Open Bug 541984 Opened 15 years ago Updated 1 month ago

Test suite shouldn't use images that are not licensed under a free license

Categories

(Testing :: General, defect, P3)

defect

Tracking

(Not tracked)

People

(Reporter: glandium, Unassigned)

References

Details

Either from mozilla itself (the Firefox icon isn't MPL/GPL/LGPL):
http://mxr.mozilla.org/mozilla-central/source/toolkit/components/places/tests/unit/
http://mxr.mozilla.org/mozilla-central/source/modules/libpr0n/test/unit/

Or from really doubtful sources:
http://mxr.mozilla.org/mozilla-central/source/modules/libpr0n/test/reftest/pngsuite-transparency/
http://mxr.mozilla.org/mozilla-central/source/layout/html/tests/block/bugs/top_middle2.jpg
http://mxr.mozilla.org/mozilla-central/source/layout/html/tests/block/bugs/20020515_60x60_scooby.jpg
http://mxr.mozilla.org/mozilla-central/source/dom/tests/mochitest/dom-level2-html/files/w3c_main.png
(There are also numerous pictures with doubtful origins)

I haven't taken a look at the text files in the test suite, but I have serious doubts about their content, considering the images...
gerv, can you look into this?
Assignee: nobody → gerv
The Firefox icon files _are_ tri-licensed. (Although there are also trademark restrictions on the image, and trademarked logos are probably not the _best_ choice for test suite material.)

As for the others, if these specific files are necessary to reproduce the bug in question, we could make a good argument for fair use.

I agree it's not ideal, but I don't see it getting to the top of anyone's priority list any time soon.

Gerv
(In reply to comment #2)
> The Firefox icon files _are_ tri-licensed.

That is _not_ what
1. the fact that the branding is in other-licenses
2. the content of other-licenses/branding/firefox/LICENSE
3. the various discussions during the dispute between Debian and Mozilla
say or have said in the past.

If they were tri-licensed, there would never have been the need to use the unbranded logo in Debian Firefox builds in the first place, which led, in the end, in changing the name, since Mozilla didn't want the Firefox name without the associated logo.
> As for the others, if these specific files are necessary to reproduce the bug
> in question, we could make a good argument for fair use.

I really doubt they are strictly necessary. As for fair use, that doesn't change the fact that they are not licenseable under the terms they are distributed under.
We tri-licensed the Firefox logo files when we moved to using the MPL as our EULA about a year ago.

> 2. the content of other-licenses/branding/firefox/LICENSE

That says they are under the tri-license.
http://mxr.mozilla.org/mozilla-central/source/other-licenses/branding/firefox/LICENSE

Gerv
(In reply to comment #4)
> I really doubt they are strictly necessary. As for fair use, that doesn't
> change the fact that they are not licenseable under the terms they are
> distributed under.

Speaking of which, the icons in the search box are not tri-licensed either.
(In reply to comment #5)
> That says they are under the tri-license.
> http://mxr.mozilla.org/mozilla-central/source/other-licenses/branding/firefox/LICENSE

That's really not how I read it. (And it hasn't changed since its inclusion in mercurial, in 2007)
 > That's really not how I read it. (And it hasn't changed since its inclusion in
> mercurial, in 2007)

Oh %$&*. I forgot to check that change in. And the tree's currently on fire. I'll add it to my list of things to do.

The new text is:

"These files are under the tri-license, as below. However, please note that you are not granted any trademark rights or licenses to the trademarks of the
Mozilla Foundation or any party, including without limitation the
Firefox name or logo."
(Tri-license header)

Who says the icons in the search box aren't tri-licensed? They may have trademark restrictions, but this is where a "we can't use anything with trademark restrictions" policy would get really silly.

Gerv
(In reply to comment #8)
>  > That's really not how I read it. (And it hasn't changed since its inclusion
> in
> > mercurial, in 2007)
> 
> Oh %$&*. I forgot to check that change in. And the tree's currently on fire.
> I'll add it to my list of things to do.
> 
> The new text is:
> 
> "These files are under the tri-license, as below. However, please note that you
> are not granted any trademark rights or licenses to the trademarks of the
> Mozilla Foundation or any party, including without limitation the
> Firefox name or logo."
> (Tri-license header)

Waw. That changes everything.
 
