<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">2016-06-30 17:59 GMT+02:00 luca pedrielli <span dir="ltr"><<a href="mailto:luca.pedrielli@gmail.com" target="_blank">luca.pedrielli@gmail.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000"><span class="">
<div>Il 30/06/2016 17:04, Jean-Claude Vanier
ha scritto:<br>
</div>
<blockquote type="cite">
<pre>Two questions:
-- currently we have 4 repositories; this number will increase for
each new release, according to the current model;
the proposed changes will keep 4 repositories forever, as there
will not be "new" release. So, at start, the benefit, in terms of
space required, is not that important.
Is that correct?
-- imagine we are, let's say ten years in the future:
"Legacy" (=2013) is still the same?
"LTS" evolves "slowly" and still features kde4?
"Stable" features "plasmic13" (the evolution of kde project in ten years)
"Cooker" is still cooking
I guess there will be a kind of shift and old releases will have to
be abandoned and we should adopt, from the beginning, a rule for that.</pre>
</blockquote>
<br></span>
+1<br>
<br>
The proposal of TPG is interesting.<br>
<br>
<span lang="en"><span>I
do not find</span> <span>great difficulty</span> <span>for the</span>
<span>ordinary work</span> <span>of everyday</span><span>.</span><br>
<span>but what happens</span> <span>in the case of</span> <span>big</span>
<span>technical news</span> <span>?:</span><br>
<br>
<span>kde4-</span><span>></span> <span>plasma,</span> <span>python2-</span><span>></span>
<span>python3</span><span>,</span> <span>gcc-</span><span>></span>
<span>clang</span> <span>are</span> <span>only</span> <span>some</span>
<span>news</span> present <span>in</span> <span>3.0<br>
<br>
</span><span>how can</span> <span>news</span> <span>like these</span> <span>"</span><span>rolling"</span>
<span>in stable</span> <span>easily</span><span>?</span></span><span class=""><br>
<br>
<br></span></div></blockquote><div><br></div><div>This is a bow for QA team :)<br><br></div><div>Till QA tests won't say that new feature won't break system, then it can not be published. Idea is to have as much automatic tests as it is possbile. Ofcourse this means effort but in time perspective there is chance to have thing working with less human input.<br><br></div><div>For example:<br><br></div><div>LTS with python2 is about to be updated with python3. This means python3 can be pushed for LTS repository same time when all dependent packages to python3 were builded for LTS, and without problems can replace old python2 packages and related rpms.<br></div><div> </div></div></div></div>