From: Eric S. R. <es...@th...> - 2000-03-08 00:52:58
|
James Simmons <jsi...@ac...>: > Looks good. Something nice is maybe add is previous and present > approaches to a new console system such as Vojtech work > (http://www.suse.cz/development/input), I'll add a pointer to the "Related Resources" section. > Evstack (look in archive) The EvStack site at pepsi.visus.com has apparently disappeared. > and > probably other solutions that will come out of the woodwork. What > are/where the basic ideas behind them. What problems they had/have. Then > we can explain our approach. How we are attempted to solve these problems > they are/have encountered. I'm not qualified to write these explanations. If somebody on the list is, rough up something (plain text is OK) and I'll webify it. > As we figure out what we are going to do we can write a long term page > since the things on the page right now could go in with the current 2.3.X > kernels. Well...deliverables 1 and 2 maybe. Hm. We need names for these. I dub them "Sapphire" (2.2.x emulation patch) and "Emerald" (2.3.x emulation patch). The more advanced stuff in deliverable 3 ("Ruby") really is new features and will probably have to wait for 2.5.x if I know Linus. I think we need to first develop a modest, easily comprehensible set of Sapphire and Emerald patches that bug-fix the emulation, without trying to hit the larger issues yet. By doing this, and getting these patches accepted, we'll open the door for more ambitious things later. I'm working on Sapphire right now. I've downloaded Dominik's 2.2.1 patch and am reconciling it with 2.2.12. My objective is to produce a minimal patch that doesn't include a whole bunch of formatting changes. I'll test that patch on my stock Red Hat 6.1 system. Once I have that verified, I'll generate a patch for console_codes(4) that updates it appropriately. Dominik, you could speed things up by sending me a functional description of your changes. Next step after that would be to generate a minimal Emerald patch, test it, and check that the updated console_code(4) page correctly describes it. Once we have these things, we can should be able to deliver a neat package to Linus and Alan that will bring both kernel lines up to snuff in the emulation department and document that properly. Then we can take on the bigger stuff. -- <a href="http://www.tuxedo.org/~esr">Eric S. Raymond</a> Everything you know is wrong. But some of it is a useful first approximation. |