> Who says the icons in the search box aren't tri-licensed? They may have
> trademark restrictions, but this is where a "we can't use anything with
> trademark restrictions" policy would get really silly.

Somehow, I find it hard to believe that the google, yahoo, amazon, answers.com, creative commons, wikipedia and ebay logos/favicons are tri-licensed.

The fact there is no licensing text in these searchplugin files doesn't help.
Depends on: 541761
FWIW, the Wikipedia icon is copyrighted by the Wikimedia Foundation, all rights reserved: <http://en.wikipedia.org/wiki/File:Wiki.png>
Gerv, could you reply to comment 4 and comment 9 ? Thanks.
My view on the questions of the images in the test suites is:

(0) we should be writing our own minimal, minimally-copyrighted testcases whereever that is possible (both as legal and as QA best practice) 
(1) fair use probably protects us where we can't do 0 and 
(2) in practice no one will care.

My view on the question of the search images is that in the short term, it would be good to have a LICENSE file wherever they live that notes that those icons may have different licenses. In the medium term, the way forward is to contact the companies or organizations in question to ask them to release the 16x16 versions of their logos as free software, without prejudice to their trademarks. Either or both of these issues should have separate bugs filed about them.

I would warn, however, that fixing this is not at the top of anyone's priority list here at Mozilla, as far as I know. If this is a task you'd like to take on, that would be great :-)

Gerv
(In reply to comment #12)
> I would warn, however, that fixing this is not at the top of anyone's priority
> list here at Mozilla, as far as I know. If this is a task you'd like to take
> on, that would be great :-)

I might do something about the reftest images, though I quite don't know yet, but for the searchplugins icons, I think I'll just go the simple route and replace the data: urls with plain urls. It's not like it's a big problem that these don't show up properly when there is no net connectivity... since the searchbox won't be working properly anyway.

Mike
FWIW, a bunch of reftest images come from http://www.libpng.org/pub/png/pngsuite.html, which in turn come from http://www.schaik.com/pngsuite/pngsuite.html. The copyright and license terms for these are definitely doubtful.
Mike: why "definitely doubtful"? Or do you mean "unknown"?

Gerv
(In reply to comment #15)
> Mike: why "definitely doubtful"? Or do you mean "unknown"?

Let's put it this way: The copyright is doubtful (i really doubt he created the NeXT icons, for instance) and the license is unknown.
> (i really doubt he created the NeXT icons, for instance)

And apparently, the set of icons from the png suite have been created on a NeXTStation, so my guess is that the NeXT icons are really a conversion of the original NeXT icons from the NeXT operating system.
The w3c dom tests are apparently under a BSD 3 clause license.

So, apart from the NeXT icons, here are the most obvious cases of copyright and license doubts:
layout/html/tests/table/images/*
layout/html/tests/images/polarBears.jpg
layout/html/tests/SlideShow/*
layout/doc/raptor.jpg
layout/html/tests/printer/images/raptor.jpg
layout/html/tests/printer/general/raptor.jpg
layout/html/tests/table/images/raptor.jpg
layout/html/tests/FrameSetNav/raptor.jpg
content/xml/tests/books/*
testing/performance/talos/page_load_test/svg/images/kyoto_2.jpg
AFAICS (but I can be wrong), the tests in layout/html and content/xml are not used, the raptor.jpg image in layout/doc could be replaced by any image with the same size, and I couldn't find the mercurial repository containing testing/performance/talos.
In case it's not clear: anyone who wants to make a patch to replace any possibly non-free test suite images with definitely-free images where the test still pass, is welcome. They might want to make a new bug for each group, though, as the reviewer will differ depending on which tests it is.

The search engine favicons are a different problem, and have to be tackled another way.

Gerv
I'm not going to be resolving this. See comment 20 for a way forward if someone wants to.

Gerv
Assignee: gerv → general
Severity: normal → S3

Moving here for further triage.

Assignee: general → nobody
Severity: S3 → --
Product: Core → Testing
Version: Trunk → unspecified

The severity field is not set for this bug.
:jmaher, could you have a look please?

For more information, please visit BugBot documentation.

Flags: needinfo?(jmaher)
Severity: -- → S3
Flags: needinfo?(jmaher)
Priority: -- → P3
You need to log in before you can comment on or make changes to this bug.