1 |
On Tue, 2003-10-21 at 22:11, Eric Sammer wrote: |
2 |
> Donnie Berkholz wrote: |
3 |
> > It works as is anyway, so what's the point? |
4 |
> |
5 |
> The point was that if the version of the synaptics driver changes prior |
6 |
> to the release of the xfree .15 ebuild, it would require those of us |
7 |
> using it to recompile xfree at least one extra time. |
8 |
|
9 |
There's no requirement to recompile xfree unless you _need_ a feature of |
10 |
the newer synaptics driver. Yes it is nice to have the latest version, |
11 |
but there's no purpose unless you require a fix in it. |
12 |
|
13 |
I'm glad you can appreciate my counterargument. |
14 |
|
15 |
> The alternate |
16 |
> argument is that it would force those not using it to recompile X and |
17 |
> yes, that is obviously a much larger number. Also, it seemed redundant |
18 |
> to have a synaptics ebuild that is just sitting there and have the xfree |
19 |
> ebuild doing it "by hand." |
20 |
|
21 |
I committed the synaptics ebuild in advance so I would have one less |
22 |
thing to do later. The xfree "by hand" synaptics is legacy, since as I |
23 |
said it wasn't possible to compile the synaptics driver externally until |
24 |
0.11.8, which was released just under two weeks ago. |
25 |
|
26 |
> It makes maintaining it difficult by having |
27 |
> it in two places. |
28 |
|
29 |
There is nearly zero maintenance cost within the xfree ebuild for the |
30 |
synaptics driver (simply change the driver version and go), and similar |
31 |
holds true for the ebuild. I expect any extra maintenance caused between |
32 |
13 October (when I committed synaptics) and whenever 4.3.99.15 is |
33 |
released to be negligible. |
34 |
|
35 |
> There is also the fact that xfree 4.3.99.* is still |
36 |
> masked so I didn't think releasing a new version of the ebuild would be |
37 |
> a big deal as it is uber-unstable. |
38 |
|
39 |
In actuality, it is not uber-unstable. I've been running 4.3.99.x for a |
40 |
fair while and have yet to encounter a crash that is 4.3.99.x-specific. |
41 |
It's masked by virtue of being a development series. |
42 |
|
43 |
> It's rather subjective, and as someone who has to use the combination in |
44 |
> question (xfree 4.3.99.x, synaptics, and the 2.6 kernel series (due to |
45 |
> an IGP320M chipset)), it would be easier to have the xfree ebuild use |
46 |
> the synaptics ebuild rather than pull it in itself. |
47 |
|
48 |
You don't need to use 2.6 for IGP chipsets. See the 2.4 patches: |
49 |
http://bugs.xfree86.org/show_bug.cgi?id=314 |
50 |
|
51 |
> It's not that big of |
52 |
> a deal, really. I can do it and just use it in PORTAGE_OVERLAY if it's |
53 |
> not something that gentoo is interested in. |
54 |
|
55 |
This isn't Gentoo, this is me. You're welcome to grab the 4.3.99.14-r1 |
56 |
ebuild from my overlay if you choose [1], but I don't consider it |
57 |
appropriate for committing. Emerge it with `INPUT_DEVICES="synaptics" |
58 |
emerge xfree` or preferably set INPUT_DEVICES in make.conf. |
59 |
|
60 |
1. http://dev.gentoo.org/~spyderous/overlay/ |