Re: [Pharo-users] How to tell Versionner to load certain additional packages?

[Christophe Demarey]

Hello Sergi,

Le 28 mars 2014 à 16:09, Sergi Reyner a écrit :

 I created a ConfigurationOfArtie 
 (http://smalltalkhub.com/#!/~SergiReyner/Artie) that loads Bootstrap, 
 Magritte3, Pillar and ApplicationSecurity. I´d like it to load some other 
 packages like Magritte-Seaside or Bootstrap-Magritte. How to tell it to do so 
 with Versionner?

You can select the dependent project the package depends on (ex: Magritte3) and 
click on the edit button (the pen).
It allows to add loads directives.
For example, you can add default, Magritte-Seaside.

It is not yet documented as it is a recent add. If needed, you can also update 
the dependent project version with the right-click menu on the dependent 
project.

Best regards,
Christophe.



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [ANN] BabyMock 2

[Christophe Demarey]

Hello,

Thanks. Very nice library!

I have a question: is it possible to expect a method to throw an Exception?

I would like something like:
protocol describe
once: mock recv: #aMethod ;
signal: anError.
I did not find anything in the documentation.

I don't want to test that the mock signals an error but I want to test a code 
that needs to take into account an exception.

Best regards,
Christophe.

Le 11 mars 2014 à 12:30, Attila Magyar a écrit :

 I'm pleased to announce the 2.0 version of BabyMock. BabyMock is a visual
 mock object library that supports test-driven development.
 
 This version has a new syntax which is incompatible with the old version.
 Therefore it has a new repository
 http://smalltalkhub.com/#!/~zeroflag/BabyMock2
 (BabyMock 1 is still available at its old location, but in the future I'd to
 focus on the development of BabyMock2, so don't expect too many changes
 regarding the old version).
 
 Changes in 2.0
 
 - A new, extensible DSL (no more should/can)
 - Improved error messages, history of messages, detailed information about
 argument mismatches
 - An improved, Spec based GUI
 - Clicking on a mock opens an inspector on the expectations
 - Clicking on a message opens an inspector on the message
 - Object methods can be mocked by defaults
 - Blocks can be executed after receiving a message by the mock. The block
 has access to the arguments of the incoming message.
 - Any argument matcher
 - Cleanups and simplifications in the code
 
 I hope you don't mind the changes regarding the syntax. Personally I think
 it has lot more pros than cons.
 
 More information
 
 http://smalltalkhub.com/#!/~zeroflag/BabyMock2
 
 p.s. 
 It needs Pharo3.0.
 
 Attila
 
 
 
 --
 View this message in context: 
 http://forum.world.st/ANN-BabyMock-2-tp4748530.html
 Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [ANN] BabyMock 2

[Christophe Demarey]

Thanks Guillaume.
It works well.

With BabyMock2, the syntax is 
= aBlock

Le 9 avr. 2014 à 11:13, Guillaume Larcheveque a écrit :

 I think you can have your mock send an exception by using #answers:aBlock and 
 signal the exception in the block
 
 
 2014-04-08 15:58 GMT+02:00 Christophe Demarey christophe.dema...@inria.fr:
 Hello,
 
 Thanks. Very nice library!
 
 I have a question: is it possible to expect a method to throw an Exception?
 
 I would like something like:
 protocol describe
 once: mock recv: #aMethod ;
 signal: anError.
 I did not find anything in the documentation.
 
 I don't want to test that the mock signals an error but I want to test a code 
 that needs to take into account an exception.
 
 Best regards,
 Christophe.
 
 Le 11 mars 2014 à 12:30, Attila Magyar a écrit :
 
  I'm pleased to announce the 2.0 version of BabyMock. BabyMock is a visual
  mock object library that supports test-driven development.
 
  This version has a new syntax which is incompatible with the old version.
  Therefore it has a new repository
  http://smalltalkhub.com/#!/~zeroflag/BabyMock2
  (BabyMock 1 is still available at its old location, but in the future I'd to
  focus on the development of BabyMock2, so don't expect too many changes
  regarding the old version).
 
