handling of pull requests from new contributors

[Philip Race]

On 10/12/20, 11:38 AM, Stuart Marks wrote:
> Hi, I wanted to mention that we're getting PRs from new contributors, 
> and they're more-or-less running into a brick wall. See this comment 
> for example:
>
> https://github.com/openjdk/jdk/pull/115#issuecomment-706100552
>
>> It is insane how difficult it is to open a Pull Request for this 
>> project. I think it is bad for the Java platform, you need to attract 
>> more contributors for the project to keep it alive...
>>
>> I am stuck with the process to open a new issue, the reset link to 
>> create a new account does not work. 
>

I'd like to point out that he must have breezed past this message on the 
JBS login page.

 >Everyone with OpenJDK Author status or above has a JBS account which 
may be
 > used to create and edit bugs. Those without accounts can view bugs 
anonymously.


I don't know exactly where he found the reset link but I think that we 
may need
to put a message more prominently and more clearly on the JBS login pages
since I think it is more useful at the point of use than in the guide.

Something like:
Note : You cannot create a JBS account for yourself. If you want to 
create one
you first need to have contributed fixes and then one will be created 
for you.
This means that if you want to create the bug you want to fix initially 
you need to use
the anonymous bug creation URL : bugreport.java.com
For more info please refer to <contributing>.


-phil.

> Naturally, there are a bunch of issues occurring in parallel here, but 
> I think one goal would be to have sufficient material in the Guide so 
> that we can simply point new contributors to it.
>
> (Also, when the Guide is in good-enough shape, the JDK README file 
> should have a pointer to it.)
>
> One issue is the OCA. The bots seemed to have taken care of that 
> reasonably well, as the submitter did go through the OCA submission 
> process and the bots removed the OCA label.
>
> A second issue, one that's currently blocking this PR, is that it's 
> been failed by jcheck because the commit message needs to refer to an 
> issue. There's no issue, so the submitter tried to create an issue. 
> But one needs an account to create an issue, so the submitter tried to 
> create an account. They found the right issue tracker (JBS) but 
> there's no way to create an account from there, and there's nothing 
> that points them to information about needing an Author role to create 
> a JBS account, and what it takes to get the Author role.
>
> But really, what's necessary is that new contributors need to find a 
> sponsor for fixes before they can get the Author role. Frankly, I 
> don't know how to do that. I guess, ask on the mailing list?
>
> Note that when a PR is stuck in this state, no mail is sent to any of 
> the mailing lists. I don't think that's wrong; doing so would result 
> in a lot more noise on the mailing lists. But it also means that PRs 
> can get stuck in this state and nobody is notified. For now, I've 
> asked my team to look for stale PRs. There aren't very many yet, but a 
> handful are building up and some are over a month old already.
>
> Here's a query URL for finding PRs that have failed their checks:
>
> https://github.com/openjdk/jdk/pulls?q=is%3Apr+is%3Aopen+status%3Afailure
>
> Anyway, I'm not sure how we're tracking stuff that needs to go into 
> the Guide. (Notes in the README?) Wherever it is, these sorts of 
> issues need to be added there.
>
> s'marks