[gentoo-pms] License for code snippets included in PMS

[gentoo-pms] License for code snippets included in PMS

[Ulrich Mueller]

The PMS document is relased under the CCPL-Attribution-ShareAlike-3.0
license. This may raise some issues for package managers that want to
use some of the included code snippets. Portage and paludis are
licensed under GPL-2, and pkgcore seems to be heading towards BSD.

Currently, most of the code snippets are short (under 10 lines) and
are probably not copyrightable. But this might change in future.

Could we put all code snippets under some unrestrictive license
please? Like CC0-1.0-Universal?

Ulrich

Re: [gentoo-pms] License for code snippets included in PMS

[Ulrich Mueller]

>>>>> On Sun, 20 Feb 2011, I wrote:

> Could we put all code snippets under some unrestrictive license
> please? Like CC0-1.0-Universal?

"Qui tacet consentire videtur."

Patch is included below, please review.

Ulrich


diff --git a/credits.tex b/credits.tex
index 459fc1b..931d0b4 100644
--- a/credits.tex
+++ b/credits.tex
@@ -19,6 +19,9 @@ inclusion.
 This document is released under the Creative Commons Attribution-Share Alike 3.0 Licence. The full
 text of this licence can be found at \url{http://creativecommons.org/licenses/by-sa/3.0/}.
 
+All code snippets included with this document are dedicated to the public domain by Creative
+Commons CC0 1.0 Universal, see \url{http://creativecommons.org/publicdomain/zero/1.0/legalcode}.
+
 \section*{Reporting Issues}
 
 Issues (inaccuracies, wording problems, omissions etc.)\ in this document should be reported via

Re: [gentoo-pms] License for code snippets included in PMS

[Petteri Räty]

[-- Attachment #1: Type: text/plain, Size: 458 bytes --]

On 02/24/2011 09:49 AM, Ulrich Mueller wrote:
>>>>>> On Sun, 20 Feb 2011, I wrote:
> 
>> Could we put all code snippets under some unrestrictive license
>> please? Like CC0-1.0-Universal?
> 
> "Qui tacet consentire videtur."
> 
> Patch is included below, please review.
> 
> Ulrich
> 

To be on the safe side I think we should get permission from everyone
who has worked on the code snippets before applying this patch.

Regards,
Petteri


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 900 bytes --]

Re: [gentoo-pms] License for code snippets included in PMS

[Ulrich Mueller]

>>>>> On Sun, 13 Mar 2011, Petteri Räty wrote:

>>> Could we put all code snippets under some unrestrictive license
>>> please? Like CC0-1.0-Universal?

> To be on the safe side I think we should get permission from
> everyone who has worked on the code snippets before applying this
> patch.

Here's the complete list from git history:

   ebuild-functions.tex
      default src_unpack (5 lines):
      Stephen Bennett, Fernando J. Pereda

      default src_configure (5 lines):
      Ciaran McCreesh

      default src_compile for EAPI 0 (8 lines):
      Stephen Bennett, Ciaran McCreesh, Fernando J. Pereda

      default src_compile for EAPI 1 (8 lines):
      Stephen Bennett, Ciaran McCreesh, Fernando J. Pereda

      default src_compile for EAPI 2 (5 lines):
      Ciaran McCreesh

      default src_install for EAPI 4 (16 lines):
      Ciaran McCreesh

   pkg-mgr-commands.tex
      einstall command (7 lines):
      Stephen Bennett, Ulrich Müller

   eclasses.tex
      EXPORT_FUNCTIONS example (8 lines):
      Stephen Bennett, Ciaran McCreesh

   env-saving.listing (21 lines):
      Ciaran McCreesh

Some of the code is of course taken from elsewhere, like ebuild.sh
(some of the default phase functions) or skel.ebuild (einstall).
And probably anything below 10 nonempty lines is not copyrightable
anyway (at least that's the rule the FSF applies).

As far as my changes are concerned, I release them under
CC0-1.0-Universal.

