<div class="gmail_extra"><div class="gmail_quote">On Tue, Dec 4, 2012 at 10:21 PM, Magnus Therning <span dir="ltr"><<a href="mailto:magnus@therning.org" target="_blank">magnus@therning.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
> This are my suggestions:<br><div><div class="h5">
><br>
> 1. Magnus keep a [haskell-base] repository. This MUST be an Arch repo<br>
> in sync with a github repo.<br>
> 2. Both repositories (Arch and git) should be available for<br>
> maintainers of other repository.<br>
> 3. Maintainer of [haskell-web] (me) MUST update it's repo to keep in<br>
> sync with base in a reasonable time.<br>
> 4. After this time a “global maintainer” (Magnus, Ramana, me, whoever)<br>
> can grab all packages from both repositories and put them in a new<br>
> one: [haskell]<br>
> 5. Only [haskell] is intended for end users.<br>
><br>
> A “reasonable time” could be 3 days. I saw that Magnus updates the<br>
> repo on Wednesday and on Saturday/Sunday: so before the next update<br>
> [haskell-web] should be in sync.<br>
><br>
> I developed [haskell-web] in a way that it will not duplicate packages<br>
> from [haskell]: I use them as DistroPkg. Updating is easy with<br>
> `cbladmin` [^1], there's no need to modify `cblrepo`, maybe you could<br>
> merge (and develop) some ideas from there. But this is not the main<br>
> point. The main point is to be in sync, or else I'm just wasting my<br>
> time, as [haskell-web] will never be really useful.<br>
<br>
</div></div>The big piece that's missing above is how to handle failures to<br>
upgrade a package caused by a dependant not allowing the upgrade. Do<br>
we hold off the base-test repo then?<br>
I currently upgrade everything I can in one transaction, what if one<br>
of the packages requires holding back due to a dependant. Would that<br>
require a partial roll-back in the base-test repo?<br></blockquote><div><br>I don't understand the issue here. The only requirement is that the [haskell-base] Arch repo and github repo are always in sync with each other.<br>
<br>I don't think a change to haskell-base necessarily needs to be only version updates, there might be rollbacks if they are necessary. But whatever changes are made to haskell-base, eventually haskell-web and other haskell- repos will be made to work with it (in a reasonable time), and then when a good state is reached, [haskell] itself can be updated. So it's always in a good state. It might mean waiting longer for updates to make their way from hackage to [haskell].<br>
<br><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
It's questions like these that I don't have a good answer to. And<br>
yes, both situations have occurred in the past. They both cause a bit<br>
of work (though not so much if they are caught already at 'clbrepo<br>
updates' time).<br>
<div class="im"><br>
/M<br>
<br>
--<br>
Magnus Therning OpenPGP: 0xAB4DFBA4<br>
email: <a href="mailto:magnus@therning.org">magnus@therning.org</a> jabber: <a href="mailto:magnus@therning.org">magnus@therning.org</a><br>
twitter: magthe <a href="http://therning.org/magnus" target="_blank">http://therning.org/magnus</a><br>
<br>
</div>I invented the term Object-Oriented, and I can tell you I did not have<br>
C++ in mind.<br>
-- Alan Kay<br>
</blockquote></div><br></div>