- Pharo-dev - lists.pharo.org

[Pharo-project] Linus Torvaldsâs Lessons on Software Development Management
by Ben Coman Oct. 17, 2011

Oct. 17, 2011

While I am waiting for enough time to devote to learning more Smalltalk/Squeak/Pharo, I came across this article [1] which reflected upon recent discussion on Fuel version format compatibility, as well as on other projects. Perhaps it is that it is such a great environment for experimenting with ideas that it encourages developers to do so, but it seems that functionality often becomes orphaned as the trunk progresses. I see some efforts provide a more professional platform, and understand that involves 'some' ripping/replacement of unmaintainable code, but which contradicts the philosophy of 'no regressions' mentioned in the article - which in practice I imagine a hard line to walk. Linux is the poster child of open source development success, and while it is more mature that Pharo such that the presented philosophy might not fit exactly, perhaps something can be learned. I hope you find parts of it interesting. cheers, Ben [1] http://h30565.www3.hp.com/t5/Feature-Articles/Linus-Torvalds-s-Lessons-on-S…

3 2

[Pharo-project] How to use freetype?
by Igor Stasenko Oct. 17, 2011

Oct. 17, 2011

I feel a bit confused. I want to do something as easy as: font := Something fontByName: 'Arial'. but the problem is that i cannot find anything like that in freetype. There is FreeTypeFontProvider fontFor: aLogicalFont familyName: familyName but i puzzled, why it needs a logical font and why i should create it? Is there a general way to programmatically say: give me font 'Arial' pt size 9, or nil, if font cannot be found? -- Best regards, Igor Stasenko.

3 3

Re: [Pharo-project] Can this be improved in Glamourous?
by Marcus Denker Oct. 17, 2011

Oct. 17, 2011

On Oct 17, 2011, at 9:37 AM, StÃ©phane Ducasse wrote: > can you check > > https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.3/ > Because it should contain the latest shout and rb for 1.3 Looks good... (after a quick look). The only problem is that it has one undeclared so we have one failing test. https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.3/lastCompletedBuild/t… Marcus -- Marcus Denker -- http://marcusdenker.de

1 0

[Pharo-project] IS Smalltalk Source also an object?
by Guido Stepken Oct. 17, 2011

Oct. 17, 2011

Hi, there! I had a nice conversation about Smalltalk among software architects: I claim smalltalk to be the only programming system, where source is an object too, with all advantages and disadvantages! Am i right? What consequences follow out of that? regards, Guido Stepken

1 0

Re: [Pharo-project] Can this be improved in Glamourous?
by Marcus Denker Oct. 17, 2011

Oct. 17, 2011

On Oct 16, 2011, at 10:07 PM, Tudor Girba wrote: > Hi, > > If you give me the script, I build one based on the Core. > Husdon does: Gofer new squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfPharo'; load. -- Marcus Denker -- http://marcusdenker.de

4 9

[Pharo-project] Shout-Examples are cool :)
by StÃ©phane Ducasse Oct. 17, 2011

Oct. 17, 2011

Hi everybody Alain did a good job to describe four nice example of how to use Shout. Finally we get some examples we can learn from. The examples are in Shout-Examples category. Stef

1 0

Re: [Pharo-project] Can this be improved in Glamourous?
by Marcus Denker Oct. 17, 2011

Oct. 17, 2011

On Oct 16, 2011, at 11:14 PM, StÃ©phane Ducasse wrote: >> We need to fix the Metacello of Shout to load the latest version so we can build a Full image based >> on 13315, this should solve some of the problems. > > I published a new configurationOfShout load for 1.3 the latest version published by alain. > Now I'm sick so I will sleep. > I have started a build of 1.3 Full: https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.3/ -- Marcus Denker -- http://marcusdenker.de

2 1

Re: [Pharo-project] Can this be improved in Glamourous?
by Marcus Denker Oct. 16, 2011

Oct. 16, 2011

On Oct 15, 2011, at 11:13 PM, StÃ©phane Ducasse wrote: > > On Oct 15, 2011, at 10:58 PM, Tudor Girba wrote: > >> Hi, >> >> On 15 Oct 2011, at 22:36, Mariano Martinez Peck wrote: >> >>> Hi Tudor. Today I wanted to give it a try to 'GT-Inspector' 'GT-Playground' hence I did: >> >> I am happy you decided to give them a try :). But, the Playground is not yet ready :). >> >>> (ConfigurationOfGlamoroust project version: #bleedingEdge) load: #('GT-Inspector' 'GT-Playground' ). >>> >>> In a Pharo 1.3 release. To my surprise, it took like half an hour to install, and it even was broken because it installed a different version of refactoring browser that is not working in 1.3. >> >> This is due to a broken ConfiguraitonOfRB. I would actually be very interested in a fix in this area. > > what is broken? because I spent one full sprint to fix it. the fixed version needs a new Core (13315), while the Full image right now is 13307. The RB between the two is not compatible. We need to fix the Metacello of Shout to load the latest version so we can build a Full image based on 13315, this should solve some of the problems. Marcus -- Marcus Denker -- http://marcusdenker.de

3 3

Re: [Pharo-project] Can this be improved in Glamourous?
by Marcus Denker Oct. 16, 2011

Oct. 16, 2011

On Oct 16, 2011, at 9:36 PM, Tudor Girba wrote: > Hi, > > I load it in a fresh Pharo 1.3 13287 image just like the squeaksource page says: > 13287 is very (very!!!) old.. the last version of the Full image is 13307, the last Core image (which we need to build a full image of, after fixing the Shout MetaCello package) is 13315. Marcus

2 1

[Pharo-project] External arrays (DWORD, DOUBLE, etc.)
by Schwab,Wilhelm K Oct. 16, 2011

Oct. 16, 2011

Hello all, I have some classes that started out as place holders to allow me to load code I wrote for Dolphin, and over time, have become almost useful :) One thing I really needed was a contiguous array of doubles; there is a LOT of code that uses them. Enter ExternalArray with many subclasses, most of which translate #at: and #at:put: into appropriate methods of ByteArray (#doubleAt:put:, etc.). The numbers end up being stored in a byte array, so the indexing looks at position and element size. Access is very slow, but most of the time, I simply allocate a block of memory (e.g. create a byte array, which is fast) and then pass it to an external C/C++ function to do the real work. The passing is the subject of this message. I end up (for FFI) describing the arguments as void*, passing the byte array to the function. Is there a way to tell FFI that an instance of DOUBLEArray can be treated as a double*? Bill

2 1