Is Apple Now Blocking Contributions To GCC? | Hacker News Reader

Is Apple Now Blocking Contributions To GCC?

phoronix.com

36 pointsspcmnspff30 comments

Apple isn't blocking contributions, they just don't care about them anymore since they're going with LLVM/Clang. And since they don't care they don't have the legal framework in place anymore to do copyright assignments.

The FSF could have these changes if they wanted if they didn't require copyright assignment.

Personally I don't contribute to FSF projects because copyright assignments are a PITA. I can see how a company that doesn't need to contribute to GCC anymore just doesn't care enough to initiate the requisite legal work.

What part of the copyright assignment process do you find bothersome?

I've contributed many thousands of lines of code to free software projects, but I've never had to sign some waiver, walk down to the post office and send an inter-continental letter in the mail to go along with a patch I wrote.

I'd had have to do that if I submitted a non-trivial change to the FSF, just finding an alternative program that I can use and fix is easier, so I've done that instead of sending patches to them.

Compare a normal workflow, like submitting a bug to Git:

* Commit your bugfix

* 'git send-email --to=git@vger.kernel.org -1'

With the GNU model:

* Commit your bugfix / make a patch

* Send your patch to the maintainer of the project

* The maintainer has to look in a central file if you already assigned copyright to the FSF

* If not, they send you an email 'would you be willing to do this'

* You reply, "sure, send me the form"

* They send you this: https://gforge.inria.fr/scm/viewvc.php/misc/www/request-disc...

* You fill in the form, and thus hand over your snail mail address to the maintainer and the FSF

* They send you another form (snail mail!)

* You fill it in and send it back (snail mail!)

* They receive your form and add it to the list

* Your bugfix is now allowed to go into the GNU repository

Yes, but, after the first shot:

* Commit your bugfix / make a patch

* Send your patch to the maintainer of the project

* The maintainer has to look in a central file if you already assigned copyright to the FSF. Since you did, you don't have to do anything.

I'm confused though: Apple already have assigned copyright. Yet, they say they will stop doing it, as if it wasn't automatic, even when done once. Either they assign copyright in an unusual way, or they actively stopped it. Somebody knows which it could be?

Just because Apple has a copyright assignment on file (i.e. saying that they are able to assign copyright to the FSF) does not mean that every piece of code Apple writes--even if it's connected to GCC--is assigned to the FSF.

So Apple can write GCC patches, Apple can release GCC patches, but it does not follow that those patches are assigned. I guess that means they've actively stopped.

Sort of the relationship of Android with the Linux kernel, for me this is the same as publishing a GPLed patch to OpenSSH or something like that, this is immense disrespect towards the developer community of a project, which does have its culture.

FSF has done this as well by changing the license of their software from GPL2/LGPL2 to GPL3/LGPL3 and turning the GameSWF (released under public domain) into the GPLed Gnash.

Are you saying that forks are rude? But they're the point of both Free Software and Open Source!

If I recall correctly, the standard FSF copyright assignment form lets you choose between assigning copyright on a particular set of changes, or for all (present and) future changes.

You forgot the bit where if you work for a company you need to get approval from 18levels of management and corporate lawyers on 3 continents who have never heard of free software - and you are on the receiving end of the political flak.

Or if you work for a US company and are writing FOSS on your own time - you still risk getting fired because one of the managers or lawyers (who have never heard of FOSS) think it breaks your employment contract and it's a cheap way of reducing head count.

Has this ever really happened? This sounds more like apocryphal legends meant to scare young developers about that big, bad manager in the closet.

Being fired - don't know.

But if you are a minion at, say MSFT, imagine how many layers of management a sign-off would go up before somebody said - yes you can assign copyright to that little utility to Gnu.

I personally know of two people from one company that were fired for contributing a patch to a popular open sourced IDE that the company used for development, of which the bug hampered the development for the company. It was a strongarmed political move for management to retain control over the personal actions of their employees.

This is all anecdotal anyways. I doubt it happens all the time, but these two were forced to sign an NDA on their way out. It makes me wonder how many times it does happen where no one is allowed to talk about it.

Do you know if, in this case, they worked on the patch on company time, or on their own time?

Sending physical mail seems really odd to me by now. I haven't turned on autopayment for half of my student loans because they require me to fill out a form and mail them in. It's such a hassle; I have to find time during the day to go purchase a single envelope and a single stamp, fill out the form, and then go drop it off. So much effort.

I'm sadden to say that I'm in the same boat as you. Too lazy to go to the post office, I just login to the website and do a manual payment for my loans.

I think, I'm going to just do it now that I realize how pathetic it is.

Not to derail this too much, but I still pay most all of my bills by check and snail-mail envelope. Until some companies stop charging a "convenience fee" that is more than the cost of a stamp and 1 minute of my time per month, of which I'm probably severely underestimating its worth, I'll continue to do so.

The headline is misleading. Chris Lattner's post actually just points out that FSF's own copyright-assignment policy may be a blocker:

"...In practice, since the FSF cares about copyright assignment, this probably means that..."

http://www.gnu.org/licenses/why-assign.html

direct links to thread: http://gcc.gnu.org/ml/gcc/2010-09/msg00109.html

"Apple does not have an internal process to assign code to the FSF anymore. I would focus on the code that is already assigned to the FSF." http://gcc.gnu.org/ml/gcc/2010-09/msg00132.html

Thread in context: http://gcc.gnu.org/ml/gcc/2010-09/threads.html#00109

So let's say a man asks for some food, and Apple says "well, we've got some beef jerky here, but I know you're a strict vegan so you'd probably be more interested in the tofu we set aside for you yesterday", and the headline is "Apple blocks starving man from food".

Er, no. Apple built off the existing code base. This isn't charity, they chose to use free software in the first place.

And actually, when they made that choice, it was Apple that was starving. But I'll let someone else draw up the parable. ;-)

Actually actually, it was NeXT who made that choice.

And the license itself does not require copyright assignment. Rather, the FSF does.

... and then FSF changed the license. GCC is under GPLv3 now, no?

Although the underlying story is interesting, I'm not going to vote up this particular article due to the horribly misleading title.

I am under the impression that phoronix usually does poor "journalism". I still need to understand why file systems get tested using compression and compilation benchmarks.

Phoronix. The tabloid of free software geek news website. Their best known speciality: sensationalist articles based on ridiculous benchmarks.

Seems to me that the original article writer is flamebaiting and being deliberately inflammatory with his assumptions. I'd say the main issue here is that Apple does not agree with the GPL license family (which we all know since long; they've voiced this before), and that they don't really care for GCC anymore now that they have set aim for Clang-LLVM.

I think your opinion is too strong, OS X still ships with many GPL licensed software, bash, smbclient, nmblookup and I'm pretty sure that OS X server includes the GPL3 licensed Samba, OS X still comes with gcc, and I believe it will for the next two releases at least. CUPS, the printer daemon that Apple bought some years ago is still a GPLed software and it appears that Apple has done nothing to change that.

Things can change and Apple is a great contributor to LLVM, but Apple appear to have no trouble with the GPL, only trouble with gcc, but that is easy to understand.

The current issue with Apple and GPL stems from the App Store. FSF says you can't limit distribution of the code, which is reasonable since they own the license to the code. Apple says we want to limit the distribution of the app, which is reasonable since they own the channel.

Just like I tell my kids, if you can't get along with the neighbor, don't go play in their back yard. Sounds like Apple is taking that advice and people in the free community are offended.

Reply on news.ycombinator.com