On 14 January 2016 at 13:49, Sam Ruby <[email protected]> wrote: > On Thu, Jan 14, 2016 at 8:33 AM, sebb <[email protected]> wrote: >> On 14 January 2016 at 13:25, Sam Ruby <[email protected]> wrote: >>> On Thu, Jan 14, 2016 at 6:00 AM, sebb <[email protected]> wrote: >>>> On 14 January 2016 at 01:58, Sam Ruby <[email protected]> wrote: >>>>> On Wed, Jan 13, 2016 at 8:50 PM, sebb <[email protected]> wrote: >>>>>> On 14 January 2016 at 01:38, Sam Ruby <[email protected]> wrote: >>>>>>> Commit 5e97b142358e2aaa8957c0ae1608c79e8fd0bb5a: >>>>>>> return only one copy >>>>>>> >>>>>>> >>>>>>> Branch: refs/heads/master >>>>>>> Author: Sam Ruby <[email protected]> >>>>>>> Committer: Sam Ruby <[email protected]> >>>>>>> Pusher: rubys <[email protected]> >>>>>>> >>>>>>> ------------------------------------------------------------ >>>>>>> lib/whimsy/asf/committee.rb | + - >>>>>> >>>>>> This is a library change. >>>>>> >>>>>> Surely that should involve an update to asf.version? >>>>> >>>>> I'll update asf.version and upload a new gem, and deploy on >>>>> whimsy.apache.org once you are happy with the contents of >>>>> http://whimsy-test.apache.org/public/. >>>> >>>> I just did a comparison between whimsy & whimsy-test. >>>> The only diffs are the gem version and the spurious boardcommittees entry. >>>> So I'm happy with the output. >>>> >>>>> Meanwhile, we probably should remove the gem version info from the >>>>> output as we move away from the idea of this being gem based? >>>> >>>> It would be useful to identify the installed code version somehow. >>>> >>>> This would be trivial with SVN (svn info filename or $revision$). >>>> Not so easy with Git. >>>> >>>> One possibility is: >>>> >>>> $ git rev-list -n 1 HEAD >>>> 5e97b142358e2aaa8957c0ae1608c79e8fd0bb5a >>>> >>>> That seems to be a local operation, so should be fast. >>> >>> WFM. >>> >>>> Another idea is >>>> >>>> $ git rev-list --count HEAD >>>> 1080 >>>> >>>> However I suspect the count breaks down if any commits are squashed. >>>> >>>> Maybe combine the two? >>> >>> I don't see the need to combine the two. In particular, you can't >>> change the count without changing the git hash. Hashes are unique and >>> immutable. Doing a squash would change the hash. >> >> Yes, but the count makes it easier to see how recent the version is. >> Otherwise one needs to search the rev-list to find out. > > How about some or all of: > > git show --quiet HEAD
>>>> commit 9cc7ee8135b61f16f5b16272f653f87731c6fe5c Author: Sebb <[email protected]> Date: Thu Jan 14 13:27:45 2016 +0000 Moved to Github <<<< The date would help with staleness checks. However it may be Not sure the author or comment are useful, given that they only refer to part of the whole. So perhaps something like: $ git show -format="%h %cI" -s HEAD 9cc7ee8 2016-01-14T13:27:45+00:00 or $ git show --format="%H @ %cI" -s HEAD 9cc7ee8135b61f16f5b16272f653f87731c6fe5c @ 2016-01-14T13:27:45+00:00 >> But the count is not usable on its own unless squashing is disabled. > > Let's not go there. :-) > > - Sam Ruby > >>>> This would identify what has been deployed without needing to login. >>> >>> Cool. >>> >>> - Sam Ruby >>> >>>>>>> ------------------------------------------------------------ >>>>>>> 2 changes: 1 additions, 1 deletions. >>>>>>> ------------------------------------------------------------ >>>>>>> >>>>>>> >>>>>>> diff --git a/lib/whimsy/asf/committee.rb b/lib/whimsy/asf/committee.rb >>>>>>> index 45613af..6beb007 100644 >>>>>>> --- a/lib/whimsy/asf/committee.rb >>>>>>> +++ b/lib/whimsy/asf/committee.rb >>>>>>> @@ -101,7 +101,7 @@ def self.load_committee_info >>>>>>> end >>>>>>> end >>>>>>> >>>>>>> - @committee_info = list.values >>>>>>> + @committee_info = list.values.uniq >>>>>>> end >>>>>>> >>>>>>> def self.nonpmcs
