Closed 2colours closed 1 year ago
@vrurg Could this be related to your recent build system changes WRT RakuAST?
@patrickbkr I don't know what and how triple
does to answer your question. I'd wonder why something happens inside nqp-configure
in first place.
BTW, the most likely cause could be renaming of master
to main
in nqp-configure
repo. But normally it is checked out by hash by Configure.pl
and this change shouldn't affect builds.
I don't know what has changed exactly but it did work now.
If any of you know what could have addressed the problem, the issue could be closed.
Could it be that it (reproducibly) fails on the first run, but succeeds on the second one?
I'm not set up to do "first runs" repeatedly. :) But it rather seems to me that this might be fixed. It succeeded on a different system as well, and when I reported the failure, pretty sure I've tried to run it 2-3 times.
I'm inclined to close this issue as non-reproducible. Things were very much in motion around those days; unless something similar happens to someone else, it's probably just noise.
rakubrew rakubrew-version
?:rakubrew v32 Build type: fatpack OS: linux
env
or inshim
mode? (runrakubrew mode
to find out):env
rakubrew home
to find out):$HOME/.rakubrew
While building Rakudo with
triple
, after successfully building NQP, amake
system call is made inside~/.rakubrew/versions/moar-HEAD-HEAD-HEAD/3rdparty/nqp-configure/lib/NQP
where there is no Makefile or anything. This terminates the whole installation process.I know that storage can be a bottleneck on this machine but I still have like 700MB on the home mount (out of 5GB) so I suppose that shouldn't be the problem in this case.
Sidenote: the whole reason I needed to try this is that
rakubrew build
downright fails with "not a git directory" out of the blue. Something is really messed up with how things are put into place and I definitely have seen this working on the very same setup (built the RakuAST branch a couple of months ago).