Repo Surgery

[Ulrich Spörlein]

Am Mo., 25. Mai 2020 um 11:57 Uhr schrieb Konstantin Belousov <
kostikbel at gmail.com>:

> On Mon, May 25, 2020 at 11:28:59AM +0200, Ulrich Spörlein wrote:
> > Am Mo., 25. Mai 2020 um 03:00 Uhr schrieb Warner Losh <imp at bsdimp.com>:
> >
> > > https://svnweb.freebsd.org/base?view=revision&revision=29297
> > >
> > > Removed tetris from the Makefile. Then a subsequent commit removed the
> game
> > > to the attic. Then Peter deleted the files.
> > >
> > > So, sources before this revision are unbuildable if you don't build
> them
> > > NOGAMES.
> > >
> > > What do you think we should do about it?
> > >
> > > My gut tells me this is a 'The project removed tetris at the request
> of the
> > > holders of the IP to tetris completely, including the CVS history. The
> cvs
> > > -> svn conversion also omitted it. The git reconstruction also omits
> it,
> > > which means the tree is unbuildable before hash XXXXX without NOGAMES
> or a
> > > hack to src/games/Makefile.'
> > >
> > > We also did this with autofs, and maybe others.... So this is not
> unique.
> > >
> > >
> > To be clear, does that mean if (!) we bring back 1.0, that we'd also need
> > to blacklist tetris accordingly?
> >
> > And regarding buildability: does anything that old even still build on a
> > modern compiler? I.e., how far back do builds still work?
> Build works if you extract jail/chroot e.g. with 1.1.5 and build there.
> We do support a.out even on amd64.
>
> Last time I did it on 4-core machine, world built in less then 1 minute.
> It was relatively long time ago.
>

Oh, that's good to know. I didn't consider the "VM" route.

btw, the code is still there in the archive ISOs on the ftp, so the "tacked
on" release ISO images would bring this back, currently.