ltib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Ltib] Release numbers on spec-files


From: Stuart Hughes
Subject: Re: [Ltib] Release numbers on spec-files
Date: Mon, 20 Jul 2009 09:42:20 +0100
User-agent: Thunderbird 2.0.0.16 (X11/20080707)

Hi Svein,

The policy is that for packages that are installed on the build machine (hostcf) any change requires an update to at least the release number. This is required as for host packages version/released update in used to gate the installation so as to make sure installing an old BSP won't downgrade a newer one's host components.

On the target side, ideally when spec files are update the release should be bumped up, but this is not mandatory. The reasons is that LTIB doesn't look at these numbers when deciding to install, just the fact that the spec file is newer than the target rpm.

I'm not sure what Freescale intend to do with respect to merge-outs, I used to handle this. If people want them (and I think they would be useful) it may help if people could get in contact Freescale and ask them to submit their updates to this forum (by way of patches).

Regards, Stuart

Svein Seldal wrote:
Hi all

Is there a policy when the release number is bumped up in a spec file? Correct me if I'm wrong, but isn't the idea of release numbers that if you change the spec file, adds patches, etc. then you should bump the release number? Because it _could_ (humble me) seem like there's a lot of updates where this isn't practiced...

I'm sitting here trying to merge two ltib trees; One Freescale BSP (based on an older unknown ltib version) and a LTIB CVS head, and I see a lot of changes (albeit most of them small) in spec files but often no change in release numbers. In this case it's actually difficult to learn the source of the changes: It may come from changes in the fsl branch or in the ltib CVS branch. (If I had the two branches common ancestor, I'd be able to separate the source of the change, but that's a different discussion.)







reply via email to

[Prev in Thread] Current Thread [Next in Thread]