1 |
On Wed, 21 Jan 2015 01:13:08 +0000 (UTC) Duncan wrote: |
2 |
> Andrew Savchenko posted on Tue, 20 Jan 2015 23:59:23 +0300 as excerpted: |
3 |
> |
4 |
> > On Tue, 20 Jan 2015 12:17:35 -0800 Christopher Head wrote: |
5 |
> >> On January 20, 2015 12:47:03 AM PST, Alexis Ballier |
6 |
> >> <aballier@g.o> wrote: |
7 |
> >> >So, you're telling me that if you have a list of 90 cpu extensions, |
8 |
> >> >you will from time to time open that list to see if there is a 91st |
9 |
> >> >one added ? I think most people won't even notice, at best they'll |
10 |
> >> >look for the changelog. |
11 |
> >> |
12 |
> >> No, actually, I’m advocating the exact opposite. I’m saying that, as |
13 |
> >> long as the list file is kept up to date, then I will look at those 90 |
14 |
> >> flags when I first install and never again. If a 91st flag appears some |
15 |
> >> day, then as long as the file was maintained as I described in an |
16 |
> >> earlier message (i.e. flags are added as soon as manufacturers announce |
17 |
> >> features), I already know I can reliably ignore the new flag. After |
18 |
> >> all, |
19 |
> >> if the flag didn’t exist when I installed the system, then my CPU must |
20 |
> >> necessarily not have that feature—unless CPUs are in the habit of |
21 |
> >> sprouting new instructions after you buy them! |
22 |
> > |
23 |
> > Not exactly. CPUs are not in a habit, but software is. Some brand new |
24 |
> > instuction set may be supported in (any of) packages with some delay. |
25 |
> > Thus it is possible that instruction set supported by your CPU will |
26 |
> > appear in the list of cpu flags after your ininial install. |
27 |
> |
28 |
> PMFJI... |
29 |
> |
30 |
> chead's idea is (I believe) simply to have the description file updated |
31 |
> with all current hardware feature flags as soon as they are known (he |
32 |
> said announced, but sometimes they change between announcement and |
33 |
> actually appearing in hardware, so "known", as in "known to actually |
34 |
> appear in hardware", would seem to be better). |
35 |
> |
36 |
> That way, no matter what the software supported at the time and what |
37 |
> flags were thus actually used in packages, when someone first installs |
38 |
> gentoo on a new machine, or when they first upgrade their CPU to |
39 |
> something with new features, they can run the script and update their |
40 |
> use_expand to match their hardware _ONCE_, without worrying about whether |
41 |
> or when a package with support for it might appear. |
42 |
> |
43 |
> If no package with that support ever appears, no harm done, that entry in |
44 |
> the use_expand is simply never used. |
45 |
> |
46 |
> OTOH when some package /does/ get support for new hardware instructions |
47 |
> and adds the appropriate flag, it'll appear in portage's output, but |
48 |
> because the use_expand was already set when gentoo was installed or the |
49 |
> cpu upgraded, the user won't have to worry about needing to look it up |
50 |
> and decide whether to set it, again, it'll already be done, back when the |
51 |
> _hardware_ changed, _not_ sometime later, when the _software_ changed to |
52 |
> support the new hardware. |
53 |
> |
54 |
> Of course if the user upgrades hardware after a package supports a |
55 |
> feature, they'll have to upgrade their use_expand setting appropriately |
56 |
> or miss support for the new instructions, but that's always the case. |
57 |
> Just if handled as chead suggests, it'd be the case ONLY when the |
58 |
> hardware is updated, instead of every time a package upgrades its own |
59 |
> support. |
60 |
> |
61 |
> Correct, chead? Does that make things clearer, aballier and bircoph? |
62 |
|
63 |
Yeah, thanks. I misunderstood original intention. It's all clear |
64 |
now. |
65 |
|
66 |
Best regards, |
67 |
Andrew Savchenko |