| 1 |
On Wednesday 02 April 2008, Walter Dnes wrote: |
| 2 |
> > To fix: unmerge the existing openmotif, merge the new one |
| 3 |
> |
| 4 |
> But, but, but... isn't the whole point of "--update --world" to |
| 5 |
> *UPDATE* to the latest available version? Why is it that openmotif |
| 6 |
> requires *MANUALLY* deleting the old version, when 99% of |
| 7 |
> non-slottable apps and libs allow portage to delete the old version |
| 8 |
> and replace it with the new version? I'm obviously missing something |
| 9 |
> here. |
| 10 |
|
| 11 |
In general that is the way it works. Neil posted to that effect a few |
| 12 |
hours ago. However, just because it usually works that way doesn't mean |
| 13 |
it always should or will work that way. There's this pesky thing |
| 14 |
called "real life" and the pesky "corner cases" that come with it. |
| 15 |
|
| 16 |
Sometimes a new version of a package simply cannot co-exist with an |
| 17 |
older version because of the way it's coded. Portage first builds a |
| 18 |
package and installs it to a sandbox environment to ensure that the |
| 19 |
package is buildable. Only when tis is established and proven, will it |
| 20 |
delete the old version and move the new one onto the live filesystem. |
| 21 |
If you have this co-exist problem with that package, you have to |
| 22 |
manually unmerge the older one, as portage will not take the risk of |
| 23 |
leaving you with an unusable package (that correctly is a decision that |
| 24 |
a human must take, it is TheUnixWay(tm) and |
| 25 |
TheOneTrueWayOfComputing(tm). |
| 26 |
|
| 27 |
The lesstif/openmotif mess is slightly different. Openmotif has an odd |
| 28 |
licensing scheme, so odd the Fedora (I think) has decided to not |
| 29 |
support it at all. The relevant Gentoo dev wants to set things up so |
| 30 |
that you have to have one or the other but not both. The sanest way to |
| 31 |
deal with this is to force the user to make a decision and implement |
| 32 |
it, and to not have to put up with whatever choice the dev happens to |
| 33 |
think is ideal. This could have been better documented. In fact, there |
| 34 |
should have been some documentation about it at all, but there isn't - |
| 35 |
just a mysterious blocker. I only know it because I read the new |
| 36 |
comments in $PORTDIR/profiles/packages.mask after every sync :-) And I |
| 37 |
see those comments are no longer there :-( |
| 38 |
|
| 39 |
It's just one of those things that we have to put up with occasionally |
| 40 |
in real life |
| 41 |
|
| 42 |
-- |
| 43 |
Alan McKinnon |
| 44 |
alan dot mckinnon at gmail dot com |
| 45 |
|
| 46 |
-- |
| 47 |
gentoo-user@l.g.o mailing list |
Archives