| commit | 225fcf4477e33bc2fb50191e63905a32c63cca8d | [log] [tgz] |
|---|---|---|
| author | David Ostrovsky <david@ostrovsky.org> | Tue Sep 15 20:21:11 2015 +0200 |
| committer | David Ostrovsky <david.ostrovsky@gmail.com> | Sat Sep 19 05:44:44 2015 +0000 |
| tree | 65465ef533f76699dd21a60161f8085003c344a9 | |
| parent | 091de60661ff024cbde000e1253f8480d12ff01c [diff] |
Update buck to the latest version This version includes a lot of new features and improvements. The highlights are: * Switched to top-down-building, which should generally make builds faster. The old behavior can be bypassed by passing --deep on the command line: [1] * New query command was added (inspired by Bazel): [2] * Performance improvement in file globbing Extend tools/eclipse.py to pass the --deep option to ask Buck to execute bottom-up build when generating the Eclipse project. This is needed otherwise after using the buck clean command only the top rule artifact would be fetched from the cache and the buck-out/gen/lib folder would remain empty. To allow two genrules, located in the same BUCK file, to have the same output artifact name, genrule output is namespaced with genrule name: [3]. That why switching to new Buck version requires buck clean. Otherwise the build would fail: [4]. [1] https://github.com/facebook/buck/commit/217cec33bcd0b0eb7ebd0c444938495379267d60 [2] https://buckbuild.com/command/query.html [3] https://github.com/facebook/buck/commit/c92ef212b53fff08a8452649b4d4faadc6b89b11 [4] http://paste.openstack.org/show/468992 Change-Id: If336ea86970dd95e3af13b92e7665ecaef369105