geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan D. Cabrera" <>
Subject Re: Old tags in Geornimo
Date Mon, 31 Oct 2005 22:18:17 GMT
Rodent of Unusual Size wrote, On 10/31/2005 1:32 PM:

>Hash: SHA1
>Alan D. Cabrera wrote:
>>Tags imply that they are supported.
>Really?  I've never encountered that before.  Maybe against a
>V1_0_0 tage, but then you just tell someone 'nope, not supported.'
>M1 means 'milestone one,' right?
That would not be a friendly way to go.  I am arguing that we remove the 

>Another thing to consider is people who build against the
>idiosyncrasies of a particular tagged version.  Making the
>tag harder to find/check out does them no service.  This
>sort of thing was (is?) seen a lot with the Apache APR
>code.  Things like Subversion and the Apache httpd code
>are released with specific versions of APR expected.  So
>what if they're obsolete?  Make the tag harder to find
>and it's harder to use.
I would use this as an argument  for removing milestone tags  as quickly 
as possible.  People should not be building against a milestone tag, 
only a released verison.  Maybe if it was important for a particular 
project to have a particular snapshot we would cut a supported tag, and 
most likely a branch if we really liked them, but, imho, using milestone 
tags is not the way to go.

>One situation in which I could see this being relevant is
>a company wanting to incorporate code.  They might want to
>pick a particular point in time in order to do their
>'pedigree' checking of the IP issues.
Using a milestone tag would not be the way to go here.  Subversion 
version numbers work quite nicely here.

>So it's not just the meaning to Geronimo that's relevant
>here, but the meaning of the tags to consumers/users of
>the code.
Agreed.  I am arguing that *milestone* tags are not the way to support 
the above scenarios.

>If the only reason to do anything with the old tags is
>the concern that someone might consider them as implying
>support, I'd suggest finding a way to negate that impression
>that doesn't involve affecting the history or ease of
>finding it.
To what end will someone dig up, say, M3?  A supported tag e.g. v1_0_0 
or v1_0_5 I can see.


View raw message