  Changes in 2.0
 
  - A new, extensible DSL (no more should/can)
  - Improved error messages, history of messages, detailed information about
  argument mismatches
  - An improved, Spec based GUI
  - Clicking on a mock opens an inspector on the expectations
  - Clicking on a message opens an inspector on the message
  - Object methods can be mocked by defaults
  - Blocks can be executed after receiving a message by the mock. The block
  has access to the arguments of the incoming message.
  - Any argument matcher
  - Cleanups and simplifications in the code
 
  I hope you don't mind the changes regarding the syntax. Personally I think
  it has lot more pros than cons.
 
  More information
 
  http://smalltalkhub.com/#!/~zeroflag/BabyMock2
 
  p.s.
  It needs Pharo3.0.
 
  Attila
 
 
 
  --
  View this message in context: 
  http://forum.world.st/ANN-BabyMock-2-tp4748530.html
  Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
 
 
 
 
 
 -- 
 Guillaume Larcheveque
 



smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] Versionner screencast

[Christophe Demarey]

Hello,

I made a small screencast on Versionner to explain how to use it: 
https://www.youtube.com/watch?v=S5Dbmmln8tA

Regards,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [ANN] BabyMock 2

[Christophe Demarey]

Hi,

Le 17 avr. 2014 à 19:22, Attila Magyar a écrit :

 Hi Christophe,
 
 Sorry for my late response. You can use the following syntax to signal an
 error in response to an incoming message:
 
 protocol describe
once: mock recv: #msg;
signals: Error.

I did not notice this one. Thanks!

I have another point. Sometimes, I would like to test a method doing some other 
calls to methods of the same class.
In this case, I would like to mock all the class methods but the method I would 
like to test.
Do you sometime meet the same need? Is there a way to do that?

Thanks,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [ANN] BabyMock 2

[Christophe Demarey]

Le 23 avr. 2014 à 15:37, Attila Magyar a écrit :

 Hi Christophe,
 
 I'm not sure I fully understand, are you referring to partial mocking? Where
 you have a real object but with some mocked methods? If this is the case,
 then no, this is not supported. Some people consider this as a test smell,
 and I agree with them. Maybe the object under test is too big/complex/has
 multiple responsibilities, and extracting some parts from it could solve the
 problem. I haven't seen legitimitate use for partial mocking so far, where
 one couldn't solve the problem with redesign/refactoring in a better way.
 Until this happens I'd like to skip this feature.

I understand your point of view and I have quite the same ... (I also had the 
same discussion with Guillaume)

... but sometimes, even in a well-designed class, you may need to test a very 
small part of this class, and in this case, you need a real object with some 
mocked methods.
It is not crucial because you can test your class as a whole (if well designed, 
only a few methods are involved) but it prevents to test a smaller part.
Then, maybe the work to support a real object with some mocked methods does not 
worth it.

Best regards,
Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [ANN] BabyMock 2

[Christophe Demarey]

Le 23 avr. 2014 à 22:25, Attila Magyar a écrit :

 Christophe Demarey wrote
 ... but sometimes, even in a well-designed class, you may need to test a
 very small part of this class, and in this case, you need a real object
 with some mocked methods.
 
 I assume that there is a method need to be tested, but it calls towards an
 other either public or private method of the same object. And you'd like to
 mock that second method. Am I understanding correctly? 

right

 Why can't just allow the first method to call the second one? Is it slow,
 complicated to setup or requires some external resource to execute? Does it
 help to extract the logic from the second method to an other object? Without
 a concrete example, I'm just speculating.

My point here is just to test a method in isolation.
If MyClassmethodA calls MyClassmethodB and there is an error in 
MyClassmethodB, I don't want to have the test on methodA failing but just the 
test on MethodB.
With cascade testing, you will get a lot of failing tests and you need to 
investigate to find the culprit.

As you said, if the class is well designed, it is not a big trade-off to unit 
test a small set of methods if they are short and related but I have the 
feeling that it would be good to be able to do that in some cases.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versionner: Specifying un-configured packages as dependencies

[Christophe Demarey]

Hi,

Le 20 mai 2014 à 15:01, Johan Fabry a écrit :

 Hi all,
 
 I’m using Versionner to make a ConfigurationOf and things are working quite 
 well (thanks for that!) until I want to add the Roassal2Spec package as a 
 dependency. The dialog box only shows packages with a configuration, and 
 Roassal2Spec does not have one. It’s just one package. How can I add this guy 
 to my configuration?

The only way (and not recommended) to add it with Versionner would be to add it 
as a package of your project but it will not work as it is hosted on another 
repository.
Why not adding a new config for this project (in your project repository) or, 
better, in the roassal repository?

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versionner: Specifying un-configured packages as dependencies

[Christophe Demarey]

Le 20 mai 2014 à 22:53, Sean P. DeNigris a écrit :

 Christophe Demarey wrote
 but it will not work as it is hosted on another repository
 
 Not that I'm recommending it, but if necessary, you could always copy it
 into your project's repo to get around that (or I think you can specify a
 repository in the #with: block of the package spec.

To me, it is preferable to add a configuration (even if very small) for each 
dependency outside your project. It is the way it should be done (that's why 
Versionner only allows that) but they are always other ways or workarounds.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versionner: Specifying un-configured packages as dependencies

[Christophe Demarey]

Le 22 mai 2014 à 01:02, Johan Fabry a écrit :

 
 On May 21, 2014, at 5:07 AM, Christophe Demarey christophe.dema...@inria.fr 
 wrote:
 
 To me, it is preferable to add a configuration (even if very small) for each 
 dependency outside your project. It is the way it should be done (that's why 
 Versionner only allows that) but they are always other ways or workarounds.
 
 Thanks for the answers, my trick now was to ask Alex to include the required 
 package inside the configuration of Roassal. :-)
 
 In general I am not so sure that adding configurations for one simple package 
 are the way to go though. It seems like adding extra layers of indirection 
 for reasons that are unclear to me.

The reason is that any package* should be self described. Your configuration 
describes your project. If you use another piece of software, it should also be 
self-described.
Let's take the following example:
for now, it's easy for you to add a direct dependency to Roassal2Spec, just as 
if it was a part of your project ... = no problem
... but it is not. Maybe tomorrow, Alex will need to add a dependency to one or 
more packages for Roassal2Spec = you will have problems because you will not 
get these dependencies. Of course, you can still add them to your configuration 
but you can easily see that you will end with a configuration including all 
flatten dependencies, and with a lot of dependencies not directly related to 
your project (transitive dependencies). Flatten dependencies is really 
something that we should avoid, else maintenance of dependencies will be a hell.
That's why I'm really in favor of having configurations, even for one single MC 
package inside the configuration.

* I mean, not a Monticello package but a piece of software that you want to 
deliver independently. It may be a whole project or part of it. It may also 
refers to one or many Monticello packages

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Unintuitive behavior of class-side initialize

[Christophe Demarey]

Le 27 août 2014 à 19:04, p...@highoctane.be a écrit :

 Yeah don't do super initialize unless you know why on the class side.
 
 That being said class side initialize is pretty cool when loading new 
 packages and setting everything up automagically.
 
In my opinion, it would be better to have a package initialization method for 
that purpose. 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Packaging Pharo for many distributions at once

[Christophe Demarey]

Le 24 sept. 2014 à 11:00, Thierry Goubier a écrit :

  Hi Damien,
 
 I would be interested in a zeroinstall [http://0install.net/] version :) A 
 cool system because it is user-level (no need to go into system admin mode).

It looks interesting but requires the user to install 0install before being 
able to install Pharo. A big drawback ...

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Packaging Pharo for many distributions at once

[Christophe Demarey]

Le 24 sept. 2014 à 18:51, Damien Cassou a écrit :

 On Wed, Sep 24, 2014 at 12:47 PM, Thierry Goubier
 thierry.goub...@gmail.com wrote:
 All non-native package managers have that issue.
 
 
 nix included

I know ...
That's why native packages are still interesting even if it's difficult to 
manage various package descriptions and build.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versioneer. nice. mmMmmm...

