On Wed, Dec 2, 2020 at 11:16 AM Andrew Purtell <apurt...@apache.org> wrote:
> Like this? > > https://gist.github.com/apurtell/f5959f36b0b13d5e92f35b22549020cc > > > Looks good to me. 2.3.0 CHANGELOG doesn't display because it is too big -- it is an accumulation of all changes up to 2.3 (good IMO) whereas 2.2+2.1, etc. are just the changes in those releases versus the previous version -- but you can view the 'raw version' of the 2.3 doc which is good enough I think. S > On Wed, Dec 2, 2020 at 10:04 AM Nick Dimiduk <ndimi...@apache.org> wrote: > > > On Tue, Dec 1, 2020 at 12:16 PM Andrew Purtell <apurt...@apache.org> > > wrote: > > > >> I think by git tag would work. Given the URL can be different from the > >> visible anchor text we could use an exact git sha too, if there is > concern > >> that tags might be changed either accidentally or intentionally in a way > >> that breaks embedded links in old changelogs. > >> > > > > A link back to the CHANGES file in GitHub of a release tag, along with > the > > tag's sha at the time of link generation, would achieve my request that > the > > point to a rendered version of the changes file where possible. Thank > you! > > > > On Tue, Dec 1, 2020 at 11:57 AM Nick Dimiduk <ndimi...@apache.org> > wrote: > >> > >> > I am in favor of referring off to the changes files of older releases. > >> > Would that be by git tag, or to the files in the distribution > archives? > >> > > >> > I don’t think these changes files are for marketing as such. However, > I > >> > think they are intended to be human-readable (if not for humans, then > >> > who/what, and why?). Not having a rendered version easily discoverable > >> is a > >> > barrier to this goal, and shrinking the file such that it renders in > >> GitHub > >> > is a low-cost approach to achieve that goal. > >> > > >> > Thanks, > >> > Nick > >> > > >> > On Mon, Nov 30, 2020 at 20:04 Andrew Purtell < > andrew.purt...@gmail.com> > >> > wrote: > >> > > >> > > Unless there is an objection to the plan I described below, it will > >> > happen > >> > > tomorrow on branch-2 in prep for RC. > >> > > > >> > > > >> > > > On Nov 30, 2020, at 4:46 PM, Andrew Purtell <apurt...@apache.org> > >> > wrote: > >> > > > > >> > > > > >> > > > I'm glad I checked email before beginning the RC. > >> > > > > >> > > > How about this: > >> > > > > >> > > > CHANGES.md file that ships in 2.4.0 will contain URLs pointing to > >> older > >> > > CHANGES.md for 1.0.0, 2.0.0, 2.1.0, 2.2.0, 2.3.0. > >> > > > > >> > > > CHANGES.md file that ships with 2.4.0 will list all issues > completed > >> > for > >> > > 2.4.0 > >> > > > > >> > > > CHANGES.md file that ships with 2.4.1 will list all issues > completed > >> > for > >> > > 2.4.0 and 2.4.1. > >> > > > > >> > > > etc. until 2.5.0, at which point the CHANGES.md file that ships in > >> > 2.5.0 > >> > > will contain URLs pointing to older CHANGES.md for 1.0.0, 2.0.0, > >> 2.1.0, > >> > > 2.2.0, 2.3.0, and 2.4.0, and will list all issues completed for > 2.5.0. > >> > > > > >> > > > I have felt traditionally the changes file is not where we do > >> release > >> > > upgrade marketing. > >> > > > > >> > > > If the objective is giving user-friendly and self-service answers > >> to an > >> > > operator or developer asking, "why should I upgrade? / what's new in > >> this > >> > > release?", then I humbly submit we should bring back the practice of > >> > > writing blog posts for blogs.apache.org/hbase. Speaking of which, > >> that > >> > > blog is in a somewhat sad state of disrepair with a lot of broken > >> image > >> > > links. > >> > > > > >> > > > > >> > > >> On Mon, Nov 30, 2020 at 1:02 PM Stack <st...@duboce.net> wrote: > >> > > >> On Mon, Nov 30, 2020 at 12:34 PM Nick Dimiduk < > ndimi...@apache.org > >> > > >> > > wrote: > >> > > >> > >> > > >> > So concretely, the conclusion here is that the CHANGES.md file > >> that > >> > > ships > >> > > >> > in 2.4.0 should contain entries for 2.0.0, 2.1.0, 2.2.0, 2.3.0, > >> and > >> > > 2.4.0? > >> > > >> > The CHANGES.md file that ships in 2.4.1 will contain all of the > >> > > above, plus > >> > > >> > entries for 2.4.1. > >> > > >> > > >> > > >> > > >> > > >> And the 1.0.0 changes. > >> > > >> > >> > > >> Point at a 1.0.0 CHANGES.md file rather than list the 1.0.0 > >> changes. > >> > > Ditto > >> > > >> for 2.0.0 changes. Could do pointer for older minor releases too > if > >> > too > >> > > >> many items to list... 2.1 and maybe 2.2. > >> > > >> > >> > > >> S > >> > > >> > >> > > >> > >> > > >> > >> > > >> > Are you sure that's what you want? That seems like more than we > >> > need. > >> > > >> > > >> > > >> > Thanks, > >> > > >> > Nick > >> > > >> > > >> > > >> > On Tue, Nov 10, 2020 at 5:47 PM 张铎(Duo Zhang) < > >> > palomino...