[tex-live] TeXLive has no stable source tree and resorts to DVD with binaries?

[Joseph Wright]

On 14/04/2011 09:09, Kārlis Repsons wrote:
>> Why should the well tested and stable package be
>> declared unstable in TL? And for how long?
> Sounds like no reason. Admittedly I perceived the actual deveIopment workflow 
> wrong (ah, forgot CTAN), but basically the idea is: the package authors and a 
> few users might reach a point, where they don't doubt stability question and 
> wish to put in CTAN, but the real problem is within interaction of two or more 
> packages, which would take more testers to find it... (but the immediate 
> inclusion encourages everyone to stay at where things are then after it!)
> Perhaps I just make some kind of theory here plus it's true that maintenance 
> of all-stable file trees is a real work... On the other hand, struggling with 
> more package conflicts or incompatible changes can too take time. It just then 
> might go off more quietly than what it takes to provide that inter-package 
> stabilization (is what I meant with "all-stable").
> (I hope I explained and not too long...)

You seem to think that the TeX Live people do bug fixes for stuff from
CTAN: this is not true. The reality is that the only way a (La)TeX
package gets used is if it gets released. (Remember that there is MiKTeX
as well as TeX Live, so it's not even just a TeX Live issue.) The
workflow for a package author is simple:

 - Write package
 - Upload to CTAN
 - TeX Live/MiKTeX pick up package
 - It gets distributed
 - Bugs appear
 - You (hopefully) fix them
 - Upload to CTAN and loop

> I read what Norbert stated already, the reason I still post this is to get 
> some opinions about that idea in general. Perhaps some Linux distros or Mac 
> even have their own (separate!) bugfixed versions of TeX software just for that 
> inter-package thing?

No, because that's not where things get fixed: it's down to package
authors via CTAN.
-- 
Joseph Wright