[Christophe Demarey]

Hi Ben,

Le 22 sept. 2014 à 18:23, Ben Coman a écrit :

 
 While I had done the Versioneer tutorial some time ago, I had some inertial 
 problem using it in anger.  Now I'm happy to report using it just now to 
 update the ConfigurationOfPharoLauncher was simple and smooth - which is 
 great since I've never done a Configuration by hand.  I thought it would be 
 useful for others to see how simple it was...
 
 1. Loaded latest ConfigurationOfPharoLauncher.
 2. Loaded and tested newer PharoLauncher packages.
 
 3. Opened World  Tools  Versioneer.
 4. Selected PharoLauncher and Open Project.
 5. Selected development list item.
 6. Clicked Release Version and Ok to accept auto increment of version 
 number.
 7. Clicked Commit the project
 Now there is an updated ConfigurationOfPharoLauncher in its Smalltalkhub 
 repository.
 
 Easy with Versioneer.  mmM. nice. shiny.
 cheers -ben

I'm happy to see that this tool is useful. Still improvable on many points but 
helps a lot for basic tasks.

Christophe



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versioneer. nice. mmMmmm...

[Christophe Demarey]

Hi Kilon,

Le 29 sept. 2014 à 19:24, kilon alios a écrit :

 great tool, the only thing missing is a button to load the configurations

What do you mean?

 /versions. 

To load a version, select a version in the version list, rigth-click and select 
load. Isn't it enough?



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versioneer. nice. mmMmmm...

[Christophe Demarey]

Le 30 sept. 2014 à 11:15, kilon alios a écrit :

 Not really, not if you are a begineer like me and does not come to mind to 
 right click on the entry. A GUI should have as button all the basic 
 functionality. Obviously not all functionality should be crammed into a GUI , 
 though depending on the design thats possible too. 

I understand your point. At the same time, I took care to not put too many 
buttons on the UI to avoid users to be lost. The best thing to do should be in 
the middle.

 
 Also I do feel Versioneer can be an excellent tool even replacing 
 Configuration Browser, allowing users to browse through projects in metarepo 
 repositories and choose specific version for each package to load, even 
 display information for those packages. 

I also have this feeling. We should just take care to give a way to load a 
configuration that is not in any metarepo repository.

 I love the tool as it is, but I do feel there is a lot more potential in it 
 and I am willing to contribute if that is desirable. 

Of course, all contributors are welcomed. I'm already working from time to time 
with Diego to merge his Metaceller model with Versionner model. It will open 
doors to support platform-specific code in dependencies.
I just added you to Versionner contributors. If you want to contribute, please 
ask me which version to start with because Monticello is a hell to manage 
parrallel development?

Cheers,
Christophe.



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versioneer. nice. mmMmmm...

[Christophe Demarey]

Hi Kilon,

Le 30 sept. 2014 à 15:53, kilon alios a écrit :

 Actually the best thing to do when you 2 choices is to pick the third one. In 
 this case its what I call double mode , you see it quite frequently with 
 software synthesizer, as you can imagine a synthesizer is a complex 
 electronic music instrument that can reach up to thousand of parameters 
 easily so what designers have done is to create an easy mode and an advanced 
 mode. In easy mode is for those that want to create sounds on the go, usually 
 that means around 10-30 parameters buttons and advanced mode is divided into 
 panels that can expanded collapse to control the complexity of the GUI that 
 may contain even thousands of parameters. 
 
 GUI design is a science and an art but yes you can have your cake and eat it 
 too. 
 
 I think with Versioner we can keep it small for now and then if it get a bit 
 more complex go for a double mode. 
 
 About collaboration  if you are familiar with git and github then nothing can 
 beat this workflow . You create a repo in gihub, I fork it, and I then send 
 pull requests to you, pull requests do not affect your code and you can 
 choose what to merge and not to merge giving you complete control and me the 
 safety that I wont brake anything. You can also comment the code directly 
 using github web interface and of course the issue tracker for any issue that 
 may be reported. If you are not familiar with git  or dont want to us it , I 
 have no issues committing to smalltalkhub , I will create a mirror repo in 
 github anyway for my work since this is how I work when it comes to my Pharo 
 projects.  
 
 The choice is your, I will now start studying the code and see what I can 
 bring to it. 

