On Sun, Jul 22, 2018 at 09:43:04AM +0100, Stuart Henderson wrote: > On 2018/07/22 00:57, Marc Espie wrote: > > On Sat, Jul 21, 2018 at 10:58:38PM +0100, Stuart Henderson wrote: > > > I think it's actually a problem with aclocal and junking.. > > > > > > -- > > > Sent from a phone, apologies for poor formatting. > > > > > > On 21 July 2018 21:54:14 Marc Espie <[email protected]> wrote: > > > > > > >On Sat, Jul 21, 2018 at 07:51:27PM +0300, Kirill Bychkov wrote: > > > >>Hi! > > > >>I've deinstalled gobject-introspection and ws able to build bitcon > > > >>without > > > >>errors. and it's present in amd64 packages from Jul 19 > > > > > > > >1/That's not a fix. > > > >2/That's no guarantee things work. > > > > > > > >What sthen@ describes looks like what we call a "hidden dependency". > > > > > > > >The fact bulk worked this time doesn't mean the port is fine. > > > > > > ... which means there is something completely wrong with using > > autogen.sh without looking more closely into it. > > > > It might affect similar ports. > > >From a quick skim of aclocal code, it looks like it may be scanning all > the m4 files in /usr/local/share/aclocal and then processing the files, > so I think this may be due to a change in files present between scan and > processing. If so, I'd expect it to affect anything using aclocal (but > the race window must be small or we'd see it more often). > > naddy has mentioned having noticed a similar problem in other ports > at times. > > If that's the case, we'd either need to disable junking while running > aclocal (ugh), or change aclocal to cope with disappearing files.
One simple thing to do would be to redirect aclocal elsewhere and make a snapshot before it's run.
