<div>On Thu, Mar 22, 2012 at 18:41, Daniele E. Domenichelli <span dir="ltr"><<a href="mailto:daniele.domenichelli@gmail.com">daniele.domenichelli@gmail.com</a>></span> wrote:</div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 22/03/12 03:39, Dario Freddi wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
This approach has also multiple advantages, such as:<br>
<br>
* We can decide which commit each submodule will point to, so it's<br>
safe ground for testers<br>
* Or we can allow people to fuck this and just get master of each repo<br>
* Qt already did the hard job for us and we just have to adapt our scripts<br>
* This approach preserves modularity but still allows to group quite<br>
efficiently some/all repos.<br>
</blockquote>
<br></div>
* If we find out that this approach sucks we can easily go back to the multiple repository approach<br>
* If one day some packages will enter in any kde "main" package we just need to import the add the submodule there and remove it from here.<br>
<br>
<br>
Since Dario is supporting me on git submodules I propose an alternative (but I still believe we should talk about this later, perhaps at Akademy):<br>
<br>
Instead of making several packages, we make a single "kde-telepathy" meta repository (yes kde-telepathy, not ktp :P). Inside we add some subdirectories (libs, handlers, config, utils, whatever, etc.) and inside each subdir we put the real repositories.<br>
</blockquote><div><br></div><div><div>Note that the packages are not the problem but just a result of it. The thing we're trying to solve is our repos count, which is high and this does not help it, on contrary it makes the complexity higher (depends on point of view).</div>
<div><br></div><div>--</div><div><font color="#666666">Martin Klapetek | KDE Developer</font></div></div></div>