1 |
On 20-04-2007 09:11:10 -0500, Yuri Vasilevski wrote: |
2 |
> On Fri, 20 Apr 2007 14:52:46 +0200 |
3 |
> Fabian Groffen <grobian@g.o> wrote: |
4 |
> > > Is the info you need in CVS/ ? |
5 |
> > |
6 |
> > That requires: |
7 |
> > a) an up-to-date CVS tree checkout |
8 |
> > b) a CVS tree (not a regular rsynced one) |
9 |
> > c) parsing of the CVS/Entries file |
10 |
> > |
11 |
> > So in principle, yes I think it can be done using the CVS directory, |
12 |
> > but I prefer not to. |
13 |
> |
14 |
> +1 on this one. |
15 |
> |
16 |
> I am writing some script to automate package building and use the |
17 |
> Header field to detect CVS revision changes. I know I could do it with |
18 |
> hash values, but then it would be harder to have builds from |
19 |
> different revisions of the same ebuild file time ordered. |
20 |
|
21 |
I thought about it some longer, and it's even nastier than just having |
22 |
to get the value in a different way, because you also have to keep the |
23 |
value in the file in your "local" version. This either means also |
24 |
storing the CVS directory (in SVN, funny :) ) or devoting the first line |
25 |
of every file for it... |
26 |
|
27 |
Side question: is it a problem that you cannot easily see which (CVS) |
28 |
version of an ebuild is in the rsync tree you/user have/has? I guess |
29 |
that's what the $Header: $ thing is actually for. I can't recall any |
30 |
occasion myself that I used it to track an issue with it though. |
31 |
|
32 |
|
33 |
-- |
34 |
Fabian Groffen |
35 |
Gentoo on a different level |
36 |
|
37 |
-- |
38 |
gentoo-dev@g.o mailing list |