| 1 |
felix@×××××××.com wrote: |
| 2 |
> On Sat, Nov 05, 2005 at 03:44:30AM +0900, Jason Stubbs wrote: |
| 3 |
> |
| 4 |
> |
| 5 |
>>Nobody else has shown a real example, why should I? |
| 6 |
>>... |
| 7 |
>>I am focusing on what it could do. I stated all the options in my previous |
| 8 |
>>email. |
| 9 |
>>... |
| 10 |
>>To restate: How often is it there is exactly one masked version available? |
| 11 |
>>What to do when there are two? |
| 12 |
> |
| 13 |
> |
| 14 |
> How old are you? You sound like some crotchety old fart on a rocking |
| 15 |
> chair on his porch. |
| 16 |
> |
| 17 |
> Good god. Probably once or twice a month I read about some program |
| 18 |
> that sounds interesting, run emerge -p on my amd64, it complains that |
| 19 |
> some dependency is masked, I edit /etc/portage/packages.keywords, |
| 20 |
> emerge -p again, get another complaint about some other top level |
| 21 |
> dependency, rinse, lather, repeat, until I have a half dozen additions |
| 22 |
> or give up in disgust. If you have never had this happen, then I feel |
| 23 |
> sorry for you for being so unadventurous. |
| 24 |
> |
| 25 |
> And the only way I can provide a real world example is wait til it |
| 26 |
> comes around again on the gitar, to quote Arlo, and I am not going to |
| 27 |
> waste my time trying to remember to come back to this dicsussion then, |
| 28 |
> it will be quite cold in its grave, obviously where you want it to go. |
| 29 |
> |
| 30 |
> I swear you have got to be just about the most negative pessimistic |
| 31 |
> whining poster on this list. |
| 32 |
> |
| 33 |
|
| 34 |
Because insulting a portage developer is sure to get your feature |
| 35 |
implemented. Er...wait a minute ;) |
| 36 |
|
| 37 |
Why do a half assed job that only prints 1 or 2 levels deep? Even then |
| 38 |
things can get complicated, especially when there are a lot of masked |
| 39 |
packages and dependencies ( try emerging enlightenment cvs, for example, |
| 40 |
like 15+ packages..or god.Modular X even :) ). I guess you could do a |
| 41 |
iterative DFS, although again, hard to tell when to stop iterating and |
| 42 |
give up. |
| 43 |
|
| 44 |
Nothing here is simple, there are a lot of choices here, and if there |
| 45 |
are a lot of masked packages, the depgraph gets deep very fast. If I |
| 46 |
try to emerge, say an xorg-meta build, there are 100+ masked ebuilds I |
| 47 |
have to unmask. Maybe the meta doesn't depend on exact versions and for |
| 48 |
each masked ebuild there are two versions available. Do you want |
| 49 |
portage to print all 200 canidates out and be like, hey you need to |
| 50 |
unmask these. There are a bunch of simple canidates out there for this |
| 51 |
that make it easy in concept, but there are a few canidates out there |
| 52 |
that make this very difficult, due to their size. In the example above, |
| 53 |
we only went 1 level deep in the tree and found 200 masked canidates, |
| 54 |
there is no way to print less ( unless you just hard mask some integer |
| 55 |
in there ). |
| 56 |
|
| 57 |
If you can come up with some simple way to handle these complex cases, |
| 58 |
by all means, otherwise the solution will work most of the time and die |
| 59 |
horribly on the complex cases, and IMHO, that is not acceptable. |
| 60 |
|
| 61 |
|
| 62 |
|
| 63 |
-- |
| 64 |
gentoo-portage-dev@g.o mailing list |
Archives