Heads up: IcedTea6 1.4 is on the way

[David Herron]

Andrew Haley wrote:
> David Herron wrote:
>   
>> Matthias Klose wrote:
>>     
>>> Andrew Haley schrieb:
>>>  
>>>       
>>>> Andrew John Hughes wrote:
>>>>    
>>>>         
>>>>> 2009/1/28 Lillian Angel <langel at redhat.com>:
>>>>>      
>>>>>           
>>>>>> Mark Wielaard wrote:
>>>>>>        
>>>>>>             
>>>>>>> It would be nice to at least update the NEWS file with the actual
>>>>>>> release date (and hopefully all the new goodies that went in).
>>>>>>>           
>>>>>>>               
>>>> No, this is quite impossible.  A release tarball is made and the release
>>>> is tested.  Once that's done, the tarball is golden.  You do not go
>>>> fiddling
>>>> with files inside the tarball.
>>>>         
>> Or you could take a page from other kinds of QA organizations ...
>>
>> The FCS or GA tarball is the one that's undergone testing but prior to
>> that tarball could be one or more beta or RC tarballs depending on your
>> labeling preference.
>>     
>
> That doesn't sound any different.  Obviously the tarballs must have
> RC status until one of them passes the release criteria, at which
> point it becomes the release.
>
> What's the difference?
>
> Andrew.
>   
Maybe I missed something earlier in the conversation, if so please be 
patient.

You said "a release tarball is made" and "the tarball is golden".  That 
sounded like your process is simply to cut the tarball and that's it and 
it can't be changed.  Maybe you're saying there is some testing between 
those two steps that validates you can bless the specific tarball as golden.

I fully agree once you've blessed a tarball any change is risky and can 
cause breakage.  Even simple changes like unpacking and changing a text 
file.  I've seen that happen.

- David


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/distro-pkg-dev/attachments/20090128/2990b040/attachment.html