[Newbies] Re: [Scratch] building the scratch image from the source?

Bert Freudenberg bert at freudenbergs.de
Fri Sep 28 16:23:39 UTC 2012

On 2012-09-28, at 17:00, Matthew Miller <mattdm at mattdm.org> wrote:

> Pardon me for the squeak newbie question. Is it possible to build the
> scratch image from the source files distributed in a scripted way? Do you do
> this internally when working on the project?


> In Fedora, shipping compiled binaries in the source isn't okay. The
> Scratch.image isn't an arch-specific compiled binary, but it *is* squeak vm
> bytecode, right?

It is a snapshot of the complete run-time memory of Squeak. Squeak is a version of Smalltalk, the original object-oriented system. It is not text-based. Coding in Smalltalk means loading a snapshot from disk, modifying objects in memory, and dumping the whole thing to a file between sessions. The "image" is evolved from version to version. It has some objects in it that were created decades ago. Literally. The last time the image was bootstrapped was in the 70ies.

This may sound odd, but e.g. the GPL states that sources must be distributed in their "Preferred Form For Modification". In Smalltalk's case, that preferred form is the image. All the code is available and ready to be inspected and modified using Smalltalk tools. It's just a bit different from most other popular programming systems.

Btw, we have a special list for Squeak newbie questions:


(actually, I'm going to cc this reply to that list)

- Bert -

More information about the Beginners mailing list