[Bioperl-l] [Gmod-gbrowse] xyplot data alignment problem?

[Chris Fields]

On Dec 18, 2006, at 10:15 AM, Sendu Bala wrote:

> Hilmar Lapp wrote:
>>
>> On Dec 16, 2006, at 9:37 AM, Sendu Bala wrote:
>>
>>> Lincoln Stein wrote:
>>>> This is very embarassing for me, particularly since I spent a lot
>>>> of time validating that Bio::Graphics was working properly before
>>>> the 1.5.2 release went out. How long before there is a 1.5.3
>>>> release? How about a 1.5.2.1release?
>>>
>>> I'm happy to try a point release for critical bug fixes. Why don't
>>> you commit the necessary fixes to branch-1-5-2 and let me know when
>>> you're happy, and I'll do 1.5.2.1.
>>
>> Feel free to do that, but why not make a 1.5.3 off the main trunk?
>> 1.5.2.1 may be adding more to the version confusion
>> (developer/stable/point-release/etc) than it is worth,
>
> My feeling is that 1.5.3 should be reserved for some significant  
> changes
> and new features, and not just a few bug fixes. I'd say this causes  
> less
> confusion amongst users - they can associate '1.5.2' with a certain  
> API
> and feature-set, and the specific name of the file they download and
> install (bioperl-1.5.2_100.tar.gz vs bioperl-1.5.2_101.tar.gz) won't
> matter at all to them.
>
> I also won't have to make some major announcement about it; it will
> simply be the most recent developer version of bioperl available so  
> new
> users trying to get 1.5.2 will end up getting 1.5.2.1, whilst existing
> 1.5.2 users will only feel compelled to get it if they suffer from the
> bugs fixed.
>
>
>> and there is no shame in releasing new developer versions every few
>> weeks.
>
> I think doing frequent releases are inadvisable; such a quick release
> won't have had much testing so we shouldn't encourage people to  
> install
> it: encouragement is implicit when a major new version comes out like
> 1.5.3. People who want to live on the edge can and should be using a
> CVS checkout.

I thought that 1.5.2 was considered a point release for the 1.5 dev  
series, for bug fixes along with the potential for added/experimental  
features.  Similarly, 1.6.x releases would be point releases for bug  
fixes only with all tests passing (no added features since it is a  
stable release series).  I guess one could reason that 1.5.x releases  
have both bug fixes and new features, while 1.5.x.y releases are  
simply bug fixes for the 1.5.x branch (no new features).  We probably  
should add something to the FAQ and maybe make a few changes to the  
1.5.2 wiki page.

I think having a 1.5.2.1 release is feasible as a quick one-off to  
get Lincoln's fixes in, since you would make them off the 1.5.2  
branch anyway (so I guess it could be considered a bug release from  
that branch).  It's probably not something we should make a habit of,  
but then again I'm not the Pumpkin!

chris