[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#906663: vcswatch: should not flag an UNRELEASED changelog entry as "ready for upload"



On Sun, Aug 19, 2018 at 08:48:02PM +0200, Christoph Berg wrote:
 Re: Marc Haber 2018-08-19 <[🔎] 153468174721.19286.15758469883614485339.reportbug@fan.zugschlus.de>
> > I could work in a branch, but that would mean additional rules for team
> > maintained packages which I'd rather avoid. I mean, there are valid
> > reasons, for example an upload held back due to an ongoing library
> > transition, for not immediately uploading any new commit to a master
> > branch in git.
> 
> Why are you putting that .0 version there in the first place? Not
> doing that would fix the issue.

I'm trying to avoid

Case 1: Somebody working on the branch, committing things and generating
new changelog entries for a version (x-y) that is already in the archive and
should therefore not be changed in the vcs

Case 2: Somebody working on the branch with a new version (x-y+1),
deploying test versions of the package that will not be overwritten by
later versions of the package (with the same version number) from the
archive, with the side effect that different but identically numbered
versions of the packages fly around. Additionally, the tracker would
begin to remind uploading those versions.

Both cases have already happened in my personal past and caused a good
portion of confusion and grief.

The x-y.0 version nicely solves both issues: it's strictly later (and
different!) than any uploaded version, and strictly earlier (and
different!) than the next uploaded version.

Greetings
Marc

-- 
-----------------------------------------------------------------------------
Marc Haber         | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany    |  lose things."    Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421


Reply to: