1 |
On Tue, 01 Nov 2005 02:21:31 -0700 Duncan <1i5t5.duncan@×××.net> wrote: |
2 |
| I'd suggest we preemptively incorporate an ENewsVersion or similar |
3 |
| field as well, to be 1.0, with this proposal, but should the format |
4 |
| ever need changed, that would allow for a format version 1.1 or 2.0 |
5 |
| or whatever. As the GLEP suggests that future GLEPs may add new |
6 |
| headers, this would allow them to update this version field as |
7 |
| necessary, presumably using the familiar minor version = backward |
8 |
| compatible, major version if not backward compatible, versioning |
9 |
| scheme. |
10 |
|
11 |
Sounds reasonable. |
12 |
|
13 |
| One other niggle. The encoding is specified as UTF-8, but then the |
14 |
| format is specified as RFC-822 (7-bit ASCII, pre-i18n) compatible. |
15 |
| I'm not up on i18n specs, but presumably there's a newer RFC that |
16 |
| specifies how internationalized internet messages are handled in an |
17 |
| RFC-822 backward compatible way? If so, it may make more sense to |
18 |
| reference that specific RFC, which presumably deals with UTF-8 |
19 |
| instead of specifying 7-bit ASCII as does 822. Alternatively, |
20 |
| MIME/Quoted-printable could be specified, which would allow for |
21 |
| escaped 8-bit chars, as I'd /assume/ UTF-8 requires, given the -8. |
22 |
| (I /said/ I'm not up on that!) |
23 |
|
24 |
Hrm. The "RFC-822 like" phrase is used in various places (for example, |
25 |
the GLEP specification) to describe the structure of the header |
26 |
section. I don't think it precludes us also specifying UTF-8. |
27 |
|
28 |
-- |
29 |
Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron) |
30 |
Mail : ciaranm at gentoo.org |
31 |
Web : http://dev.gentoo.org/~ciaranm |