Date: Sun, 18 Feb 2024 12:20:32 -0800 From: Mark Millard <marklmi@yahoo.com> To: Miroslav Lachman <000.fbsd@quip.cz> Cc: FreeBSD Mailing List <freebsd-ports@freebsd.org> Subject: Re: FreeBSD ports community is broken [port building configuration notes] Message-ID: <A4AA312D-F160-402F-853B-E59800EB4FB0@yahoo.com> In-Reply-To: <f6e730bf-3d4e-4318-b49a-7100f38fc3ed@quip.cz> References: <87B38D6C-1D83-4158-B03B-F4C8EA396DD1.ref@yahoo.com> <87B38D6C-1D83-4158-B03B-F4C8EA396DD1@yahoo.com> <f6e730bf-3d4e-4318-b49a-7100f38fc3ed@quip.cz>
next in thread | previous in thread | raw e-mail | index | archive | help
On Feb 18, 2024, at 11:34, Miroslav Lachman <000.fbsd@quip.cz> wrote:
> On 18/02/2024 17:52, Mark Millard wrote:
>> Aryeh Friedman <aryehfriedman_at_gmail.com> wrote on
>> Date: Sun, 18 Feb 2024 10:37:06 UTC :
>>> It should not require
>>> prodiere running on a supermassive machine to work (in many cases
>>> portmaster and make install recursion fail where prodiere works).
>> As for configuring for small, slow systems relative to
>> resource use, I provide some settings that I've
>> historically used below. Then I have some other notes
>> after that material.
>=20
> It is not just about resources required by Poudriere but also the fact =
that Poudriere almost always rebuilds too many packages almost without a =
reason... "just to be sure". You want to build update for one small =
package ended up waiting half a day for rebuild of rust, cmake, llvm...
I know.
I responded just to both of:
A) "It should not require prodiere running on a supermassive
machine to work"
and:
B) "in many cases portmaster and make install recursion
fail where prodiere works"
You choose not to quote any later material relevant to
either and to instead write about an aspect not about
the "port building configuration notes" that I wrote
about. I did not write about a grand solution to all
the tradeoffs that are now involved across different
ways of building/installing ports/packages.
I fully agree that poudriere's systematic behavior
rebuilds more than the likes of portmaster [but fails
less often].
As stands there are tradeoffs between use of portmaster
(and the like) vs. use of poudriere (/synth?). No one
has produced an alternative that avoids the tradeoffs
as far as I know. So one picks between the tradeoffs
by the choice of which way to build.
Claims that one of the two general types of approaches
should just go away so that only the other type is
used are not viable or reasonable in my view.
For me, I've used both and settled on systematic use
of poudriere, explicitly accepting the tradeoffs
involved. Others go the other way, as I used to.
My expectation is that the maintainer/commiter burden
part of the tradeoffs will always lead to portmaster
and the like having more build failures where poudriere
works. Which is part of what what I was trying to
identiify for (B). Until there is an implemented
alternative that does not have the tradeoff structure(s),
I expect the tradeoff choices will continue to need to
be made.
I was not trying to vent about the frustrations of the
various types of tradeoffs involved.
=3D=3D=3D
Mark Millard
marklmi at yahoo.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A4AA312D-F160-402F-853B-E59800EB4FB0>