[Top] [All Lists]

Re: xfsprogs/xfsdump release process

To: Felix Blyakher <felixb@xxxxxxx>
Subject: Re: xfsprogs/xfsdump release process
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Mon, 23 Feb 2009 13:29:55 -0600
Cc: Mike Frysinger <vapier@xxxxxxxxxx>, Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs-oss <xfs@xxxxxxxxxxx>
In-reply-to: <FF190F30-2ECF-49D7-B63F-EF65C076EC7A@xxxxxxx>
References: <200902221348.51905.vapier@xxxxxxxxxx> <20090223072207.GA4112@xxxxxxxxxxxxx> <DCA2D97E-F9A4-4A9E-BC39-CFFF1D0AD2CD@xxxxxxx> <200902231125.44076.vapier@xxxxxxxxxx> <FF190F30-2ECF-49D7-B63F-EF65C076EC7A@xxxxxxx>
User-agent: Thunderbird (Macintosh/20081209)
Felix Blyakher wrote:
> On Feb 23, 2009, at 10:25 AM, Mike Frysinger wrote:
>> On Monday 23 February 2009 11:10:44 Felix Blyakher wrote:
>>> On Feb 23, 2009, at 1:22 AM, Christoph Hellwig wrote:
>>>> On Sun, Feb 22, 2009 at 11:00:45PM -0600, Felix Blyakher wrote:
>>>>> Mike, you're absolutely right here. The config files got into the
>>>>> dmapi tarball by mistake, sorry for this. I verified that neither
>>>>> xfsprogs nor xfsdump have any extra files. New (clean) dmapi  
>>>>> tarball
>>>>> will be on the oss site shortly.
>>>> Please don't overwrite already uploaded tarballs.  Let's make a  
>>>> 2.2.10
>>>> release instead.  We might just use the Makepkgs script for it after
>>>> fixing it.  See the question and patch on the list for xfsprogs for
>>>> that, haven't checked dmapi works right yet.
>>> Didn't think that removing the unneeded files from the package
>>> justifies the version bump. It doesn't change anything for people
>>> who already downloaded unclean dmapi tarballs.
>>> Though, if opinion on this matter is that strong, I'd definitely
>>> bump the version, and will use updated Makepkgs.
>> the problem is for people (like Gentoo) who already fetched the  
>> tarball,
>> hashed it, and posted the resulting URL/hash to their packaging  
>> systems ...
> OK, that's definitely convincing.

Just FWIW, Fedora does similar.  Changing md5sums on a released,
versioned tarball could set off all sorts of worries for distributions,
in general...


<Prev in Thread] Current Thread [Next in Thread>