[Box-Admins] [INFO] squeaksource.org notes

[Chris Muller]

Hi Dave,

> I have added some Montecello packages to match the current running
> squeaksource.com image:
>
> 1) The SqueakSource package version that is running on squeaksource.com is
> SqueakSource.sscom-dtl.1130 in the source.squeak.org/ss repository. This is
> a branch (sscom) that captures differences versus the source.squeak.org image.

I adopted the two fixes you had in dtl.1130 into the main SqueakSource
branch, and saved it as SqueakSource-cmm.1134.

I noticed that comment on 1130 said it was to avoid Magma
requirements.  Magma is not and never has been required for
SqueakSource -- at least not on purpose.  I just looked in the
Dependency Browser and noticed a few methods had accidentally creeped
in and introduced dependencies on that package, so I removed them and
saved it as SqueakSource-cmm.1133.  I then committed another version,
cmm.1134, with fixes merged from dtl.1130.  I'm glad to get
those fixes into the production code, but those additions of #startUp
and #shutDown were unnecessary (erroneous, actually) because that
stuff is already handled in startUp: and shutDown: (the colon'ized
versions).

I DID fully consider the nature of squeaksource.com when I fixed up
the SqueakSource design years ago, including it's desire to use image
persistence and NOT Magma.  The way that's done under the new design
is:

   SSRepository storage: SSImagePersistence new

I tested this on a copy of squeaksource.com on one of my home
machines.  So if you are running dtl.1130, I hope you'll consider
switching over to cmm.1333, instead.  It works great and there's no
reason we should be forking the code.

> 2) Other support utilities (e.g. things that previously lived in workspaces
> in the image) are in package Startup-SqueakSource-dtl.9 in the
> http://www.squeaksource.com/SSImageInit repository.

I got a debugger when I tried to access those files due to some kind
of corruption or difference in date format (I was using 5.3).  But I
hope we can get any other fixes you have in them integrated into the
mainline code!

Thanks,
  Chris