1 |
On Sun, 2020-02-09 at 22:51 -0800, Zac Medico wrote: |
2 |
> On 2/9/20 10:44 PM, Michał Górny wrote: |
3 |
> > On Sun, 2020-02-09 at 22:30 -0800, Zac Medico wrote: |
4 |
> > > Hi all (especially package owners in CC), |
5 |
> > > |
6 |
> > > In various packages we have inconsistent use of || preferences for |
7 |
> > > www-client/elinks, links, lynx, w3m, and w3mmee. This means that the |
8 |
> > > default preference depends on the combination of packages that one has |
9 |
> > > installed and the order that one has installed them, leading to |
10 |
> > > unpredictable results. |
11 |
> > > |
12 |
> > > Here is a list of relevant packages and their dependencies: |
13 |
> > > |
14 |
> > > app-text/docbook-sgml-utils: || ( www-client/lynx www-client/links |
15 |
> > > www-client/elinks virtual/w3m ) |
16 |
> > > app-text/sgmltools-lite: || ( www-client/w3m www-client/lynx ) |
17 |
> > > app-text/xmlto: || ( virtual/w3m www-client/lynx www-client/elinks ) |
18 |
> > > dev-lang/mono: || ( www-client/links www-client/lynx ) |
19 |
> > > mail-client/mutt: || ( www-client/lynx www-client/w3m www-client/elinks ) |
20 |
> > > mail-client/neomutt: || ( www-client/lynx www-client/w3m www-client/elinks ) |
21 |
> > > net-irc/irssi: || ( www-client/lynx www-client/elinks ) |
22 |
> > > sys-fs/gt5: || ( www-client/links www-client/elinks www-client/lynx ) |
23 |
> > > x11-base/xorg-server: || ( www-client/links www-client/lynx www-client/w3m ) |
24 |
> > > |
25 |
> > > How about if we create some more virtuals to cover all of the relevant |
26 |
> > > cases? |
27 |
> > |
28 |
> > I don't think that's a valid case for a virtual since those tools do not |
29 |
> > provide a consistent API for other packages. It just happens that some |
30 |
> > packages explicitly support multiple choices, and this is exactly what |
31 |
> > > > indicates. |
32 |
> > |
33 |
> > The virtuals would really be arbitrary here. Developers would |
34 |
> > repeatedly fail to use them because they wouldn't naturally expect |
35 |
> > the virtual to exist. |
36 |
> |
37 |
> In that case, I suppose we'll have to apply consistency manually? Can we |
38 |
> all agree on a global order of preference for the relevant packages? |
39 |
|
40 |
That would be my idea, yes. I'd suggest going for the 'lightest' |
41 |
package first. Would you be able to figure out some kind of measure |
42 |
on how heavy each of those packages is? I suppose we need to account |
43 |
for build time and dependencies. |
44 |
|
45 |
-- |
46 |
Best regards, |
47 |
Michał Górny |