Re: when emacs 22.1 release will ready?

["Lennart Borgman (gmail)" <lennart.borgman@gmail.com>]

Thomas Hühn wrote:
> "Lennart Borgman (gmail)" <lennart.borgman@gmail.com> writes:
> 
>> What do they do to keep control over the quality? I guess they are
>> using a lot of unit testing for the kernel, or?
> 
> I suppose there are *a lot* more people trying out pre-release kernels
> than CVS Emacs or pretest versions.
> 
> Another point why this pretest thing over the last weeks might have been
> bad:
> 
> People don't take "this is our pretest tarball, please check it for
> severe errors, this is our chance to find showstoppers before the
> release" serious if there are a lot more of "this time it's for real"
> pretest tarballs after that with no release in sight. That means that
> pretest tarballs reach even less potential testers.
> 
> That's more psychology than anything else, but it might count
> nevertheless.


I agree to both your points here. It looks however to me that it is a 
mix of slightly bad circumstances that are adding together. It might be 
tempting to look for a single cause, but it is perhaps more fruitful to 
consider what the addition of a lot of small things may cause.

There is a lack of resources here and that is a crucial thing that 
interacts badly with other weak points. Every developer and tester is 
valuable then.

Normal testing is known to demand a lot of resources. Bad implementation 
of normal testing is by some writers said to demand much more resources 
than trying to go along without any testing framework.

Unit testing promoters says that unit tests instead quickly is freeing 
resources. If that is true for a project like Emacs is of course not 
self evident, but I think it can be.