Is the kernel piece going to be open source too, or are they trying to implement the _entire_ driver in userspace (except for using already-existing mainline Linux kernel components)?
I really wonder how many project creation requests start off with Daniel asking whether a mailing list is really needed. Why else would one request a project unless to get a) git repositories b) a mailing list c) bugzilla. Asking whether a mailing list is really needed but not asking whether creating a separate project is really needed then seems quite off.
Daniel has a long history of attacking radeonhd from the X.org/Fd.o side (there was this mailing list thing, there was him silently dropping radeonhd from the build script, and then it was of course his fd.o admin access that was used to hack/vandalize the long unused radeonhd repository), while not contributing much in code/software to the radeonhd versus radeon feud.
I was hoping that after RadeonHD's rather disgusting climax, things would've changed at fd.o, that the same games would not be repeated. I could not have been more wrong, and i was very stupid to even try filing for a project at fd.o.
If I understand correctly, Daniel's question was not whether a mailing list is needed, but whether a separate mailing list specifically for that driver is needed.
In the last year, daniels has not asked this question for any other project request. He has asked questions when some guys from Nokia's Meego (he used to work for that for a long time, but not until the point where Elop killed it) came and asked for a project, this until those guys gave up and went somewhere else. One project request from collabora was immediately created. Besides that, he created a few projects, without asking any noteworthy questions, in the usual timeframe. He even got briefly involved with the openchrome project request (i too was involved there, suggesting to james simmons to get a user account and use some user trees for his VIA playgrounds). In the end, openchrome, a project with little development and even less users (yes, i know, my unichrome work has even less users - i do not need to be told that), got both openchrome-devel and openchrome-users, and in the last 2 months, those mailing lists weren't used at all. In the last year, nobody was asked, not by daniels, not by anyone else, whether a mailing list really was necessary.
One would also think that a guy like me, with all the experience i have in dealing with driver projects, would know whether or not a mailinglist is needed, and would definitely not ask for one (and only one!) lightly. I have quite a few playgrounds on people.fd.o for which i wouldn't even think of getting more infrastructure. Nobody else with fd.o admin access would have asked whether a mailing list really was necessary.
That's one side of it. The other side is clearly documented in that bug. When we asked in 2007, and Egbert Eich asked the first question (he was one of the founding members of the current X.org and spent ages creating bylaws and did one of the first X.org releases after xfree86), we first got such a question back. When we then insisted, and stated that this mailinglist would see a lot of traffic quickly, he said "no, use xorg@". This was only the start of a very very nasty sequence of events.
If anyone else had asked that, my answer would have been very different indeed. But nobody else did ask that. Nor would they have dreamt of doing so.
Reply With Quote
Phoronix News Bot
Announcing The Lima Open-Source GPU Driver
Posting Permissions