Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 22 Nov 2015 10:41:23 +0000
From:      bugzilla-noreply@freebsd.org
To:        freebsd-ports-bugs@FreeBSD.org
Subject:   [Bug 204512] [NEW PORT] devel/bazel: Correct, reproducible, and fast builds for everyone
Message-ID:  <bug-204512-13-8fQ4t6jm0M@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-204512-13@https.bugs.freebsd.org/bugzilla/>
References:  <bug-204512-13@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=204512

--- Comment #7 from Peter Jeremy <peterj@FreeBSD.org> ---
I've done some more poking around and compared the proposed port with
bazel-0.1.1-installer-linux-x86_64.sh.  The latter installs a smaller bazel but
separately installs a large collection of files in
/usr/local/lib/bazel/base_workspace/ and appends that path to --package_path. 
I suspect this is the correct approach.

I have verified that adding ${WRKSRC} to --package_path make bazel work.

Having looker closer at the Makefile, I notice a couple of issues that should
be looked into:
- There's a test suite in bazel and it would be useful if that was hooked into
"make test"
- ${REINPLACE_CMD} includes the '-i' option.  Also '-i' takes an option so the
Makefile winds up using '-e' as a suffix for backup files.
- do-build deletes files in ${HOME}.  This is highly undesirable - the Makefile
should override ${HOME} so that the port does not alter any files outside
${WRKDIR}
- There are many occurrences of /bin/bash and md5sum that are not corrected. 
This definitely causes test failures but it's not clear whether it would impact
normal bazel operation.

-- 
You are receiving this mail because:
You are the assignee for the bug.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-204512-13-8fQ4t6jm0M>