© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
a60fb26
(Get the code!)
I was thinking about whether we should be using a variable like this a bit more, and I've had a chat with some of the developers about this. I'm pasting below a snipped version of the log.
The conclusion that was come to is that we should avoid using the point release variable in the documentation. In general, we should use 8.04 for the release, and we should seek to avoid using a string with a point release in it where possible.
The problem with using the point release variable is that each time a point release is issued, we would have to issue an updated package with the variable amended, and to fix all the translations that referred to the previous string (because translations don't use the variable in the same way as the English version).
On that basis, I think the best fix to this bug is to remove the hardcoded url, and use something like this:
"Download the JeOS iso file from the <ulink url="http://
(That folder will always redirects to the appropriate point release).
----
10:37:44 < mdke> When a point release is issued, does the proper name for a particular release include the point version? So is hardy called 8.04.2 or 8.04?
10:38:12 < mdke> the latter still, right?
10:38:39 < infinity> mdke: Depends on which proper name you're looking at.
10:39:03 < infinity> mdke: Check "lsb_release -a" on a hardy system, for instance.
10:39:11 < infinity> Description: Ubuntu 8.04.2
10:39:11 < infinity> Release: 8.04
10:39:25 < infinity> mdke: The former is the official name/version, which includes the .2
10:39:41 < infinity> mdke: But the release version doesn't increment for the sake of sanity of any tools that might check against that number.
10:40:09 < Amaranth> 8.04.2 sounds like it should have been released 2 April 2008 :P
10:40:20 < mdke> I see that http://
10:40:40 < mdke> what I'm looking at is whether, and if so when, in the documentation we should update version numbers
10:41:01 < infinity> mdke: In the docs? Never, IMO.
10:41:09 < mdke> in some cases it's essential, because links to isos are broken (the iso has .2 in the name)
10:41:19 < infinity> mdke: Any links to download stuff end up symlinked with 8.04->8.04.$current anyway.
10:41:41 < mdke> http://
10:41:44 < infinity> mdke: And do docs actually link directly to ISOs, rather than pages listing said ISOs?
10:42:14 < mdke> in one place, I think, bug 371773
10:42:16 < ubottu> Launchpad bug 371773 in ubuntu-docs "broken iso link on https:/
10:43:07 < cjwatson> the directory is symlinked but the ISO basename isn't
10:43:14 < infinity> mdke: Yeah, I'd call that a bug in the way the docs are written, IMO...
10:43:29 < mdke> perhaps linking to the directory is a better solution, even if it requires the user to do a bit more work
10:43:59 < cjwatson> anyway, in general I think of 8.04 as the "series" and 8.04.2 as identifying a specific point
10:44:00 < mdke> doing a SRU for the serverguide each time a point release is released is a bit of a pain, especially because it requires translators to redo that particular string too
10:45:03 < mdke> or at least running sed over the translations
10:45:55 < mdke> infinity: so you'd recommend simply pointing to the directory and telling the user to wget the relevant file from the directory?
10:46:52 < infinity> mdke: As much as that's less hand-holdy, I'm seriously failing to see how someone who's expected to understand a CLI-based server with almost nothing installed (JeOS is very minimal) wouldn't be able to sort out an http directory listing.
10:47:14 < mdke> yes, I don't think there is a risk that someone could get that wrong
10:47:22 < infinity> Oh, there's always a risk.
10:47:29 < mdke> heh
10:47:38 < infinity> But, I'm in the "let them keep both halves" camp when it's not a desktop/simple use case.
10:48:27 < mdke> ok. And it seems clear that generally in other cases, 8.04 should be used
10:48:49 < mdke> as the generic name for the distribution
10:49:12 < infinity> I see no reason to document 8.04.2 as a version number anywhere, except to name images, and in base-files, no.
10:49:19 -!- mok0 [n=mok@
10:49:45 < infinity> It would be about as silly as the Apache section in the server-guide talking about apache_
10:50:04 < infinity> (Well, not quite, but you get the absurdity reduction there, I guess)
10:50:06 < mdke> not quite that silly, surely, but I get the point
10:50:13 < mdke> :)
10:50:39 < mdke> thanks for the assistance