| 1 |
Ciaran McCreesh posted on Mon, 13 Apr 2015 17:49:13 +0100 as excerpted: |
| 2 |
|
| 3 |
> On Mon, 13 Apr 2015 06:36:44 +0000 (UTC) |
| 4 |
> Duncan <1i5t5.duncan@×××.net> wrote: |
| 5 |
>> What about "initial-build" for the flag name, and making it part of the |
| 6 |
>> next EAPI, such that PMs know how to handle it without involving the |
| 7 |
>> user having to set it, and indeed, actually ignore the flag (as a |
| 8 |
>> masked flag) in normal operation if the user /does/ set it? |
| 9 |
> |
| 10 |
> You haven't thought your cunning plan all the way through. What if cmake |
| 11 |
> is installed but broken and needing a revdep-rebuild? |
| 12 |
|
| 13 |
Two ways to deal with that. |
| 14 |
|
| 15 |
1) Unmerge cmake, so it's an initial build again. (Depending on the |
| 16 |
conflict, an unmerge/clean-merge being the simplest resolution isn't |
| 17 |
uncommon. With all of @world in subsets here, portage unmerges, with a |
| 18 |
warning either that it's still in set X, or that it's a dep, so it never |
| 19 |
leaves @world and a standard emerge @world will bring it back.) |
| 20 |
|
| 21 |
2) Masked USE flag, so users can override the same way they can other |
| 22 |
masked USE flags. |
| 23 |
|
| 24 |
Also, being a masked USE flag, it's displayed, just as masked, so a user |
| 25 |
can know it's there. |
| 26 |
|
| 27 |
Tho your point remains valid at some level as those are both "advanced" |
| 28 |
hoops that a user arguably shouldn't be /forced/ to jump thru, even if |
| 29 |
doing so might be the most convenient fix, at times. |
| 30 |
|
| 31 |
-- |
| 32 |
Duncan - List replies preferred. No HTML msgs. |
| 33 |
"Every nonfree program has a lord, a master -- |
| 34 |
and if you use the program, he is your master." Richard Stallman |
Archives