No subject

------------------<br>
<br>
The snapshot site<br>
<br>
=C2=A0 =C2=A0<a href=3D"http://www.cs.utah.edu/plt/snapshots/" target=3D"_b=
lank">http://www.cs.utah.edu/plt/snapshots/</a><br>
<br>
demonstrates how a lot is working.<br>
<br>
Here are the remaining implementation issues:<br>
<br>
=C2=A0* Generated distribution sites do not yet include a source code<br>
=C2=A0 =C2=A0option or &quot;collects.tgz&quot; for piggy-backing distribut=
ors, and the<br>
=C2=A0 =C2=A0makefile or configuration file lacks support for piggy-backing=
.<br>
<br>
=C2=A0 =C2=A0These seem straightforward to add.<br>
<br>
=C2=A0* The PLT-maintained packages are not yet reflected on<br>
=C2=A0 =C2=A0&quot;<a href=3D"http://pkg.racket-lang.org" target=3D"_blank"=
>pkg.racket-lang.org</a>&quot;.<br>
<br>
=C2=A0 =C2=A0Because all of those packages are currently in one big git<br>
=C2=A0 =C2=A0repository, it&#39;s not clear how to register the packages. G=
uesses #1<br>
=C2=A0 =C2=A0and #2 in &quot;Developing Racket like PLT developers&quot; ab=
ove are two<br>
=C2=A0 =C2=A0possible routes. Another is that we set up a process to pull f=
rom<br>
=C2=A0 =C2=A0git and bundle package sources into individual zip archive tha=
t are<br>
=C2=A0 =C2=A0registered on &quot;<a href=3D"http://pkg.racket-lang.org" tar=
get=3D"_blank">pkg.racket-lang.org</a>&quot;.<br>
<br>
=C2=A0* The `make installers&#39; support needs to be less tied to<br>
=C2=A0 =C2=A0&quot;main-distribution&quot;.<br>
<br>
=C2=A0 =C2=A0You can configure the set of packages that are built and inclu=
ded<br>
=C2=A0 =C2=A0in installers by `make installers&#39;, but that set currently=
 must be<br>
=C2=A0 =C2=A0be a subset of the packages in the &quot;pkgs&quot; directory =
of the Racket<br>
=C2=A0 =C2=A0repository. It&#39;s easy in principle to pull the packages fr=
om a<br>
=C2=A0 =C2=A0catalog server, but there will be some issues to sort out in t=
he<br>
=C2=A0 =C2=A0bootstrapping process and in ensuring a consistent snapshot.<b=
r>
<br>
=C2=A0* No support yet for generated distributions sites with binary<br>
=C2=A0 =C2=A0packages.<br>
<br>
=C2=A0 =C2=A0Probably not too difficult. I forget what went wrong last time=
 I<br>
=C2=A0 =C2=A0tried this, but a lot has been fixed since then. In any case, =
the<br>
=C2=A0 =C2=A0idea of binary packages does not seem to have gained much trac=
tion.<br>
<br>
=C2=A0* Package-dependency checking for tests.<br>
<br>
=C2=A0 =C2=A0Maybe it&#39;s just a matter of compiling tests sorting them i=
nto<br>
=C2=A0 =C2=A0suitable packages, like everything else, which is a direction =
that<br>
=C2=A0 =C2=A0we&#39;ve already started.<br>
<br>
=C2=A0* The &quot;main-distribution&quot; package needs to be cleaned up.<b=
r>
<br>
=C2=A0 =C2=A0The &quot;main-distribution&quot; package currently inclues te=
sts, and it<br>
=C2=A0 =C2=A0includes packages like &quot;honu&quot; that are not in the cu=
rrent release.<br>
=C2=A0 =C2=A0This clean-up task is related to sorting out tests.<br>
<br>
=C2=A0* Different builds modes are not yet configured with different<br>
=C2=A0 =C2=A0default package scopes.<br>
<br>
=C2=A0 =C2=A0Should be easy.<br>
<br>
I also have a long-ish list of minor repairs and usability<br>
improvements to tackle.<br>
<br>
<br>