| lifeless | persia: so | 01:12 |
|---|---|---|
| lifeless | persia: I have dug | 01:12 |
| lifeless | persia: its terrible code quality; terrible infrastructure | 01:13 |
| persia | Yes. | 01:13 |
| lifeless | crying out to be autotools or some other HLL based environment | 01:13 |
| lifeless | the .so files the iegd config thing makes are X 1.6 at best | 01:14 |
| lifeless | so 1.7 won't load them | 01:14 |
| lifeless | no obvious source :- cannot fix (except perhaps, by writing a 1.6->1.7 driver loader) | 01:14 |
| persia | Is there a quick hack available like there was with the displaylink stuff? (change a couple calling conventions for >> 6) | 01:14 |
| lifeless | I'm going to to back-burner this till 10.4 or whatever is released | 01:14 |
| lifeless | persia: if I had source, likely. | 01:15 |
| persia | Oh, right. Forgot about that bit. | 01:15 |
| lifeless | we get kernel module source; not X driver source. | 01:15 |
| lifeless | terrible. anyhow, thats just a FYI | 01:16 |
| lifeless | StevenK: ^ | 01:16 |
| lifeless | 'do not try 10.3 on lucid, its a fail' | 01:16 |
| persia | So the conclusion is that it needs a weeks work to have a sane buildsystem, etc., and a couple days work to be compatible with modern X, but needs source for either? | 01:16 |
| lifeless | persia: If I was upstream, its probably under a week to adjust the X driver to X1.7 - at most | 01:17 |
| lifeless | the install script is GPL of all things; we could start fixing that now, but no public VCS etc make that a likely losing proposition | 01:17 |
| lifeless | What it /needs/ is X 1.7 compat - it passes the selftest for the new kernel modules | 01:17 |
| lifeless | there is still a dedicated drm, but its now namespaced not to mess with other drm modules. | 01:18 |
| lifeless | everything else is dealable with debian/patches or whatever we might choose to use. | 01:18 |
| lifeless | oh, and its non redistributable - the version I got anyway. So we'd need a mscorefonts approach to doing it. | 01:18 |
| lifeless | It looks like its DKMS'able-with-some-effort. | 01:19 |
| persia | smcorefonts+dkms+build-time patching sounds like sufficient pain as to deserve to wait until other things are sorted. | 01:19 |
| lifeless | right | 01:20 |
| persia | Unfortunate that the hardware involved happens to be inexpensive. | 01:20 |
| lifeless | no point until the blob they ship can work | 01:20 |
| lifeless | and popular | 01:20 |
| persia | It's only popular because it's inexpensive. Even in environments where there *is* a working driver (e.g. hardy, jaunty, windows), the performance is apparently only acceptable. | 01:21 |
| lifeless | the newer drivers manage accelerated mpeg, 3d games at ok frame rates | 01:21 |
| lifeless | from what I've read | 01:21 |
| persia | Indeed. "only acceptable". | 01:22 |
| lifeless | true | 01:30 |
| lifeless | however there are plenty of CPU's more expensive than an entire poulsbo lapop | 01:30 |
| lifeless | later | 01:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!