1 |
On Mon, Oct 1, 2018 at 10:14 PM Marty E. Plummer <hanetzer@×××××××××.com> wrote: |
2 |
> |
3 |
> On Mon, Oct 01, 2018 at 10:08:03AM -0400, Mike Gilbert wrote: |
4 |
> > On Mon, Oct 1, 2018 at 9:17 AM James Le Cuirot <chewi@g.o> wrote: |
5 |
> > > |
6 |
> > > On Mon, 1 Oct 2018 09:00:38 -0400 |
7 |
> > > Mike Gilbert <floppym@g.o> wrote: |
8 |
> > > |
9 |
> > > > > The reason why it is in DEPEND though is that none of these tools |
10 |
> > > > > are required at runtime. They are needed at postinst and postrm |
11 |
> > > > > stages which afaik makes them DEPEND on EAPI previous to EAPI 7 and |
12 |
> > > > > BDEPEND in EAPI 7 if I'm not mistaken. |
13 |
> > > > |
14 |
> > > > This is incorrect; programs executed in pkg_postinst/pkg_postrm should |
15 |
> > > > be present in RDEPEND. Consider a binpkg installation, which would |
16 |
> > > > ignore DEPEND, but would still execute pkg_postinst. |
17 |
> > > > |
18 |
> > > > Another option: since xdg-utils.eclass functions will politely skip |
19 |
> > > > the updates if the tools are missing, we could just drop the |
20 |
> > > > dependency from xdg.eclass entirely and wait for the tools to get |
21 |
> > > > installed through some indirect means. |
22 |
> > > |
23 |
> > > Unfortunately there is no correct answer here as we didn't consider |
24 |
> > > this scenario when planning EAPI 7. We should have looked at Exherbo, |
25 |
> > > which does make this important distinction. There is a bug report open. |
26 |
> > > |
27 |
> > > https://bugs.gentoo.org/660306 |
28 |
> > > |
29 |
> > > What I will say is that we should not attempt to executing anything |
30 |
> > > from within ROOT or SYSROOT unless it known to be interpreted |
31 |
> > > (e.g. a shell script) and there is no feasible alternative. I add that |
32 |
> > > exclusion clause because I have been forced to do this with |
33 |
> > > python-config. |
34 |
> > |
35 |
> > That's a good point; both update-desktop-database and |
36 |
> > update-mime-database are compiled binaries, and should probably be |
37 |
> > executed from BROOT (/) instead. I'll revise my patchset. |
38 |
> > |
39 |
> Not to steal the show/etc, but I do have a patchset which does most of |
40 |
> this already across {gnome2,xdg}{,-utils}.eclass, though I'm as |
41 |
> conflicted as you guys are with regards to the DEPEND/PDEPEND/BDEPEND/RDPEND |
42 |
> for it as well. |
43 |
> https://archives.gentoo.org/gentoo-dev/message/e0167edbb7b48b03dc90bfbb250ae16d |
44 |
|
45 |
Thanks for pointing this out, I missed this patchset. Can you rebase |
46 |
on current master and add Signed-off-by? |
47 |
|
48 |
Or maybe start with a github pull request and we can work out the |
49 |
kinks there. Whichever you prefer. |
50 |
|
51 |
Note that I pushed my patches for xdg-utils.eclass already, and just |
52 |
sent another patch to drop EROOT from the binary paths. It would |
53 |
probably be best to drop EROOT from the gtk-update-icon-cache path as |
54 |
well. |