Thanks for all the replies by the way. Definitely helping me understand the other viewpoint (and try to poke holes :) ).
On 11/6/06, Mike Kienenberger <[EMAIL PROTECTED]> wrote:
On 11/6/06, Henri Yandell <[EMAIL PROTECTED]> wrote:
<snip...>
> How do you know you can use the code to identify and fix the bug? > > And more confusingly - how do you write a unit test for that bug > without taking too much copyright from the example code? > > Legal-wise (and my understanding of this stuff is always iffy) - I > thought that if something wasn't licensed/publically announced in the > public domain, that it was 'All Rights Reserved'. So anything that > doesn't check the 'code-grant' checkbox seems completely unusable. I'm certainly no expert on this, but I'd say if someone opens a bug report and attaches code to it saying that this is an example reproducing the bug, then they've given you the rights to use it to identify and fix the bug. In all cases we're assuming the person posting the code has the right to grant usage since any other case is invalid no matter how they check the box. Writing a unit tests using their code as a template is separate issue which I won't try to address (not qualified).
Yeah, same here. It underlies a more general one though - if something is not being granted to us, we don't know what the terms are under which it is being attached.
Code grant serves another purpose -- it's for explicitly saying "I want you to include my code in the project code, and I give you full rights to do so". > From a more ASF-selfish point of view - what are we losing if we don't > allow people to put stuff in without granting us the right to use > that? Are we going to see lots of people not attaching patches/test > code? It's going to depend on the people, but in general, yes. It's hard to get people to attach test code. It's harder still if you expect them to independently write a test case for which ASF will get a code-grant. If you're only interested in fixing problems for people who have time to make code-grant test cases instead of snippets of work they are probably doing for-hire, you're going to lose a lot of potential contributions.
Is it normal to have permission to publish snippets of for-hire work? I would have thought that posting such things would break most people's contracts/NDAs etc. Not allowing a grey-line set of attachments is a bonus to people in that position I think. It seems that mostly people would either attach/report-issues that we can use, or they shouldn't have reported/attached in the first place for whatever internal legal reason. A policy which allows things to be published publically, but not be used in code seems pretty unlikely given the general level of paranoia.
Some projects may find that acceptable, but unless there's a legal reason why it cannot be allowed, why make the process any more difficult?
Because we're making the process more difficult by allowing the option. Apparantly I have to chase up people in Bugzilla to find out which option they would have checked, and I have to start checking which option an attachment was flagged under everytime I look at an attachment in JIRA. We also have JIRAs that have not migrated to the ASF who do not have this turned on yet. So they need to be doing the same as Bugzilla users.
In my experience, it's rare that such a posted example is useful (even for test cases) outside of identifying the bug in the first place unless the reporter already wrote a standalone (and thus grantable) test case. And in those cases, ask the reporter to check the code grant button.
But you think that if they couldn't post that example, that they wouldn't have reported the bug? Hen --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
