Formal Complaint about off-list development discussions.

Jeremy Utley jeremy at
Thu Jul 1 10:40:08 PDT 2004

Jeroen Coumans said:
> Tushar Teredesai said the following on 01-07-2004 17:13:
>> The problem is that testing is not a different branch. At a particular
>> point in time, the unstable will be branched into testing and all the
>> "undiscussed changes" that went into unstable will be part of testing.
>> This concern was also brought up during the heavy discussion on hotplug.
> Yeah, the current development model is still broken. This is how it
> should be:
> * stable: released, tagged and well-tested LFS book.
> * testing: relatively stable. Merges relatively stable changes from
> unstable. Development branch for stable release.
> * unstable: developers-only playground, bleeding-edge, may not built,
> may not be coherent, may include CVS packages, (anything, actually).
> New LFS users use stable, regulars use testing and developers (and
> bleeding-edgers) use unstable. Currently, we have no testing branch,
> thus unstable fullfils that function. Which means that the current model
> is exactly the same as the previous (CVS/release) model, with all its
> problems.

And this is because the community, as a whole, wanted to wait for GCC
3.4.1 prior to cutting a new testing branch.  So, until 3.4.1 is released,
have no choice but to have no testing branch.


More information about the lfs-dev mailing list