Bug#930587: Re: Bug#930587: release-notes: fails to build for jessie and stretch

[Wenbin Lv <wenbin816@gmail.com>]

Hi Jover,


You are right. I was choosing between release-notes and www.debian.org

but I was not aware that the .dbk files will be regenerated anyway

during the build process. It seems that this should be reassigned

to www.debian.org.


The build script is at

https://salsa.debian.org/webmaster-team/cron/raw/master/parts/7release-notes

and it indeed uses git checkout to build for different releases.


Best regards,

Wenbin Lv


On Sun, 16 Jun 2019 23:01:14 +0200 Guillem Jover <guillem@hadrons.org> wrote:

> Hi!
>
> On Sun, 2019-06-16 at 14:40:53 +0800, Wenbin Lv wrote:
> > Package: release-notes
> > Severity: serious
> > Tags: ftbfs
> > Justification: fails to build from source (but built successfully in the past)>
> > Recent transition from .dbk to .po for the ca language leaves some
> > untracked .dbk files in the source directory and causes git checkout of
> > jessie and stretch to fail in the build system of the Debian website,
> > showing the following errors[1]:
> >
> > rebuilding the release notes for {jessie,stretch}
> > error: The following untracked working tree files would be overwritten by checkout:> > ca/about.dbk
> > ca/installing.dbk
> > ca/issues.dbk
> > ca/moreinfo.dbk
> > ca/old-stuff.dbk
> > ca/release-notes.dbk
> > ca/upgrading.dbk
> > ca/whats-new.dbk
>
> This looks like a bug in the website build machinery? These files are
> left behind in the same way they are for all other languages. The
> difference is that they got switched very recently.
>
> I'm assuming from your description that the website machinery, does a
> git checkout on each specific branch and then builds it, but is not
> running «make clean» before each git checkout?
>
> > which causes buster version of the release notes to be built for jessie
> > and stretch[2][3].
>
> Right!
>
> > Please remove the untracked files to fix the issue.
>
> I don't think there is nothing to fix in the release-notes package
> itself. This sould probably be reassigned.
>
> Thanks,
> Guillem
>
>