I've created an issue[1] to track the project and started the port of magritte-metamodel/magritte on my fork of the project[2]. If you follow along in Issue #105[1], I'll record progress in commits and comments on the issue ...


[1] https://github.com/magritte-metamodel/magritte/issues/105

[2] https://github.com/dalehenrich/magritte/tree/issue_105

On 5/2/19 8:08 AM, dtrussardi@tiscali.it wrote:
Ciao  Dale,


On 5/1/19 8:58 AM, dtrussardi@tiscali.it wrote:
Ciao Dale, thanks.

have no great experience in managing this type of updates.
I had some considerations.

I've looked at this a bit closer and I notice that you are referencing the magritte-metamodel/magritte[1] and that project has not been ported to GLASS yet ... A quick way to tell if a particular project has been ported to GLASS is to look at the .travis.yml[2] file and check to see if there are any GemStone versions being tested.


The GsDevKit/Magrite3[3] was ported to GLASS and the test ran cleanly as of 3 years ago[4]. The project hasn't been updated to SmalltalkCI and the last version tested was GemStone 3.3.1[5]. I just triggered a new build to see where the project stands today[6] - the travis run passed for GemStone 3.2.15 and 3.3.1 ...

Moving forward, I would think the best bet would be to port the magritte-metamodel/magritte project to GLASS and GemStone 3.2.17, 3.3.9, and 3.3.9 ... I assume that Magritte has changed from the point where I forked Magritte 3.3.0 back in December of 2015 ...

From the release history[7] of GsDevKit/Magritte3, it looks like I ported 3.2.0, 3.3.0, 3.3.1, 3.3.2, and 3.4.2. the master branch is 3.4.2. I also got a start on 3.5.0, but didn't finish the port to the point where I merged to master ...

I would think that updating GsDevKit/Magritte3 3.4.2 to support smalltalkCI and GemStone 3.4.3, would be pretty straightforward ...

I have no idea about it...... how should I proceed?

The release history of magritte-metamodel/magritte[8] indicates that 3.5.4 was released in January of this year ... I've launched another GsDevKit/Magritte3 travis run for branch version_350[9] to see what the state of the GLASS Magritte3.5.0 port is ...

It lools like the Magritte 3.5.0 test[9] passes so it is likely that we'll be able to merge the GsDevKit/Magritte3 work into the magritte-metamodel/magritte project ...

I don't have full time to commit to doing this work, but barring any major surprises between Magritte 3.5.0 and  Magritte 3.5.4 and later, it should only take a week or so of elapsed time for me to do the port ... if there are significant changes then all bets are off ...

But what do you mean? That you can port   ( the GsDevKit/Magritte3 work into the magritte-metamodel/magritte project )  within a week?

I don't have a lot of time to mess around with porting projects that no one will use, so there are two different routes that I can take:

What I have developed is based on Seaside and Magritte and for now I am not able to review and change the reference base.

I would like and would be useful to continue in this direction.

    1. Port GsDevKit/Magritte which is known to run on GemStone 3.3.x and will likely run on GemStone 3.4.3 without a lot of changes. Downside is that it is 4 years old and may not meet your requirements. Going this route is likely to take around a week elapsed time (not full time).

in this case,  I don't know if feasible but: Set the development environment based on Pharo 7.0.3 (it seems stable) and install Magritte from ConfigurationOfMagritte3 to the version compatible with GsDevKit / Magritte.

If you tell me the relative ConfigurationOfMagritte3 version of  GsDevKit / Magritte I can try to set the Pharo 7.0.3 environment in this regard.

    2. Port magritte-metamodel/magritte to GemStone ... there are GemStone packages present, in the repository, but GemStone is not in the travis lineup, so the GemStone code has not been tested for a long time (maybe more than 4 years). Going this route will likely take a week (elapsed --- not full time) just to figure out how big of a job the port will be ...

You are currently trying to use magritte-metamodel/magritte with Pharo 7.0.3, so it seems that the right answer is to spend time porting , magritte-metamodel/magritte to GemStone ... then you'll be using the same code base on Pharo and GemStone ... if there aren't significant changes to Magritte in the last 4 years then a couple of weeks would be a good guess ... and I can handle that ...

This seems to me the best solution.

That can be useful also for the porting of Pier (which if I remember correctly is based on Magritte) ????

In this case I wonder if I can be useful?

If this makes sense to you, then I will start work on the port tomorrow after I see your email:)

Other considerations?



Magritte, Pier and Related Tools ...