1 |
On May 12, 2005, at 10:11 am, Patrick Lauer wrote: |
2 |
> On Wed, 2005-05-11 at 23:58 +0100, Stroller wrote: |
3 |
>> On May 11, 2005, at 8:10 pm, Ciaran McCreesh wrote: |
4 |
>>> |
5 |
>>> * Unique ID strings for packages, zynot style. Messy as hell though, |
6 |
>>> DEPEND="foo/bar {12379812AD7382164BD87678652438FC65E43A2}" doesn't |
7 |
>>> have |
8 |
>>> the same kind of ring to it... |
9 |
>> |
10 |
>> Maybe I'm just a messy person, but I really like this. |
11 |
> So does Microsoft. The registry has many entries where 128bit (?) |
12 |
> object-IDs are used. Very interesting to debug. |
13 |
|
14 |
I'm going to ignore that. This thread started because the current |
15 |
category/name naming convention causes interesting conditions. I |
16 |
appreciate that generally Microsoft Are Not Our Favourite Software |
17 |
Company, but giving them as an example doesn't inherently make unique |
18 |
IDs bad, and 128-bit ones are not necessary in this case. |
19 |
|
20 |
Also, before I start, I'd like to say that I know I'm not qualified to |
21 |
advocate this as a serious suggestion for adoption by Gentoo, so I'm |
22 |
just explaining _why I like it_. |
23 |
|
24 |
>> It prevents upstream naming collisions |
25 |
> But reduces readability for humans to zero. We don't want that. |
26 |
|
27 |
Humans are used to dealing with indexes - we remember phone numbers |
28 |
easily, and if we're looking up several things in a large volume, then |
29 |
we're used to using bookmarks or noting down page numbers. A six figure |
30 |
decimal packageID allows for a million packages in the Portage tree |
31 |
(and I'm assuming versions will be separate, anyway), a five figure hex |
32 |
ID would allow far more. |
33 |
|
34 |
Yes, arbitrary unique IDs would require an index tool to access ebuild |
35 |
name / category data, but surely there is little choice if |
36 |
naming-collisions are to be avoided and multiple categories are |
37 |
desired? Surely any human-focused naming convention will cause |
38 |
collisions and introduce potential for confusion? The current |
39 |
categories divide collisions into separate spaces, but they don't solve |
40 |
the problem of foo-player being eligible for both the media-CDplayers |
41 |
and audio-mp3rippers categories. |
42 |
|
43 |
> At least you haven't tried to optimize it all by using XML ... |
44 |
>> but the rest of us will use |
45 |
>> `esearch -o "%p\n" "" | grep -e category -e keyword`. |
46 |
> *head explodes* |
47 |
> No. |
48 |
|
49 |
That's the first time I used that command, but it only took me two |
50 |
minutes to look up & test. Since a dedicated index tool would clearly |
51 |
be required, I'm sure it would have better & more useful syntax. |
52 |
Currently I assume that Mr Harring searches for all the applications in |
53 |
a category by typing `ls -d /usr/portage/app-category/*` - wouldn't it |
54 |
be easier to use `esearch --category country`. Not only would it list |
55 |
them all, but multiple categories per package would also allow those to |
56 |
be shown that might debatably be categorised as "western". |
57 |
|
58 |
> ...It might make portage more resilient to one kind of problem, |
59 |
> but forget debugging then. |
60 |
|
61 |
Do we have 65000 unique packages in the tree? Would a four figure hex |
62 |
"part number" be that hard to remember when you're debugging package |
63 |
names? |
64 |
|
65 |
Again: I know I'm not qualified to advocate this as a serious |
66 |
suggestion for adoption by Gentoo, so I'm just explaining _why I like |
67 |
it_. |
68 |
|
69 |
Stroller. |
70 |
|
71 |
-- |
72 |
gentoo-dev@g.o mailing list |