[webkit-dev] Is the wxWidgets port maintained?
kevino at theolliviers.com
Mon Feb 11 16:53:09 PST 2013
On Feb 11, 2013, at 2:32 PM, Ryosuke Niwa wrote:
> On Mon, Feb 11, 2013 at 2:13 PM, Brent Fulgham <bfulgham at gmail.com> wrote:
> On Mon, Feb 11, 2013 at 1:51 PM, Ryosuke Niwa <rniwa at webkit.org> wrote:
>> Asking people to ignore wx is basically the same as having the code outside the tree. Having the code outside the tree would make this change effective instead of having one more rule to the project.
> My big concern is not losing history while moving the wx files out of the tree and into a git branch, and also at the same time not losing the ability to merge the rest of trunk safely. Though I haven't investigated the process yet, so I don't know how easy or hard it may be.
> I kindly ask you to let us remove it from trunk in that case.
> Given that we have 400+ contributors, having a port that's only maintained by a single person on trunk imposes a significant maintenance burden on others even if you said that people can break Wx port because not everyone reads this thread and wouldn't necessarily know your intent.
> I would be sad if the WebKit project started to exclude ports simply because they were (largely) maintained by a single person. I do agree that it's not fair to other ports if the primary developer is not working to keep the port building, but life happens and sometimes a lone developer cannot stay on top of the rapidly-changing source base.
> Having these ports in the main tree is a good thing, in my opinion, because it helps avoid fragmentation, and often the refactoring required to enable these ports provides architectural benefits to the project as a whole.
> I'm not opposed to keeping those refactoring in trunk.
> As I said on https://lists.webkit.org/pipermail/webkit-dev/2013-January/023563.html, the problem is the ratio of contribution each port makes to the core code vs. the cost it imposes on the project.
> Having hundreds, if not thousands, of engineers sync Wx code in WebKit trunk whenever they run "svn up" or "git fetch" is a significant cost on its own, and we need a good justification for it.
I know you're speaking more in general here, but just to be clear, over the past 6 months, the data hit that WebKit engineers have taken on account of the wx port is close to 0 bytes. The only time svn up or git fetch has pulled anything related to wx is when other developers attempted to update the wx port based on changes made to common code. As things stand, I'd not only be okay with, but actually prefer, that no other port maintainers do anything to try and fix the wx build. This way the port is not causing anyone else a burden except occasionally when I do land some fixes. As I said in my reply to Brent, I think a "no build bot / no shared maintenance" project rule is not only reasonable, as a practical matter it is required.
Anyway, I'm happy to move it off of trunk if there's a straightforward way to, say, not lose history and such with git and I will try and check into how to make that work more later. I just wanted to point out that I don't think wx has really been having a lot of negative impact on WebKit developers, and if it ever does start having more of an impact, we'll probably by then have a buildbot set up and be taking more of an active role more in line with other active ports anyway.
> - R. Niwa
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the webkit-dev