Stamatis, Francis, thank you very much for the feedback. I will work on that.
Ruben On Tue, Oct 6, 2020 at 11:03 PM Francis Chuang <[email protected]> wrote: > Hi Ruben, > > Thanks for getting the released rolled out. > > In general, when making changes to the website: > - Master is the source of truth. > - Commit to Master first. > - Cherry pick into Site. > - Build Site and publish. > > If a commit is on site, but not on master, then the person making the > change made a small mistake and you should make sure the commit is > cherry-picked into master. > > After a release, we need to make Site equal Master since Master is the > source of truth. > > I believe you run "git reset --hard master" on Site to force it to equal > master. Note that this completely overwrites site with all the commits > from Master and drops commits that were in Site but not in Master. > Therefore it is important to make sure any commits that were on Site but > not on Master are "fixed" by making sure they are on Master first. > > Once that's done, just publish the site following the instructions in > site/README.md > > Hope that helps! > > Francis > > On 7/10/2020 7:08 am, Ruben Q L wrote: > > Hi all, > > > > I have to publish the site after the release 1.26.0, it is the first time > > that I do such a thing, and I don't want to mess things up (especially > > since I am not a git expert). > > > > I do not understand the current situation between 'master' and 'site' > > branches. It looks like they have diverged: 'site' branch is currently "2 > > commits ahead, 99 commits behind master" [1], is this normal or expected? > > The ahead commits seem to be: > > - > > > https://github.com/apache/calcite/commit/8cf0701dac5aad7d695709d8e35957c261f8ae82 > > - > > > https://github.com/apache/calcite/commit/b65944772c7234612fc5a8c84384d40a5cfa6d84 > > > > The second one is an interesting case, it seems that the exact same > change > > was also committed to master (as a different commit): > > - > > > https://github.com/apache/calcite/commit/0920796bb917ee8420d2534233486ca0413b4430 > > I guess this was a cherry-pick? Is this normal? In that case, why the > other > > one was not applied to master? > > > > My question is, how should I proceed? Shall I merge master into site? > site > > into master? both? I have tried both locally, and they cannot be > > fast-forwarded (merge --ff-only). > > > > Then, which branch should I use to re-build the site and re-generate the > > javadoc before pushing it into the calcite-site repo? Master? Site? > Either > > of them once they are aligned? > > > > Thanks, > > Ruben > > > > [1] https://github.com/apache/calcite/tree/site > > >
