| 1 |
Chris Gianelloni posted <1120255303.13274.4.camel@×××××××××××××××××.net>, |
| 2 |
excerpted below, on Fri, 01 Jul 2005 18:01:43 -0400: |
| 3 |
|
| 4 |
> Last time that I checked, each arch needs at least one non-cascaded |
| 5 |
> profile, due to older versions of portage not working with cascaded |
| 6 |
> profiles. Either this, or users will not be able to upgrade from old |
| 7 |
> installations that have not been upgraded for some time. |
| 8 |
|
| 9 |
IIRC (because I made a similar argument, tho likely not as well) when this |
| 10 |
came up the last time, the decision was to create a rescue site with one |
| 11 |
such profile per arch, getting them out of the tree and uncluttering it. |
| 12 |
|
| 13 |
Either that, or combine it with the rescue portage project, such that they |
| 14 |
can download and untar a functionally usable cascade understanding portage |
| 15 |
version to get them out of between the rock and the hard place, using the |
| 16 |
same functionality now in place for when portage itself crashes. This |
| 17 |
alternative may actually already be in place only we hadn't thought about |
| 18 |
it. |
| 19 |
|
| 20 |
The idea being... If someone has been offline for two years or whatever, |
| 21 |
it's really sort of unreasonable for them to expect a problem free upgrade |
| 22 |
in the first place. They can post a question on the lists or forums or |
| 23 |
irc, and be directed to the proper location and procedure as necessary. |
| 24 |
Alternatively, and /not/ that much less practically if you think about it |
| 25 |
anyway, given the number of packages they'll have to update if they've |
| 26 |
been offline for well over a year, they can simply download the latest |
| 27 |
LiveCD and start over with a clean install, even stage-3 plus GRP, if they |
| 28 |
are impatient, and be caught-up with far less hassle than attempting to do |
| 29 |
it in-place, starting from such an old snapshot that there's been no |
| 30 |
testing nor real consideration of the upgrade path in the first place, |
| 31 |
thereby creating far more hassle than necessary for themselves, getting |
| 32 |
everything working. |
| 33 |
|
| 34 |
IOW, just as Gentoo as it exists today isn't really suitable for nor does |
| 35 |
it support the multi-year "freeze-frame" snapshots plus security-only |
| 36 |
update routine of strict enterprise policy, because that simply doesn't |
| 37 |
fit the continual update community focused distribution model Gentoo has |
| 38 |
chosen, so a similarly outdated "offline for two years" Gentoo |
| 39 |
installation cannot be expected to be able to update as if it were last |
| 40 |
updated a week ago. The Gentoo model does not support such, nor, without |
| 41 |
dividing scarce developer resources, can it be made to do so, regardless |
| 42 |
of whether those last non-cascading profiles remain in place or not. The |
| 43 |
rescue portage may well work, but even then, there will be other issues. |
| 44 |
If it's been a year, chances are a from-latest-stage-X upgrade will be |
| 45 |
about even hassle compared to an upgrade in place. If it's been a year |
| 46 |
and a half, things favor the from-stage-X upgrade. If it's been two years |
| 47 |
or longer, things VASTLY favor the from-stage-X upgrade. That's just the |
| 48 |
way it is, with Gentoo as it exists today. If it's not the profiles |
| 49 |
causing the issue, it'll be some other incompatibility causing headaches, |
| 50 |
and whether or not they can be overcome, from-stage-X is simply going to |
| 51 |
be less of an issue, and be easier to support, because others will have |
| 52 |
likely run into similar problems, so the answers will be easier to find. |
| 53 |
|
| 54 |
-- |
| 55 |
Duncan - List replies preferred. No HTML msgs. |
| 56 |
"Every nonfree program has a lord, a master -- |
| 57 |
and if you use the program, he is your master." Richard Stallman in |
| 58 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
| 59 |
|
| 60 |
|
| 61 |
-- |
| 62 |
gentoo-dev@g.o mailing list |
Archives