On Tue, 12 Nov 2019 at 17:53, Matthew Flatt <mfl...@cs.utah.edu> wrote:
> At Thu, 7 Nov 2019 23:14:52 +0000, "'Reuben Thomas' via Racket Users"
> wrote:
> > I have a CONFIG_SITE environment variable which contains the following
> line:
> >
> > test "$prefix" = NONE && prefix="$HOME_LOCAL"
> >
> > (HOME_LOCAL is set to $HOME/.local)
> >
> > Things then go wrong because racket/src/start/configure.ac code relies
> on
> > prefix still being set to NONE in order to finally set prefix to `cd
> > "${srcdfir}/.." && pwd`.
> >
> > Is there some reason that prefix cannot be passed in from the top-level
> > Makefile for an in-place build?
>
> I think it would be ok to add `--prefix=NONE` to `CONFIG_IN_PLACE_ARGS`
> in the top-level makefile (alongside `--enable-origtree`). Does that
> solve the problem?
>
Thanks for the suggestion! That doesn't work, because when configure is
run, it reads CONFIG_SITE, loads my config file that checks `$prefix =
NONE` (it is!), which thus overrides it.
Adding `--prefix=`cd ../.. && pwd` to `CONFIG_IN_PLACE_ARGS` works for
setting `prefix`, but of course configure will then fail to set
`inplacebuild`, because to do so it tests `${prefix} = "NONE"`. Is there
some reason configure can't test `enable_origtree`? Is there some time the
build system is supposed to build in-tree even without --enable-origtree?
If so, could it test `"${prefix}" = "NONE" -o "${enable_origtree}" = "yes"`?
--
https://rrt.sc3d.org
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/racket-users/CAOnWdojw%2BpNP%2B%2BSjHqz5gbZwLuZcj7G97OG6zOG%3D168hnepe6g%40mail.gmail.com.
