1 |
> -----Original Message----- |
2 |
> From: George Shapovalov [mailto:george@g.o] |
3 |
> Sent: Sunday, December 07, 2003 11:40 AM |
4 |
> To: gentoo-portage-dev@g.o |
5 |
> Subject: Re: [gentoo-portage-dev] portage-ng concurse entry |
6 |
> Was: Updated Portage project page |
7 |
> |
8 |
> |
9 |
> On Saturday 06 December 2003 17:44, Jason Stubbs wrote: |
10 |
> > It's not getting ahead of things! That's a requirement that's not |
11 |
> > covered yet. "Package definition should be powerful but |
12 |
> simple with a |
13 |
> > small learning curve" or something to that effect. |
14 |
> |
15 |
> Hm, isn't it a bit too late to change ebuild format, with us |
16 |
> sitting on 7000+ |
17 |
> ebuilds? The only reasonable way to do so is to make it structurally |
18 |
> compatible and create a converter tool. Even then this is a |
19 |
> major endeavor |
20 |
> that would require a very good reason (nothing short of |
21 |
> deadly limitations of |
22 |
> the present format, which I woudn't say is the case). |
23 |
> Furthermore, this would |
24 |
> require wide publicity and even votes if we do not want to |
25 |
> alienate users, as |
26 |
> this is the change that definitely will affect them (take a |
27 |
> look at number of |
28 |
> new ebuild submissions ;)). |
29 |
|
30 |
There's another requirement! "The package system needs to be backward |
31 |
compatible or have a compatibility layer for existing ebuilds" or |
32 |
something to that effect. Having to rewrite them all just for a package |
33 |
manager migration would not only be insane but would be, dare I say, |
34 |
impossible - by the time they are all rewritten, they'd all have become |
35 |
obsolete. If there is backward compatibility, they should be able to be |
36 |
phased out in less than a year (assuming the format is in fact |
37 |
replaced). |
38 |
|
39 |
> But then I don't really see the problem with present format. |
40 |
> bash involvment |
41 |
> is really necessary only during the pkg_* and src_* steps, |
42 |
> when a lot of |
43 |
> other stuff is going to happen anyway, so this is hardly a |
44 |
> bottleneck. To get |
45 |
> definitions of various vars and dependency information out is |
46 |
> trivial and can |
47 |
> be done in anything. That bash is involved in this step at present is |
48 |
> unfortunate, but there were reasons for it and it definitely |
49 |
> may be undone |
50 |
> even for the present portage. |
51 |
|
52 |
Personally, I don't think the current ebuild format is so bad either |
53 |
while the dev team was relatively small. But there are a number of |
54 |
places that mistakes can be made making for bad QA. I think it does make |
55 |
sense to split the file into the pkg_*/src_* stuff and the various vars |
56 |
as per current requirements spec. I would also prefer to see the |
57 |
pkg_*/src_* stuff abstracted so that an ebuild is never a security risk. |
58 |
ATM, pkg_postinst is the worst that I can think of from a security |
59 |
viewpoint. |
60 |
|
61 |
Regards, |
62 |
Jason Stubbs |
63 |
|
64 |
|
65 |
-- |
66 |
gentoo-portage-dev@g.o mailing list |