Just to confirm. Please reply +1 if you're okay for me to commit the revert / re-commit of METRON-646 (PR#441).
Thanks, Kyle On Mon, Feb 27, 2017 at 9:25 PM, Casey Stella <[email protected]> wrote: > Yeah, I'd think there does not exist a time in which the email should be > null. I'd rather just error out if you can't find it and the committer > doesn't put one in. > > I do agree that the most sensible way to pull the commit name is to pull it > from the repo's commit history. Here's a 1-liner to use if you dont' feel > like coming up with it yourself > > git clone https://github.com/kylerichardson/incubator-metron.git --depth=1 > --branch METRON-646 --single-branch METRON-646 >& /dev/null && cd > METRON-646 && (git log | grep Author | awk -F: '{print $2}' | sed 's/^ > //g') && cd .. > > > > On Mon, Feb 27, 2017 at 9:15 PM, Nick Allen <[email protected]> wrote: > > > Sure. It could validate the email address before letting you proceed. > > > > It tries to get the email from the author's Github profile. If the > author > > doesn't make one public, it will come back as 'null' and prompt you to > > change it. Of course, it will just use 'null' if you don't provide an > > alternative. > > > > Most of the time I have to enter the email address manually because not > > many people make their email public. Even better would be to pull the > email > > from the author's own commits in the PR. That would reduce how often we > > have to manually input an email. > > > > > > On Mon, Feb 27, 2017 at 8:53 PM, Casey Stella <[email protected]> > wrote: > > > > > Nick, what are your thoughts on adjusting the script to error out or > > prompt > > > for an email address if one can't be found? > > > > > > On Mon, Feb 27, 2017 at 8:51 PM, Nick Allen <[email protected]> > wrote: > > > > > > > I think revert and commit again is the best way to go. Not a big > deal. > > > > > > > > On Mon, Feb 27, 2017 at 6:55 PM, Casey Stella <[email protected]> > > > wrote: > > > > > > > > > I think it should be changed, but I'm not sure how to change it. I > > > think > > > > it > > > > > should be changed because our git history is our legal trail of > > > > > attribution. Mucking with it is relatively serious business. > > > > > > > > > > As to how, normally I'd say git commit --amend --author > > > "kylerichardson < > > > > > [email protected]>" if we act before the next commit and a > > git > > > > > rebase otherwise, but it's pushed and rewriting history for a > push'd > > > > commit > > > > > has consequences. Not the least of which the scary force'd push. > > The > > > > > challenge here is that all forked repos during this period between > > the > > > > > wrong commit and the correction commit will be based on a dead > > > branch. I > > > > > guess I would vote for 1, the revert and then the re-commit. > > > > > > > > > > I'd like to understand a bit more about how this happened. Ryan, > can > > > you > > > > > walk it through how you did the commit so we can avoid it in the > > > future? > > > > > > > > > > Casey > > > > > > > > > > > > > > > On Mon, Feb 27, 2017 at 4:04 PM, Kyle Richardson < > > > > > [email protected]> > > > > > wrote: > > > > > > > > > > > Ok, so here's the story... Ryan was nice enough to commit my > recent > > > PR > > > > > and > > > > > > for whatever reason my github username but not my email address > > > appears > > > > > in > > > > > > the commit author (see below). > > > > > > > > > > > > commit 41fc0ddc9881d9cfdd8bae129c0bb7800a116d4c > > > > > > Author: kylerichardson <null> > > > > > > Date: Mon Feb 27 11:38:55 2017 -0600 > > > > > > > > > > > > METRON-646 Add index templates to metron-docker > (kylerichardson > > > via > > > > > > merrimanr) closes apache/incubator-metron#441 > > > > > > > > > > > > My question is can it be left as is or does it need to include > the > > > > email > > > > > > address per apache? > > > > > > > > > > > > If it needs to be changed, what are the acceptable options? > > > > > > > > > > > > (1) commit a revert and re-commit; maintains a record of > everything > > > > > > (2) rebase one back, update, and force a push; like it never > > happened > > > > > > (3) another option I haven't considered? > > > > > > > > > > > > -Kyle > > > > > > > > > > > > > > > > > > > > >
