1 |
On Nov 18, 2011 9:27 PM, "Willie Wong" <wwong@××××××××××××××.edu> wrote: |
2 |
> |
3 |
> On Thu, Nov 17, 2011 at 07:41:21PM +0000, James wrote: |
4 |
> > > Now, why can't the USE descriptions be like the kernel option |
5 |
> > > descriptions and have something like what Pandu wrote included? |
6 |
> > |
7 |
> > I added this to root's .bashrc a long time ago: |
8 |
> > |
9 |
> > # USE flag settings hack by Ciaran McCreesh: |
10 |
> > explainuseflag(){ sed -ne "s,^\([^ ]*:\)\?$1 - ,,p" $(portageq |
11 |
> > portdir)/profiles/use.{,local.}desc; } |
12 |
> > alias ef="explainuseflag" |
13 |
> > |
14 |
> > |
15 |
> > Then simply use the alias for a quick check to learn about all the |
16 |
different |
17 |
> > uses of a given flag: |
18 |
> > |
19 |
> > 'ef graphite' |
20 |
> > |
21 |
> > # ef graphite |
22 |
> > Enable support for non-Roman fonts via media-gfx/graphite2 |
23 |
> > Enable support for non-Roman fonts via media-gfx/graphite2 |
24 |
> > Add support for the framework for loop optimizations based on a |
25 |
polyhedral |
26 |
> > intermediate representation |
27 |
> > |
28 |
> > Then drill down into the a specific package's use flag meaning, using |
29 |
the |
30 |
> > aforementioned 'equery u' delineated by Albert. |
31 |
> |
32 |
> You people seem to miss my point. I know perfectly well how to find |
33 |
> the USE descriptions. It is just that the USE description, in this |
34 |
> case (as in many others) isn't terribly useful. |
35 |
> |
36 |
> "Add support for the framework for loop optimizations based on a |
37 |
> polyhedral intermediate representation" means absolutely gibberish to |
38 |
> me. |
39 |
> |
40 |
> But if one were to add an additional one or two lines a la Pandu, |
41 |
> about how it is supposed to make " gcc-4.5.3 use a newer method to |
42 |
> detect parallelism, thus (potentially) makes programs compiled by gcc |
43 |
> to have better multithreaded performance" and perhaps even a Kernel |
44 |
> help page style "It is mostly stable. If unsure, say Yes." |
45 |
> |
46 |
> It would be ever so much more helpful for people who would like to |
47 |
> find out what new flags do before deciding whether or not to follow |
48 |
> the default recommended by the devs which are set into the profile. |
49 |
> |
50 |
> (I'm not saying this type of hand holding is necessary for all flags: |
51 |
> "enable support for non-Roman fonts via media-gfx/graphite2" is |
52 |
> perfectly understandable, as are most other flags about features a |
53 |
> "user" is likely to interact with. But for some of the more "system" |
54 |
> type flags (see also that python/perl flag business from the recent |
55 |
> months), I think the USE descriptions can stand some improvement.) |
56 |
> |
57 |
|
58 |
I agree with you (and not because my name is mentioned :-P). |
59 |
|
60 |
I got lucky with USE "graphite": gcc's homepage is quite clear; a 15-minute |
61 |
reading convinced me to try graphite. But there are still lots of other USE |
62 |
flags that sent me on hours of goose-chase before I can enable/disable with |
63 |
conviction. |
64 |
|
65 |
I'm not sure where to put the more detailed explanations, though; perhaps a |
66 |
$PN.usedesc file in the package's directory? Kind of a complement to the |
67 |
.ebuild file(s). |
68 |
|
69 |
Rgds, |