| 1 |
On 03/29/10 23:05, Dag Sverre Seljebotn wrote: |
| 2 |
> Michael Haubenwallner wrote: |
| 3 |
>> Dag Sverre Seljebotn wrote: |
| 4 |
>> |
| 5 |
>>> There's lot of half-baked distribution solution available (sagemath.org, |
| 6 |
>>> femhub.org, enthought.com, pythonxy.com), but Gentoo seems like a better |
| 7 |
>>> choice for my project for several reasons. |
| 8 |
>>> |
| 9 |
>> |
| 10 |
>> After a short look on all those links and after reading bug#240622 I get |
| 11 |
>> the impression (may be wrong), that Gentoo Prefix _is_ what you're |
| 12 |
>> looking |
| 13 |
>> for - as a _replacement_ for sagemath (and others): |
| 14 |
>> It just seems necessary to isolate the math-only package(s) out of |
| 15 |
>> sagemath and write the ebuild(s) which uses already existent Gentoo |
| 16 |
>> package-definitions (=ebuilds) as dependencies. |
| 17 |
>> |
| 18 |
> That's what I'm hoping, anyway. |
| 19 |
> |
| 20 |
> The ebuilds won't be the problem, the ones I need are already in Gentoo |
| 21 |
> (and if not, François Bissey, who maintains a Sage overlay for Gentoo, |
| 22 |
> has agreed to help me get to know Gentoo). |
| 23 |
> |
| 24 |
> The summary seems to be (and it's how others have put it as well): |
| 25 |
> - Gentoo prefix has the advantage of a much more mature package |
| 26 |
> distribution system (and community around it) |
| 27 |
> - The ones mentioned has (what is in the community percieved as) the |
| 28 |
> advantage of using LD_LIBRARY_PATH (being easily movable, bundleable) |
| 29 |
> and using the system compilers and libtool (so they feel less like an |
| 30 |
> "OS within an OS" and more like a software distribution mechanism -- one |
| 31 |
> could argue this is not a good argument though). |
| 32 |
|
| 33 |
just to chime in on the movability: i successfully moved whole prefix |
| 34 |
installations to other directories, and they worked fine (hint: i used |
| 35 |
chpathtool to adjust paths) the only limitation is, that the path has to |
| 36 |
be the exact same length (or less long, but that one didn't work for me |
| 37 |
yet, but that could be fixable). of course this is not really |
| 38 |
convenient, but it should be possibly for you to create |
| 39 |
|
| 40 |
"/home/user/prefix-with-long-path1" |
| 41 |
and |
| 42 |
"/cluster/disk25/user/prefix-short" |
| 43 |
|
| 44 |
those would be exchangeable. Of course chpathtool is possibly still not |
| 45 |
perfect, and i observed the one or the other segfault with certain files |
| 46 |
(none of those lately though). However you're free to fix it ;) |
| 47 |
|
| 48 |
Cheers, Markus |
| 49 |
|
| 50 |
> |
| 51 |
> It seems I'm hoping for the best of both worlds, which I might not be |
| 52 |
> able to get (or, not without a lot of work, and I guess I should at |
| 53 |
> least gain much more Gentoo experience first). |
| 54 |
> |
| 55 |
> Dag Sverre |
| 56 |
> |
Archives