#280 fedora-review doesn't seem to know who owns `/usr/share/glib-2.0/`
Closed: not-a-bug Opened by terrycloth.

I'm currently doing a package review for the GNOME Shell extension Netspeed. I feel like fedora-review should already know that the directories /usr/share/glib-2.0/ and its subdirectory /usr/share/glib-2.0/schemas/ belongs to glib --- I haven't run into this issue for my own 5 different extensions recently accepted in the Fedora repos. But for Netspeed, fedora-review complains that no package owns the directory. Specifically, review.txt says this:

[ ]: Package must own all directories that it creates.
     Note: Directories without known owners: /usr/share/glib-2.0/schemas,
     /usr/share/glib-2.0

Netspeed's spec file list these dependencies:

BuildRequires:  gettext
BuildRequires:  gnome-common
Requires:       gnome-shell-extension-common >= %{min_gs_version}

In the %install section of the spec, it installs the glib-schema thus:

mkdir -p %{buildroot}%{_datadir}/glib-2.0/schemas
install -Dp -m 0644 schemas/org.gnome.shell.extensions.netspeed.gschema.xml \
    %{buildroot}%{_datadir}/glib-2.0/schemas/

The spec includes the glib-schema in the %files section:

%{_datadir}/glib-2.0/schemas/org.gnome.shell.extensions.netspeed.gschema.xml

And fedora-review includes this file in its generated files.dir:

/usr/share/glib-2.0/schemas/org.gnome.shell.extensions.netspeed.gschema.xml

I'm still fairly new to Fedora packaging, but I think the Netspeed RPM spec is doing everything right here. Is this a bug in fedora-review, or am I missing something?


Note the text here: fedora-review does not know the owner. The text means exactly that, not that the library actually lacks an owner. The purpose of this test is to pinpoint directories which needs manual inspection, not to establish a "no owner" fact.

Bottom line here is that f-r cuts some corners to avoid an even worse performance. There has been attempts to make a complete ownership test, but it just takes too long time.

Thanks for reporting! That said, I'm closing this as not-a-bug.

Metadata Update from @leamas:
- Issue close_status updated to: not-a-bug

Metadata