Re: the nature of pkgsrc support for OS versions (release vs stable)

[Greg Troxel <gdt%lexort.com@localhost>]

Manuel Bouyer <bouyer%antioche.eu.org@localhost> writes:

>> 2) This MAKE_JOBS=1 forcing is not understandable since it is already
>> MAKE_JOBS_SAFE=no on NetBSD.  Please file a PR about this.  I would like
>> to have a clear understanding on whether this MAKE_JOBS=1 is needed for
>> 9.0_RELEASE or also 9.0_STABLE.   (I wonder if the MAKE_JOBS_SAFE needs
>> to be someplace else in the Makefile.)
>
> I fear I can't fill a usefull PR for this. we can't run such tests on pkgbuild,
> and the test box where I could check for this has "only" 3Gb ram, so
> rust won't build with MAKE_JOBS > 1 anyway.

I am not asking you to debug this.  You are the one observing the
problem so I am merely asking you to create a PR saying what happened in
your environment, and describing your workaround.  It may well be that
it never gets fixed, but I am really uncomfortable with having
workarounds for problems that are not in the PR database.  So a PR is
useful, even if you don't see it as leading to a fix.

(I also don't follow why, if you need to have MAKE_JOBS=1 on that box
because of RAM (fair eoough), that you needed to have this workaround
since it would have been avoided.)

>> 1) move pkgbuild to STABLE (and document this as how it should be)
>
> no, this is not "how it should be". We can't afford to update the chroot every
> now and then. admins needs to be asked to update the chroot when needed.

Ideally the bug would have been fixed before the release :-)

What I meant is that given a critical bug, bulk machines need to be
updated to past the bug.  I didn't mean to suggest that they be updated
every week.