On Apr 11, 2011, at 12:09 AM, Amir Ansari wrote:

It certainly works on Pharo 1.1. I haven't yet upgraded to 1.2, so I can't confirm that it works on the newest version of Pharo. Anyway, it's not just Squeak!

On Sun, 10 Apr 2011 22:08:47 -0600 Elliot Finley efinley.lists@gmail.com wrote:

...

bump. Is Magma only supported on Squeak?

You guys really need to stop promising that someday it will be easy to install Magma in Pharo and either: 1) put up a Metacello configuration that works under Pharo, 2) put up a wiki page with precise, tested instructions for getting it installed, or 3) stop saying that Magma is not a Squeak-only product and admit that you have no intention of actually supporting it under Pharo. Your current scheme is humiliating new users, and you're only getting feedback from the ones that get on the mailing list before giving up.

— Daniel Lyons

Daniel, just to be clear.

Magma is not yet supported on Pharo 1.2, but it is part of my plans to do so. In fact, I plan to look into it this month.

Please note that I expect Pharo to eventually produce resources (developers or money) to assist with Magma support and upgrades as Pharo becomes more and more divergent from Squeak in the future.

- Chris

On Mon, Apr 11, 2011 at 4:49 PM, Daniel Lyons fusion@storytotell.org wrote:

First, let me apologize for my explosion.

In Pharo 1.2, the correct way to install Magma _appears_ to be following these instructions in Pharo in the 'Installing external projects' workspace:

Gofer new        squeaksource: 'MetacelloRepository';        package: 'ConfigurationOfMagma';        load.

((Smalltalk at: #ConfigurationOfMagma) project latestVersion) load

This results in some version of Magma being installed, I'm not clear which. The result of running the test suite is that of 384 tests, 344 pass with 1 expected failure, ten failures and 29 errors. I don't yet know how to interpret the test runner screen, but it looks like the failures are in MaIdentityKeyWeakValueDictionaryTest, MaWeakIdentityKeyDictionaryTest, MaWeakKeyDictionaryTest, UnixProcessUnixFileLockingTestCase, UnixProcessWin32FileLockingTestCase, and WBTestBarrier. The errors seem to be in MaClientServerTestCase, MaWeakIdentityKeyDictionaryTest, WBTestBarrier and WBTestSystemChanges. I suspect those are all the missing and:and:and: selector on Boolean, because I got a lot of messages about that while running the tests.

I don't know if this means I have a working Magma or not. I would like to know.

If this is right, this information could be placed on the wiki. I see how to do that, and if it's right, I see what should go there, so if no one objects I would like to put this information on the wiki.

Thanks. And again, I am very sorry about my outburst.

-- Daniel _______________________________________________ Magma mailing list Magma@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/magma

On 14 April 2011 07:45, Daniel Lyons fusion@storytotell.org wrote:

On Apr 12, 2011, at 10:34 AM, Chris Muller wrote:

...

Magma is not yet supported on Pharo 1.2, but it is part of my plans to do so.  In fact, I plan to look into it this month.

http://lists.squeakfoundation.org/pipermail/magma/2011-March/001747.html http://lists.squeakfoundation.org/pipermail/magma/2011-February/001718.html

...

Please note that I expect Pharo to eventually produce resources (developers or money) to assist with Magma support and upgrades as Pharo becomes more and more divergent from Squeak in the future.

I appreciate that you're busy and I can understand not wanting to maintain a port of your software to a platform you don't use, but as long as Magma development remains this quiet, what good will it do for Pharo to provide developers? If it were "hey, this piece is broken under Pharo, please someone from Pharo take a look at it," it would be a lot easier I think for people to get involved and help shoulder some of the burden. It is (or should be) much easier to get help when you ask for it than when you wish for it silently, looking into the void.

I'm not an expert at Smalltalk or Pharo and *far* from being one in Magma, but I *am* a real programmer and I know how to fix bugs. Step one is knowing what they are.

My 2c. I don't see any reasons why Magma cannot work on Pharo. Its only a question of putting an effort on fixing all minor incompatibilities.

Btw, it would be interesting if we could set up a magma loader on Hudson server. It could also run test suite(s) after building image. But i don't have time on it right now, so if there anyone interested, let me know or write a message to pharo mailing list.

Ok, I'll announce when I have the Pharo version..

On Thu, Apr 14, 2011 at 3:57 AM, Igor Stasenko siguctua@gmail.com wrote:

I just talked with Marcus about this.

Its only a matter of giving an instructions, what should be loaded and in what order.

So we can set up a Hudson job to do that, and it will build pharo + magma image on every update. And then of course it can run Magma tests after build.

-- Best regards, Igor Stasenko AKA sig. _______________________________________________ Magma mailing list Magma@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/magma

Please see:

http://wiki.squeak.org/squeak/2639

Your answer is at the bottom of that page.

On Mon, Apr 11, 2011 at 1:59 AM, Amir Ansari fractallyte@csi.com wrote:

Hi List

Sorry for this long post, but I'm having a problem and just can't figure out where I'm going wrong...

I created a new class, 'FruitBowl', with variables 'color' and 'fruits', and added an instance of it to a database.

MyDB commit: [                MyDB root at: 'fruitbowls' put: (MagmaCollection new                        addIndex: (MaSearchStringIndex attribute: #color);                        addIndex: (MaKeywordIndex attribute: #fruits);                        yourself) ].

fruitbowls := MyDB root at: 'fruitbowls'.        myBowl := FruitBowl new color: 'silver'; fruits: #('apple' 'pear' 'plum'); yourself.

MyDB commit: [ fruitbowls add: myBowl ].        MyDB release.

I can query the database without any problem:        query := fruitbowls where: [ :each | each fruits includesAllOf: 'pear' subStrings ].

However, if I replace the collection of fruits and commit:        MyDB commit: [                myBowl fruits: #('walnut' 'hazelnut' 'chestnut') ].        MyDB release.

The new collection seems to be there:        reader := fruitbowls read: #color.        (reader at: 1) fruits

But I can no longer query:        query := fruitbowls where: [ :each | each fruits includesAllOf: 'walnut' subStrings ].        query size

I must be doing something fundamentally silly - can anyone point out what it is?

Thanks!

Amir _______________________________________________ Magma mailing list Magma@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/magma

Oh dear! For the example I provided, 'noteOldKeysFor' isn't working...

My original post was rather long and detailed, but the summary is: my MagmaCollection contains an array (of strings). When I replace the array with a new set of strings, I can see they're there, but a keyword query returns zero results.

I tried this (and various permutations):

MyDB session commit: [ MyDB session noteOldKeysFor: myBowl. "<- added this line, as indicated in the wiki" myBowl fruits: #('walnut' 'hazelnut' 'chestnut') ].

This query then fails:

fruitbowls where: [ :each | each fruits includesAllOf: 'walnut' subStrings ].

Any idea where I'm going wrong?

Thanks!

Amir

On Tue, 12 Apr 2011 11:38:42 -0500 Chris Muller asqueaker@gmail.com wrote:

...

Please see:

http://wiki.squeak.org/squeak/2639

Your answer is at the bottom of that page.

I have it working now. It's essential to invoke the method 'noteOldKeysFor:' within a reader (just as shown in the wiki). Here's the code for my example:

MyDB commit: [ reader := fruitbowls read: #color. reader do: [ :each | MyDB session noteOldKeysFor: each. each fruits: #('walnut' 'hazelnut' 'chestnut') ] ].

Thanks for your help, Chris!