1 |
On Thu, Mar 04, 2010 at 10:08:06PM +0100, Sebastian Pipping wrote: |
2 |
> 1. Copy |
3 |
> ======= |
4 |
> - Duplicate any traces of dev-util/${PN} in profiles/ to dev-vcs/${PN} |
5 |
> (fgrep -Rw "dev-util/${PN}" profiles/) |
6 |
> |
7 |
> - Copy complete package dev-util/${PN} to dev-vcs/${PN} |
8 |
> (watch out CVS directories) |
9 |
> |
10 |
> |
11 |
> 2. Switch |
12 |
> ========= |
13 |
> - Update eclasses |
14 |
> fgrep -w "dev-util/${PN}" eclass/*.eclass |
15 |
> |
16 |
> - Update reverse dependencies |
17 |
> http://tinderbox.dev.gentoo.org/misc/dindex/dev-util/${PN} |
18 |
> http://tinderbox.dev.gentoo.org/misc/rindex/dev-util/${PN} |
19 |
> ^^^ |
20 |
> fgrep -w "dev-util/${PN}" */*/*.ebuild |
21 |
|
22 |
I'd suggest |
23 |
|
24 |
pquery --revdep dev-util/${PN} --raw --repo $PORTDIR |
25 |
|
26 |
instead- throw in a '--attr inherited' if you want to look for |
27 |
eclasses that are implicated in addition, but it's a safer bet then |
28 |
relying on fgrep (since it's literal metadata searching) and doesn't |
29 |
have the potential for staleness tinderbox does. Speed wise it'll be |
30 |
slower then grep, but not hugely so- reasonably up to date cache, it's |
31 |
<9s on my hardware so... there's no reason to not do it properly. |
32 |
|
33 |
|
34 |
> - Append "move dev-util/${PN} dev-vcs/${PN}" line to |
35 |
> profiles/updates/1Q-2010 |
36 |
> (must come last, described in [1]) |
37 |
> |
38 |
> |
39 |
> 3. Remove |
40 |
> ========= |
41 |
> - Remove package dev-util/${PN} from CVS (as in [2]) |
42 |
> cd dev-util |
43 |
> cvs rm -Rf ${PN} |
44 |
> cvs ci -m "dev-util/${PN}: Remove (renamed to dev-vcs/${PN})" ${PN} |
45 |
|
46 |
|
47 |
Realistically, the dev *really* should be doing visibility scans |
48 |
after their efforts- yes bones does them too, but it's better to do |
49 |
the scan after to ensure any breakage is caught quickly. Specifically |
50 |
via pkgcore-checks, I'd suggest running |
51 |
|
52 |
pcheck -r $PORTDIR -c visibility '*' |
53 |
|
54 |
prior to the move, and a dump of that commands output after the cvs rm |
55 |
invocation above. Presuming everything was updated properly, the |
56 |
line count should be the same- output will vary slightly (previous |
57 |
complaints about dev-util/bzr being unavailable become dev-vcs/bzr |
58 |
unavailable specifically). |
59 |
|
60 |
At least for my hardware, it's a 3.5 minute scan- so there really |
61 |
isn't a reason to *not* do this check (if the mips profiles were |
62 |
cleaned up/out prior it's more like 2.5 minutes on a sidenote)... |
63 |
|
64 |
|
65 |
If folks are aware of alternate tools for doing these sort of checks |
66 |
in a timely fashion, feel free to pipe up also- I'm just suggesting |
67 |
the tools I know. |
68 |
|
69 |
|
70 |
Random sidenote, anyone looked at using an alternate vcs to do the |
71 |
work, then proxy it back? Specifically thinking of workflow like svk |
72 |
(or in this case hg cvs, |
73 |
https://wiki.mozilla.org/Using_Mercurial_locally_with_CVS ). The |
74 |
reason I ask is that via building the work up outside of cvs, then |
75 |
proxying the add/remove/modifications back into it, it should be |
76 |
possible to minimize the window of cvs breakage down to bare minimum |
77 |
while still getting the same level of QA validation for the changes. |
78 |
|
79 |
That's assuming hg/cvs behaves sanely of course, and doing a checkout |
80 |
(tip as it were) from cvs into hg is fast enough to be viable. Just a |
81 |
thought- djc, you got any comments on this usage of hg? |
82 |
|
83 |
~harring |