Ulrich

Re: [gentoo-pms] License for code snippets included in PMS

[Brian Harring]

[-- Attachment #1.1: Type: text/plain, Size: 822 bytes --]

On Sun, Mar 13, 2011 at 03:10:55PM +0100, Ulrich Mueller wrote:
> >>>>> On Sun, 13 Mar 2011, Petteri Räty wrote:
> 
> >>> Could we put all code snippets under some unrestrictive license
> >>> please? Like CC0-1.0-Universal?
> 
> > To be on the safe side I think we should get permission from
> > everyone who has worked on the code snippets before applying this
> > patch.
> 
> Here's the complete list from git history:
>    env-saving.listing (21 lines):
>       Ciaran McCreesh

This section wasn't accurate, as such I threw it out- attached is a 
brain dump of exactly what characteristics can be relied on for
env saving, and should be reasonably complete.

Could stand to have some comments added though.

As for license, CC0-1.0-universal/public domain/whatever floats your 
boat.

~harring

[-- Attachment #1.2: 0001-correct-env-saving.listing-to-be-accurate-complete.patch --]
[-- Type: text/plain, Size: 3805 bytes --]

From a02ba65376e463818aa3d7abc2f845bd5ca745ff Mon Sep 17 00:00:00 2001
From: Brian Harring <ferringb@gmail.com>
Date: Sun, 13 Mar 2011 09:09:43 -0700
Subject: [PATCH] correct env-saving.listing to be accurate/complete

Version that was in place wasn't particularly accurate nor complete.
As such, and to cover copyright/licensing concerns, a from-scratch
version was written covering array behaviours, functions, unset behaviour,
and exports.

Signed-off-by: Brian Harring <ferringb@gmail.com>
---
 env-saving.listing |   84 +++++++++++++++++++++++++++++++++++++++------------
 1 files changed, 64 insertions(+), 20 deletions(-)

diff --git a/env-saving.listing b/env-saving.listing
index 1650aa7..4391553 100644
--- a/env-saving.listing
+++ b/env-saving.listing
@@ -1,27 +1,71 @@
-GLOBAL_VARIABLE="a"
-
-src_compile()
-{
-    GLOBAL_VARIABLE="b"
-    DEFAULT_VARIABLE="c"
-    export EXPORTED_VARIABLE="d"
-    local LOCAL_VARIABLE="e"
+GLOBAL_SCOPE_VARIABLE="initial"
+VAR_TO_WIPE=initial
+declare -a ARRAY_VARIABLE=( a b )
+
+my_function() {
+	echo "exists"
+}
+
+src_compile() {
+	[[ ${GLOBAL_SCOPE_VARIABLE} == initial ]] || \
+		die "GLOBAL_SCOPE_VARIABLE must be preserved"
+
+	export GLOBAL_SCOPE_VARIABLE=set_twice
+
+	[[ ${VAR_TO_WIPE} == initial ]] || \
+		die "VAR_TO_WIPE must be preserved"
+
+	# now we wipe the var, to ensure that the env saving doesn't preserve it.
+	unset VAR_TO_WIPE
+
+	local LOCAL_SCOPE_VARIABLE=unseen
+
+	DEFAULT_SCOPE_VARIABLE=set
+
+	{ [[ ${#ARRAY_VARIABLE[@]} == 2 ]] && \
+		[[ ${ARRAY_VARIABLE[0]} == a ]] &&  \
+		[[ ${ARRAY_VARIABLE[1]} == b ]]; \
+	} || \
+		die "ARRAY_VARIABLE contents must be preserved exactly, and accessible"
+	ARRAY_VARIABLE[2]=c
+
+	[[ $(my_function 2> /dev/null) == exists ]] || \
+		die "'my_function' must be preserved"
 }
 
-src_install(){
-    [[ ${GLOBAL_VARIABLE} == "a" ]] \
-        || [[ ${GLOBAL_VARIABLE} == "b" ]] \
-        || die "broken env saving for globals"
 