Sorry for the delay. I was about to move Versionner to git but finally, I 
prefer to wait a bit before switching.
I cleaned up Versionner repository, i.e. give right package names so that we 
can easily catch branches.
I think it will be the best way to collaborate for now: commit in a branch in 
the current repository on SmalltalkHub. Please, create a new branch for your 
contributions, and ping me if you have something to show or if you need help.

Best regards,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] inria survey

[Christophe Demarey]

Le 6 nov. 2014 à 12:21, Bernat Romagosa a écrit :

 What should I enter under Establishment? Is it the company/institution 
 where one works/studies?

Yes, your company/institution where one works/studies



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Nautilus groups

[Christophe Demarey]

Le 10 févr. 2015 à 16:39, Martin Bähr a écrit :

 Excerpts from Christophe Demarey's message of 2015-02-10 14:28:19 +0100:
 We would like to know if some of you use Nautilus groups functionality:
 dynamic groups: 'Last modified packages', 'Most viewed classes', 'Work'
 
 do you mean 'last modified classes'?

yes, sorry

 We have the feeling that it is not used (or not a lot) and we think to 
 replace it with another mechanism (for example, let the user define his/her 
 'favorites' classes / packages).
 
 how is that different from user defined groups?

Well, it is not yet clear but at least, do not flatten all classes and keep 
packages.

[Pharo-users] Nautilus groups

[Christophe Demarey]

Hello,

We would like to know if some of you use Nautilus groups functionality:
dynamic groups: 'Last modified packages', 'Most viewed classes', 'Work'
user defined groups : from the menu, you can flatten the content of many 
packages and get all classes into the newly created group.
We have the feeling that it is not used (or not a lot) and we think to replace 
it with another mechanism (for example, let the user define his/her 'favorites' 
classes / packages).

Thanks for the feedback.

Re: [Pharo-users] Omit a method from a mc commit

[Christophe Demarey]

Hi,

Le 17 mars 2015 à 00:27, Serge Stinckwich a écrit :

 Hi all,
 
 is there any way to omit a method from a Monticello commit for example
 in the case of containing a password that you don't want to publish ?

If it is one shot, you can use kommiter to cherry pick the changes you want to 
commit.

Regards,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] GTTools and Mongo

[Christophe Demarey]

Really nice addition!

Thanks

Le 10 mars 2015 à 10:20, Torsten Bergmann a écrit :

 After loading MongoTalk or VoyageMongo in Pharo 4 you can additionally 
 load the package: Mongo-Pharo-Tools from mongotalk repo on SmalltalkHub.
 
 This way you can browse your mongo database (collections and documents) 
 directly from within Pharo using GTInspector tools (see screenshot). 
 
 Just by adding two methods - nice!
 
 Bye
 T.
 mongo.png



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Pillar parser rewrite: PetitParser or not?

[Christophe Demarey]

Le 11 mars 2015 à 23:07, Tudor Girba a écrit :

 I completely agree.
 
 I think there are mostly pros. It can be that the parsing will be slower, but 
 I do not think this is a critical issue for Pillar. 

not sure about the speed.
I hope we will adopt pillar for class / packages comments. Then, you need a 
fast parsing to render it nicely and quickly on the browser.
Fast/slow always depends on your comparison point. To be more precise, I would 
say we need to render (parsing + syntax highlighting) Pillar documents (at 
least one page) in less than 10 ms. 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] GTTools and Mongo

[Christophe Demarey]

Hi Phil,

Le 11 mars 2015 à 08:48, p...@highoctane.be a écrit :

 FWIW Mapless and its mongo support is quite awesome. I am using it for a 
 feature and it has tremendous potential.
 

I would be interested to have a small comparison between voyage and mapless 
(pros and cons). Do you have enough feedback to list pros/cons for each library?

Christophe

smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] Jenkins build trigger

[Christophe Demarey]

Thanks to Esteban add-on on Smalltalkhub, it is now possible to notify Jenkins 
that there is a new package published for your project.
What we do for now is to poll very often (each minute) smalltalkub to check if 
there is an update. It is a big waste of resources!
Now we can do better! It would be very nice to update your jobs to now use the 
commit hook.