@gmail.com> > >> > > >> > wrote: > >> > > >> > > >> > > >> > > +1 on what Sean proposed to include the changes started from > >> the > >> > > first > >> > > >> > > major release. > >> > > >> > > > >> > > >> > > Sean Busbey <bus...@apache.org> 于2020年11月10日周二 下午7:37写道： > >> > > >> > > > >> > > >> > > > I thought we had written up a guide before for what goes in > >> the > >> > > changes > >> > > >> > > > file, but I can't find it at the moment. > >> > > >> > > > > >> > > >> > > > For branch 2.3 I am surprised at 0.99 stuff. I would > expect: > >> > > >> > > > > >> > > >> > > > * 2.0.0 > >> > > >> > > > * 2.1.0 > >> > > >> > > > * 2.2.0 > >> > > >> > > > * 2.3.[0-z] > >> > > >> > > > > >> > > >> > > > Because that would be enough that if I was coming from the > >> prior > >> > > major > >> > > >> > > > release I could see everything that might matter getting to > >> the > >> > > >> > release. > >> > > >> > > > > >> > > >> > > > If we just include 2.3.z changes then I have to go look at > >> each > >> > > of the > >> > > >> > > > previous minor releases on the release line as well. > >> > > >> > > > > >> > > >> > > > We've talked for some time about possibly including release > >> > notes > >> > > / > >> > > >> > > changes > >> > > >> > > > for just those things in each individual release on the > >> website > >> > > before. > >> > > >> > > > Would adding something like that be sufficient for the use > >> > you're > >> > > >> > > thinking > >> > > >> > > > of? > >> > > >> > > > > >> > > >> > > > > >> > > >> > > > > >> > > >> > > > > >> > > >> > > > On Mon, Nov 9, 2020, 15:35 Nick Dimiduk < > ndimi...@apache.org > >> > > >> > > wrote: > >> > > >> > > > > >> > > >> > > > > Heya, > >> > > >> > > > > > >> > > >> > > > > The CHANGES.md file on branch-2.3 weighs in at over 1mb > >> and is > >> > > too > >> > > >> > big > >> > > >> > > > for > >> > > >> > > > > Github to render. Its content covers back to 0.99. This > >> isn't > >> > > really > >> > > >> > > > usable > >> > > >> > > > > by someone who wants to easily see what's new in the > latest > >> > > patch > >> > > >> > > > release. > >> > > >> > > > > > >> > > >> > > > > I propose we truncate these changes files to what's new > for > >> > the > >> > > >> > release > >> > > >> > > > > branch. It probably needs some more work, but the > git-jira > >> > audit > >> > > >> > script > >> > > >> > > > [0] > >> > > >> > > > > is able to generate a report of what's new (never > >> previously > >> > > >> > released) > >> > > >> > > > for > >> > > >> > > > > a target release-line branch. We could use this as the > >> basis > >> > > for the > >> > > >> > > > > CHANGES file when starting a new release-line branch. > From > >> > then > >> > > on, > >> > > >> > > Yetus > >> > > >> > > > > takes care of the patch release updates. > >> > > >> > > > > > >> > > >> > > > > What do you think? > >> > > >> > > > > > >> > > >> > > > > Thanks, > >> > > >> > > > > Nick > >> > > >> > > > > > >> > > >> > > > > [0]: > >> > > >> > > > > > >> > > >> > > > > > >> > > >> > > > > >> > > >> > > > >> > > >> > > >> > > > >> > > >> > https://github.com/apache/hbase/tree/master/dev-support/git-jira-release-audit > >> > > >> > > > > > >> > > >> > > > > >> > > >> > > > >> > > >> > > >> > > > > >> > > > > >> > > > -- > >> > > > Best regards, > >> > > > Andrew > >> > > > > >> > > > Words like orphans lost among the crosstalk, meaning torn from > >> truth's > >> > > decrepit hands > >> > > > - A23, Crosstalk > >> > > > >> > > >> > >> > >> -- > >> Best regards, > >> Andrew > >> > >> Words like orphans lost among the crosstalk, meaning torn from truth's > >> decrepit hands > >> - A23, Crosstalk > >> > > > > -- > Best regards, > Andrew > > Words like orphans lost among the crosstalk, meaning torn from truth's > decrepit hands > - A23, Crosstalk >
