1 |
On śro, 2017-04-12 at 01:57 +0300, Mart Raudsepp wrote: |
2 |
> Ühel kenal päeval, T, 11.04.2017 kell 22:10, kirjutas Michał Górny: |
3 |
> > --- |
4 |
> > glsa.dtd | 4 +++- |
5 |
> > 1 file changed, 3 insertions(+), 1 deletion(-) |
6 |
> > |
7 |
> > diff --git a/glsa.dtd b/glsa.dtd |
8 |
> > index 52be18e..22237b0 100644 |
9 |
> > --- a/glsa.dtd |
10 |
> > +++ b/glsa.dtd |
11 |
> > @@ -124,7 +124,8 @@ |
12 |
> > Description: Version of the vulnerable package. Can be a range |
13 |
> > too |
14 |
> |
15 |
> Maybe some slot attribute documentation could be added as well, similar |
16 |
> to how the original patch attached to https://bugs.gentoo.org/106677 |
17 |
> did? |
18 |
|
19 |
I'd rather not put more effort into the DTD than absolutely necessary. |
20 |
I'm halfway porting it to XML Schema. Plus, a lot of the documentation |
21 |
in the DTD is plain outdated. In fact, the patch in question has been |
22 |
written against a different DTD file, more likely a better documented |
23 |
version that was never committed. |
24 |
|
25 |
> Though not the "*" stuff, that feels wrong and should be just |
26 |
> omitted attribute for that. |
27 |
|
28 |
Not convinced about that. '*' is already used explicitly in arch='', so |
29 |
I don't see a reason to not allow the same syntax here. This is also |
30 |
what the Paludis implementation assumes, and it fits into slot operators |
31 |
nicely. |
32 |
|
33 |
Maybe I should just change the default from #IMPLIED to "*"? (i.e. make |
34 |
omitted attribute equivalent to "*" DTD-wise) |
35 |
|
36 |
-- |
37 |
Best regards, |
38 |
Michał Górny |