Here is the way to go:
go to your job configuration on Jenkins
In the build triggers section, 
enable Trigger builds remotely (e.g., from scripts) and choose an 
authentication token. 
disable [URLTrigger] - Poll with a URL
go to smalltalkhub, on your project settings page
Fill in the commit hook url field with 
https://ci.inria.fr/pharo-contribution/buildByToken/build?job=JOBNAMEtoken=TOKEN
 by replacing JOBNAME and TOKEN by the Jenkins job name and the token chosen.

In the end, the build is now triggered immediately after the commit (no more 
delay).

Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] How to silently generate and remove code

[Christophe Demarey]

Le 30 mars 2015 à 09:10, Marcus Denker a écrit :

 
 On 29 Mar 2015, at 10:32, stepharo steph...@free.fr wrote:
 
 Before that I would like that the compiler outputs classes definition in an 
 environment (that can be the default one) but that can be the one
 I want and specified from the outside of the compiler.
 
 For installing, it is the ClassInstaller that does that (and yes, it has an 
 environment).
 
 What the compiler does not have is to look up references e.g. to symbols 
 regarding to an environment, but that is because we
 don’t have that concept right now in the system in general. globals/class 
 vars are requested from the class (which defines its environment).
 
 So in general, we first need a use case, else it is not clear what to do…

A first use case could be to install new code in a sandbox, check if the 
loading is good, and then move the code to the default environment, else drop 
the sandbox.
It is also a part of the feature asked by Johan: install code in a sandbox for 
tests purposes and then drop it without perturbing the system.

smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] GTSpotter with no results

[Christophe Demarey]

Hi,

I would like to know if some of you already got this problem: when I open 
GTSpotter (with the shortcut) and I start to type quickly (before any result 
appears) in the text I'm searching for, I get no result at all. Even If I 
remove all characters from the search field, GTSpotter still presents an empty 
list.

Thanks,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] GTSpotter with no results

[Christophe Demarey]

I can reproduce it very easily by typing fast.
Maybe it is because I have an old computer ...

Le 31 mars 2015 à 11:00, Tudor Girba a écrit :

 I never got it, but I would still be interested in how to reproduce it :).
 
 Doru
 
 On Tue, Mar 31, 2015 at 10:53 AM, Andrei Chis chisvasileand...@gmail.com 
 wrote:
 I got this error once or twice but could not reproduce it.
 
 
 Cheers,
 Andrei
 
 On Tue, Mar 31, 2015 at 10:47 AM, Christophe Demarey 
 christophe.dema...@inria.fr wrote:
 Hi,
 
 I would like to know if some of you already got this problem: when I open 
 GTSpotter (with the shortcut) and I start to type quickly (before any result 
 appears) in the text I'm searching for, I get no result at all. Even If I 
 remove all characters from the search field, GTSpotter still presents an 
 empty list.
 
 Thanks,
 Christophe.
 
 
 
 
 -- 
 www.tudorgirba.com
 
 Every thing has its own flow



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Gitfiletree unstable on Windows

[Christophe Demarey]

Le 31 mars 2015 à 16:29, Peter Uhnák a écrit :

 Metacello new
   baseline: 'FileTree';
   repository: 'github://dalehenrich/filetree:pharo4.0_dev/repository';
   load: 'Git' 
 I thought that you needed to have GitFileTree loaded to have github:// and 
 Git support. Is that false? Does that mean that github:// (bitbucket://) 
 doesn't use git at all?

right. It downloads a zip file with all sources from github and then a simple 
filetree repository is available.

Christophe.



smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] Versionner repository moved to PharoExtras

[Christophe Demarey]

Hi,

I just moved the Versionner repository from 
http://smalltalkhub.com/#!/~demarey/Versionner/ to 
http://smalltalkhub.com/#!/~PharoExtras/Versionner/.
It will ease contributions and it will be easier to find with other tools / 
library of Pharo.

Regards,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] ZnUrl directory

[Christophe Demarey]

Hi,

I just tried to evaluate:
(ZnUrl fromString: 'file:///tmp/') directory 
and got 'tmp/' as result.

