[gentoo-dev] Re: EAPI usage

[Duncan]

Rich Freeman posted on Thu, 30 Aug 2012 20:38:11 -0400 as excerpted:

> My main concern is doing bumps all the time just for their own sake.

Yes.  That's why I didn't tackle that side at all.  But I've seen the 
"PM's can never drop support for an EAPI once adopted" thing before, and 
while there's quite a possibility I'm missing something as I'm no PM 
expert, it does seem to me that rings hollow; that an EAPI drop COULD be 
done, without too much disrupting either users or devs (PM or otherwise).

But as the experts say otherwise, there probably /is/ something I'm 
missing, which is why I asked.

Meanwhile, I'll definitely allow that there's often a big chasm between 
"possible" and "worth the trouble", too, and it's quite within the realm 
of reason that it's simply "not worth the trouble" at this point, even if 
very much possible, and even likely worth the trouble once we get upto 
say 10 EAPIs or some such.

Meanwhile(2), I (cautiously) support the idea I've seen before of 
deprecating and gradually removing at least EAPI-1, and probably 2 and 3 
as well over time.  People /have/ pointed out that core system packages, 
toolchain, etc, may well need to stay at EAPI-0 virtually "forever".  
That's the exception I mentioned with EAPI-0 thus being an exception as 
well, thus the focus on 1-3.  But once 1-3 are out of the tree for a 
sufficient period, I really /don't/ see why the method I described can't 
be used to drop their support from the PMs, as well, and expect that 
regardless of whether it's worth tackling as a project starting today, at 
some point, it'll be worth doing.

OTOH, I can see someone, possibly concerned about the historical 
implications (so "gentoo historians" at least, can try long deprecated 
ebuilds and see how they work), might wish to maintain support for every 
EAPI "forever".  But I don't believe it should be mandatory, and in 
practice, I'd venture that due to simple code rot once there haven't been 
any packages of a particular EAPI in the tree or in wide circulation for 
awhile, even if support /does/ officially continue, it'll likely be 
broken if anyone tries to use it, say five years or a decade later.  Once 
that starts being a major concern, why /not/ just dump it.  The 
historians can go find an old stage tarball with an old PM that supported 
the EAPIs they're interested in, if it comes to that.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman