[Vm-dev] Squeak/Pharo VM fork history (was: RE: [Pharo-dev] Squeak and Pharo speed differences)

Robert robert.withers at pm.me
Sun May 17 18:51:01 UTC 2020


On 5/17/20 2:30 PM, Robert wrote:

> On 5/17/20 2:03 PM, Jakob Reschke wrote:
>
>> Supporting something that Pharo invented, to continue being able to
>> target both Squeak and Pharo or simply work together, is mostly left
>> as an exercise to the interested or "needy" Squeaker. For example
>> developing equivalent tools to deal with Git repositories, porting the
>> new FFI and maybe the new compiler... *Active* sharing between Pharo
>> and Squeak seems to have been handled inadequately in the past, and it
>> seems to have ceased in the VM as well. Pharo shapes its future, and
>> maybe the committee does not like or cannot afford to put much effort
>> into the relationship with Squeak. Not doing it is also part of
>> freedom, sure.
>
> It's shortsighted. And it does not honor Pharo's heritage. Squeak needs Tonel & Tool support, to interoperate. Pharo needs Monticello maps. A worthy goal and ought be registered higher on priority.
>
> If the core technical issue is the use of CMake versus autoconf and so on, then why not support both build systems? Well, how can we do that without huge complexity and double the work of config specification, one may ask me? I reply, it requires common code mgmt systems, Tonel and Monticello in both Squeak & Pharo. Otherwise, we got nothing.
>
> The answer is that we construct a Smalltalk-based build system with a swap-able back-end to generate & emit CMake config or autoconf config. Run ./mvm which would call

>> squeak VMBuildSystem makeVMWithAutoconf
>> pharo VMBuildSystem makeVMWithCMake
>
> Similar to groovy, allow Smalltalk scripts to define the build.

The reference to groovy was the build system known as gradle. I propose we construct a gradle equivalent in Smalltalk.

---

Kindly,
Robert
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20200517/7b1b967f/attachment.html>


More information about the Vm-dev mailing list