[arch-general] How stable are the new version number formats on eg. filesystem, usbutils, etc.
All, Updating different minimum dependency package version info for tde PKGBUILDs, I note there have been a number of 'version number format' changes for various packages. E.g.: filesystem 0.x.y-z ==> 2013.05-2 usbutils 0.x.y-z ==> 006-1 There is a big difference going from filesystem>=0.7.3 to filesystem>=2013. When I run across packages like this where the version has changed format -- Are they likely going to stay with the new format? Or will they likely revert back to major.minor-rel numbers at some time? -- David C. Rankin, J.D.,P.E.
On Tue, Jan 14, 2014 at 6:39 PM, David C. Rankin <drankinatty@suddenlinkmail.com> wrote:
All,
Updating different minimum dependency package version info for tde PKGBUILDs, I note there have been a number of 'version number format' changes for various packages. E.g.:
filesystem 0.x.y-z ==> 2013.05-2 usbutils 0.x.y-z ==> 006-1
There is a big difference going from filesystem>=0.7.3 to filesystem>=2013. When I run across packages like this where the version has changed format -- Are they likely going to stay with the new format? Or will they likely revert back to major.minor-rel numbers at some time?
-- David C. Rankin, J.D.,P.E.
This is handled with the `epoch` part of the version now, so it's a non-issue. Arch generally doesn't make use of minimum version requirements at all because partial upgrades are not supported.
On 15 January 2014 07:39, David C. Rankin <drankinatty@suddenlinkmail.com> wrote:
All,
Updating different minimum dependency package version info for tde PKGBUILDs, I note there have been a number of 'version number format' changes for various packages. E.g.:
filesystem 0.x.y-z ==> 2013.05-2
This has been the format since at least April 2008. [1]
usbutils 0.x.y-z ==> 006-1
And this since January 2011. [2]
There is a big difference going from filesystem>=0.7.3 to filesystem>=2013. When I run across packages like this where the version has changed format -- Are they likely going to stay with the new format? Or will they likely revert back to major.minor-rel numbers at some time?
There is always a reason for the change, enforced by upstream or necessitated by a move to a different release source (e.g. VCS). I can't recall filesystem's history, but it might have been due to the need for a version format that makes sense for such a package. And the versioned dep you are citing is probably a leftover that got updated. Anyway, as Daniel already said, changes like this shouldn't bother you now that epoch is used to force upgrades in the event vercmp returns negative. [1] http://goo.gl/AA6Xn9 [2] http://goo.gl/5Dm5k3 -- GPG/PGP ID: C0711BF1
participants (3)
-
Daniel Micay
-
David C. Rankin
-
Rashif Ray Rahman