Am Donnerstag 06 November 2008 15:53:53 schrieb Sergiu Ivanov:
> > And since ebuilds are very easy to maintain, it's far easier to keep a
> > system
> > current with them.
>
> Is there any advantage over Debian binary packages when using ebuilds?
file to the version of the application.
But that also means, every user has to compile the applications himself (I
like that, but it takes some processing time).
I see. It seems to me that this necessity won't be very repellent, so
it doesn't look like a problem. Especially for Gentoo folks ;-)
> Note that I have no special preferences with debs and I'm not sitting
> on a Debian system. What I pursue with this question is whether a
> Gentoo GNU/Hurd would be easier to maintain up-to-date with usual
> Gentoo repositories and whether it would be possible to avoid
> situations like we are in at the moment: some packages are broken and
> a lot of stuff does not work (emacs, for example).
If we get maintainers for the Hurd stuff, then yes.
But for that we'd still need people who maintain them, and thought it's far
less work than doing a deb package (just rename the file), the update still
has to be tested, so someone has to build it which takes about the same time
as making the deb package.
So I assume the main advantage is the high geek concentration in the Gentoo
community :)
OK, this sounds pretty interesting :-) As soon as there are more
geeks, there might be more (possible) developers and more bug reports,
which is good :-)
> BTW, I once tried to build emacs from source on Hurd and the attempt
> failed. I used to think that if I have the source code of a program
> for Linux, not using some kernel interfaces, I could easily build it
> on the Hurd, but it proved to be false...
The emacs ebuild is quite complex, so I can well imagine that building it
isn't that simple... (I attached the ebuild, maybe it can give you a look on
the difficulty inherent to emacs - the ebuild is one with useflags ("I want
that feature, but don't want this one") and custom patches).