| 1 |
El lun, 02-03-2009 a las 13:01 -0800, Zac Medico escribió: |
| 2 |
> || ( ( x11-libs/qt-gui:4 x11-libs/qt-webkit:4 ) x11-libs/qt:4 ) |
| 3 |
> |
| 4 |
> It's a variation of bug 161953 [1]. For this particular variation, |
| 5 |
> at the moment I don't think there's a good way to distinguish that |
| 6 |
> the choice on the left is the better choice. |
| 7 |
|
| 8 |
>From my point of view, I think that portage should do something like |
| 9 |
calculating how much a dep is already in the system, for example, in |
| 10 |
this case, I have 50% of ( x11-libs/qt-gui:4 x11-libs/qt-webkit:4 ) |
| 11 |
already installed on my system while 0% of x11-libs/qt:4. Then, |
| 12 |
( x11-libs/qt-gui:4 x11-libs/qt-webkit:4 ) would be a better option |
| 13 |
because, hopefully, a user that has already installed a portion of this |
| 14 |
dep, will prefer to maintain it installed |
| 15 |
|
| 16 |
But, this is only theoretical as I am not a programmer and I don't know |
| 17 |
if this could be implemented |
| 18 |
|
| 19 |
> However, if we |
| 20 |
> implement PROPERTIES=virtual [2] then we can use that to tag the |
| 21 |
> qt-4 ebuild as a virtual and then we should get better behavior due |
| 22 |
> to virtual lookahead mechanism that has been implemented for bug |
| 23 |
> 141118 [3]. In the absence of PROPERTIES=virtual support, we'd have |
| 24 |
> to use a more complex approach such as the "avoid redundant |
| 25 |
> upgrades" algorithm suggested for bug 260225 [4]. |
| 26 |
> |
| 27 |
> [1] http://bugs.gentoo.org/show_bug.cgi?id=161953 |
| 28 |
> [2] |
| 29 |
> http://archives.gentoo.org/gentoo-dev/msg_9d449a18a96a25a547fcfd40544085cf.xml |
| 30 |
> [3] http://bugs.gentoo.org/show_bug.cgi?id=141118 |
| 31 |
> [4] http://bugs.gentoo.org/show_bug.cgi?id=260225 |
| 32 |
> - -- |
| 33 |
> Thanks, |
| 34 |
> Zac |
| 35 |
|
| 36 |
This seems another option but, implementing previous one, would allow to |
| 37 |
cover more possible cases |
| 38 |
|
| 39 |
Best regards :-) |
Archives