examines *every* transitive dependency.

This might a good time to remind everyone that every single program compiled by a standard GHC is linked against an LGPL library (the Gnu multi-precision integer library) - unless you take care first to build your own copy of the compiler against the integer-simple package instead of integer-gmp. As far as I know, there are no ready-packaged binary installers for GHC that avoid this LGPL'd dependency. http://hackage.haskell.org/trac/ghc/wiki/ReplacingGMPNotes Just saying. Regards, Malcolm

Michael Snoyman

11:54 a.m.

On Sat, Dec 15, 2012 at 4:25 PM, Malcolm Wallace wrote:

...

On 13 Dec 2012, at 10:41, Petr P wrote:

...
In particular, we can have a BSD package that depends on a LGPL package, and this is fine for FOSS developers. But for a commercial developer, this can be a serious issue that is not apparent until one examines *every* transitive dependency.

This might a good time to remind everyone that every single program compiled by a standard GHC is linked against an LGPL library (the Gnu multi-precision integer library) - unless you take care first to build your own copy of the compiler against the integer-simple package instead of integer-gmp. As far as I know, there are no ready-packaged binary installers for GHC that avoid this LGPL'd dependency.

http://hackage.haskell.org/trac/ghc/wiki/ReplacingGMPNotes

Just saying.

The difference between a library being (L)GPLed and this GMP issue is that, in the latter case, we have an escape route. I know of at least two companies which are actively considering switching entirely to simple-integer because of this issue. If a widely used package (e.g., cpphs) is not available under a permissive license, there's not such escape route available to users. (And note that I'm not actually *happy* about the GMP situation, but at least we have a possible solution.) I would strongly recommend reconsidering the licensing decision of cpphs. Even if the LICENSE-commercial is sufficient for non-source releases of software to be protected[1], it introduces a very high overhead for companies to need to analyze a brand new license. Many companies have already decided BSD3, MIT, and a number of other licenses are acceptable. It could be very difficult to explain to a company, "Yes, we use this software which says it's LGPL, but it has this special extra license which, if I'm reading it correctly, means you can't be sued, but since the author of the package wrote it himself, I can't really guarantee what its meaning would be in a court of law." Looking at the list of reverse dependencies[2], I see some pretty heavy hitters. Via haskell-src-exts[3] we end up with 75 more reverse dependencies. I'd also like to point out that cpphs is the only non-permissively-licensed dependency for a large number of packages. I can give you more detailed information about my commercial experience privately. But I can tell you that, in the currently situation, I have created projects for clients for which Fay[4] would not be an option due to the cpphs licensing issue. Michael [1] I'm not sure of that, since IANAL. [2] http://packdeps.haskellers.com/reverse/cpphs [3] http://packdeps.haskellers.com/reverse/haskell-src-exts [4] http://packdeps.haskellers.com/licenses/fay

Malcolm Wallace

6:10 p.m.

On 15 Dec 2012, at 16:54, Michael Snoyman wrote:

...

I would strongly recommend reconsidering the licensing decision of cpphs. Even if the LICENSE-commercial is sufficient for non-source releases of software to be protected[1], it introduces a very high overhead for companies to need to analyze a brand new license. Many companies have already decided BSD3, MIT, and a number of other licenses are acceptable.

Well, if a company is concerned enough to make an internal policy on open source licences at all, one might hope that they would perform due diligence on them too. For instance, the FSF have lawyers, and have done enough legal work to be able to classify 48 licences as both "free" and GPL-compatible, a further 39 licences as "free" but non-GPL-compatible, and 27 open source licences that are neither "free" nor GPL-compatible. This kind of understanding is what lawyers are supposed to be for. Making them look at another (short) licence is not really a big deal, especially when it closely resembles BSD, which they have already supposedly decided is good. My suspicion, though, is that most of the companies who even think about this question are small, do not have their own lawyers, and are making policy on the hoof, motivated purely by fear. I also suspect that they do not even have the resources to read the licence for each library in its entirety, to determine whether it is in fact BSD3 or MIT, as claimed, or whether someone has subtly altered it. Also, I think I could be pretty confident that there are many shipping products that contain genuine BSD-licensed code, but which do not comply with its terms.

...

It could be very difficult to explain to a company, "Yes, we use this software which says it's LGPL, but it has this special extra license which, if I'm reading it correctly, means you can't be sued, but since the author of the package wrote it himself, I can't really guarantee what its meaning would be in a court of law."

Like I say, if someone claims the software to be BSD-licensed, someone has to read the licence text itself anyway, to determine whether the claim is true. Pretty much every copy of the BSD licence text differs anyway, at least by the insertion of the authors' names in various places, and sometimes there are varying numbers of clauses.

...

Looking at the list of reverse dependencies[2], I see some pretty heavy hitters. Via haskell-src-exts[3] we end up with 75 more reverse dependencies. I'd also like to point out that cpphs is the only non-permissively-licensed dependency for a large number of packages.

I'm glad that cpphs is widely used. I'm also glad that it remains free, and I disagree with you that its dual-licence model is non-permissive. I would like to encourage more Haskell developers to adopt free licensing. Don't be bullied by BSD evangelists! BSD is not the only way to a good citizen of the community! Your libraries can be delivered to clients as products, without you having to give up all rights in them! It's not like I'm saying to companies "if you make money out of my code, you have to pay me a fee". All I'm saying, to everyone, is "if you notice a bug in my code and fix it, tell me". This is fully compatible with allowing people to release my code to their clients inside products.

...

I can give you more detailed information about my commercial experience privately. But I can tell you that, in the currently situation, I have created projects for clients for which Fay[4] would not be an option due to the cpphs licensing issue.

If you are complaining about the crazy policies that many companies adopt about the use of free software within their business, then I have plenty of sympathy for that too. I know of one which has a policy of "no use of open source code whatsoever", but runs thousands of linux servers. :-) Also, many companies with large numbers of software engineers on staff apparently prefer to buy crappy commercial products and pay handsomely for non-existent support, instead of running high-quality open-source software with neither initial nor ongoing costs, and where bugfixes are often available the same day as you report the bug. But hey ho. Corporate policy is usually made by people with neither technical nor legal expertise. As regards cpphs, if you don't want to use it because of its licences, that is your choice. You can always use some other implementation of the C pre-processor if you wish. GHC has always refused to distribute cpphs, on the basis of its GPL licence, and instead chose to distribute GNU's gcc on Windows. :-) (I hope you see the irony!) Regards, Malcolm

Brandon Allbery

12:39 p.m.

On Sat, Dec 15, 2012 at 9:25 AM, Malcolm Wallace wrote:

...

This might a good time to remind everyone that every single program compiled by a standard GHC is linked against an LGPL library (the Gnu multi-precision integer library) - unless you take care first to build your own

This is less relevant though, because gmp is not a Haskell library so linking against it doesn't pull significant chunks of its source code into your program. Indeed, many GHC distributions dynamic-link against it, minimizing the legal concerns; and IIRC the gmp license explicitly allows the use GHC makes of it as long as the gmp symbols are not re-exported as part of its API (which they aren't). -- brandon s allbery kf8nh sine nomine associates allbery.b@gmail.com ballbery@sinenomine.net unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

4540

Age (days ago)

4542

Last active (days ago)

List overview

Download

21 comments

9 participants

participants (9)

Brandon Allbery
Brent Yorgey
Daniel Trstenjak
Felipe Almeida Lessa
Henk-Jan van Tuyl
Malcolm Wallace
Michael Snoyman
Petr P
Vincent Hanquez