Continuing the discussion about GRASS 6.4.4 in #757308 on debian-gis@.
I'm currently working on addressing the many lintian issue affecting the
package. I don't think we should upload a new grass revision until the
majority of these issues have been fixed.
On 08/20/2014 07:53 AM, Hamish wrote:
> Hamish wrote:
>>> grass 6.4.4.packaging is currently (basically) ready in DebianGIS
>> But not by using git-import-orig. The upstream branch hasn't been
>> updated with the grass_6.4.4.orig.tar.gz contents.
> it was done in the master branch,
We use the git-buildpackage toolkit for packages maintained in git, and
that workflow requires the upstream sources in the upstream branch.
Which are then merged into the master branch where the debian directory
also lives. Just invoking `git-import-orig --pristine-tar
--uscan|<path-to-orig.tar.gz>` would have done that automatically.
I know that you're a fan of Subversion, but please read up on the git
workflow that the rest of the team uses.
> tagging is currently incomplete, but known todo and discussed privately.
It would be nice to have this on the list so the rest of the team is
aware of the ongoing issues too. Can you elaborate on the incomplete
>> You should run lintian showing all tags after your build to get an
>> idea of what is left to fix before I would consider an upload.
> er, you really think we don't do that and know exactly what they are?
> We've been busy packaging this software for more than a dozen years!
> Not masking false positives and "real but won't-fix in this major
> version" lintian issues is not a bug IMHO.
To be honest, I don't think that you know what the lintian issues with
the grass package are. They've been unaddressed for years, which shows a
lack of maintenance for the package.
Since you're not listed in the Uploaders of the package, I don't think
you consider yourself a maintainer of the package and just an upstream
developer. And that's fine, that's where your strength lies.
We know that frankie is a busy person and wasn't able to keep up with
all the work in the Debian GIS team, that's why I'm trying to improve
and grow the team to address that.
There has been some great team work recently, and it would be great to
get that around the grass package too. It's the historic roots of the
team after all.
> No doubt the main rules file could use a bit of a refresh here and there
> (supporting 'make -j' would be nice), but let's continue this
> conversation on the DebainGIS list, not in a ticket about transitioning
> to wxpython3.0.
It's not only the missing targets that need to be added to the rules
file for policy compliance. It also needs to stop using
hardening-includes to fix the lintian error it produces:
E: grass source: build-depends-on-obsolete-package build-depends:
hardening-includes => use dpkg-buildflags instead
Properly using the hardening flags would fix 494 lintian issues (mostly
missing use of relro).
Are you aware of the shell/tcl dance in animate.tcl that makes lintian sad?
E: grass-gui: shell-script-fails-syntax-check
Shouldn't this TCL script also exec wish like the others? That line is
currently commented out, and from the looks of it a bug that should be
Another thing that could be fixed upstream, or just in the packaging is
moving all the images the usr/share/grass64 and symlink them to
usr/lib/grass64 for FHS compliance. This change would fix 483 lintian
There are also issues with the copyright file, that can be fixed at the
same time as rewriting it using copyright-format 1.0. Something upstream
can address here is to use a proper copyright statement that includes
the years in the COPYING file.
Running `licensecheck --copyright -r` on the package source reports a
lot of files without copyright statements, this is ideally fixed
The many man page issue appear to be caused by html2man which like
Doxygen is not very good at converting to proper man pages. Fixing the
man page issues would get rid of another 134 lintian issues.
GPG Key ID: 4096R/E88D4AF1
Fingerprint: 8182 DE41 7056 408D 6146 50D1 6750 F10A E88D 4AF1