-    [[ ${DEFAULT_VARIABLE} == "c" ]] \
-        || die "broken env saving for default"
+src_install() {
+	[[ ${GLOBAL_SCOPE_VARIABLE} == set_twice ]] || \
+		die "GLOBAL_SCOPE_VARIABLE modifications in src_compile must be preserved"
+
+	[[ -n $(declare -xp GLOBAL_SCOPE_VARIABLE 2> /dev/null) ]] || \
+		die "GLOBAL_SCOPE_VARIABLE must be exported due to src_compile exporting it"
+
+	[[ ${LOCAL_SCOPE_VARIABLE-unset} == unset ]] || \
+		die "a local defined variable should be impossible to propagate"
+
+	[[ ${VAR_TO_WIPE-unset} == unset ]] || \
+		die "VAR_TO_WIPE was unset during src_compile; this wiping must be preserved"
+
+	[[ ${DEFAULT_SCOPE_VARIABLE} == set ]] || \
+		die "DEFAULT_SCOPE_VARIABLE must be preserved"
+
+	[[ -n $(declare -xp DEFAULT_SCOPE_VARIABLE 2> /dev/null) ]] || \
+		die "DEFAULT_SCOPE_VARIABLE wasn't exported; environment saving must be preserved that fact."
+
+	[[ ${ARRAY_VARIABLE-unset} != unset ]] || \
+		die "ARRAY_VARIABLE must be preserved"
 
-    [[ ${EXPORTED_VARIABLE} == "d" ]] \
-        || die "broken env saving for exported"
+	[[ -n $(declare -ap ARRAY_VARIABLE 2> /dev/null) ]] || \
+		die "ARRAY_VARIABLE created during src_compile, must have it's array type preserved"
 
-    [[ $(printenv EXPORTED_VARIABLE ) == "d" ]] \
-        || die "broken env saving for exported"
+	{ [[ ${#ARRAY_VARIABLE[@]} == 3 ]] && \
+		[[ ${ARRAY_VARIABLE[0]} == a ]] &&  \
+		[[ ${ARRAY_VARIABLE[1]} == b ]]; \
+		[[ ${ARRAY_VARIABLE[2]} == c ]]; \
+	} || \
+		die "ARRAY_VARIABLE contents must be preserved exactly"
 
-    [[ -z ${LOCAL_VARIABLE} ]] \
-        || die "broken env saving for locals"
+	[[ $(my_function 2> /dev/null) == exists ]] || \
+		die "'my_function' must be preserved"
 }
-- 
1.7.4


[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-pms] License for code snippets included in PMS

[David Leverton]

On 13 March 2011 16:40, Brian Harring <ferringb@gmail.com> wrote:
> This section wasn't accurate, as such I threw it out- attached is a
> brain dump of exactly what characteristics can be relied on for
> env saving, and should be reasonably complete.
>
> Could stand to have some comments added though.

As discussed on IRC, the use of "declare" here is problematic, because
inside a function it acts like "local", even if said function is part
of the PM that sources the ebuild.  This would put a pretty nasty
constraint on the PM implementation, and wouldn't work in eclasses
anyway because of the "inherit" function.  Arrays can still be used
simply by not using "declare" and just assigning them, which doesn't
have the implicit local attached, so all that's needed is to change
"declare -a ARRAY_VARIABLE=( a b )" in the patch to simply
"ARRAY_VARIABLE=( a b )".

(See also http://bugs.gentoo.org/show_bug.cgi?id=309369, in which this
caused problems with loading environment.bz2 files.)

As an aside, bash 4.2 has "declare -g" to override the implicit local,
but it'll be a while before we can start using that in ebuilds.

Re: [gentoo-pms] License for code snippets included in PMS

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 624 bytes --]

On Sun, 13 Mar 2011 15:10:55 +0100
Ulrich Mueller <ulm@gentoo.org> wrote:
> Here's the complete list from git history:
>       Ciaran McCreesh

For anything I wrote from scratch, I'm fine with any licence that
allows package manager authors to relicence it under whatever it is
they use themselves. Although, because of

> And probably anything below 10 nonempty lines is not copyrightable
> anyway (at least that's the rule the FSF applies).

I remain unconvinced that any of this is relevant, and by sticking an
explicit licence on the code we're probably just making things worse...

-- 
Ciaran McCreesh

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-pms] License for code snippets included in PMS

[Ulrich Mueller]

>>>>> On Sun, 13 Mar 2011, Ciaran McCreesh wrote:

> For anything I wrote from scratch, I'm fine with any licence that
> allows package manager authors to relicence it under whatever it is
> they use themselves.

Thanks.

> I remain unconvinced that any of this is relevant, and by sticking
> an explicit licence on the code we're probably just making things
> worse...

Well, CC0-1.0-Universal isn't really a license, it's rather a
copyright waiver. Comparable to WTFPL-2, but with a wording that may
be more appropriate in some situations. ;)

Ulrich