I'm a bit suprised. Shouldn't it be '/tmp/' ?
Be fore trying to fix I would like to be sure what is the expected behavior.

Thanks,
Christophe.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Eating your own dog food (Site/Blog)

[Christophe Demarey]

There is a plan to have pillar rendering with TextModel. I don't think it will 
support the full pillar syntax.

Le 8 avr. 2015 à 11:16, Norbert Hartl a écrit :

 What you are after? A static site generator or a dynamic engine? Btw. what do 
 you use to render pillar inside the image to something readable?
 
 Norbert
 
 Am 08.04.2015 um 08:32 schrieb stepharo steph...@free.fr:
 
 I started to work on using pillar to do that but nothing ready for people to 
 use.
 I will work on it at Lviv or during my vacations just after.
 I want to use it for my website. Right now parts of it are automatically 
 generated from Pharo.
 
 Stef
 
 Le 7/4/15 10:56, Yuriy Tymchuk a écrit :
 Oh, I knew that there was a discussion about that, but couldn’t find where.
 
 Still, is anybody else using Marina except for Pharo website?
 
 Uko
 
 
 
 On 07 Apr 2015, at 10:45, Damien Cassou damien.cas...@gmail.com wrote:
 
 
 Yuriy Tymchuk yuriy.tymc...@me.com writes:
 
 if I want to create a site/blog in pharo… What do I use? I know that
 there is Pier, but when you check its webpage, the last update is in
 2011. Pharo website is built with Marina, but when I check the repo,
 the last commit was done by me and it was 2 month ago. Will Marina
 continue to evolve? And is anyone else using Marina?
 
 http://forum.world.st/Any-framework-to-create-a-blog-over-seaside-td4810615.html
 -- 
 Damien Cassou
 http://damiencassou.seasidehosting.st
 
 Success is the ability to go from one failure to another without
 losing enthusiasm. --Winston Churchill
 
 
 
 
 
 
 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Spec TreeModel dynamic contents

[Christophe Demarey]

Le 19 mai 2015 à 18:04, webwarrior a écrit :

 I want to make a tree using TreeModel that reacts to changes in underlying
 data - when an item changes, tree node for that item should update its state
 (recompute its children, etc.).
 
 Rebuilding the whole tree is ok, as long as selections and
 collapsed/expanded state of nodes are preserved.

yes but not very efficient if you have a big tree

 Using #updateTree however resets all these things. And there is no easy way
 to save/load them either - #selectedItems/#selectedItems: work with TreeNode
 instances that get reset, and collapsed/expanded states of nodes are simply
 unaccessible from TreeModel.

did you check the implementation of Komitter?
It keeps the state of your tree (selected items, collapsed/expanded).  So maybe 
it is doable.

 I also tried subclassing TreeNodeModel. But I found no way to update
 underlying UI widget. There is a huge mess involving TreeNodeModel,
 MorphicTreeNodeAdapter, SpecTreeNodeModel, and MorphTreeNodeMorph.
 
 The only option left is to extract all needed information from TreeNodeModel
 hierarchy (preserving the tree structure) before update, and then try to
 reapply it to newly created hierarchy.
 
 Or is there an easier way?
 
 
 
 --
 View this message in context: 
 http://forum.world.st/Spec-TreeModel-dynamic-contents-tp4827440.html
 Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] gtInspectorPresentation:, TreeModel, pragmas, visitors and polluting protocols

[Christophe Demarey]

Why not try to use TreeNodeModel to define different kind of nodes? On each 
node you can redefine the way to get children, the icon, etc.
Then you give these nodes as the tree roots.

Le 18 juin 2015 à 10:10, Peter Uhnák a écrit :

 
  I don't fully understand the TreeModel example, what would your tool do 
  with all the men/displayBlock/childrenBlock ... - methods?
 
 
 Ah, I should've explained what TreeModel is; it is a Spec UI widget to 
 display tree-like structures:
 
 
 ​
 
 So the way this works is that I give the TreeModel some root object (DCFsm) 
 and then blocks that describes some behavior.
 So in childrenBlock: [ :parent | ... ] the block is expected to return child 
 elements for the parent.
 In displayBlock: you pass a block that returns the label that you will see.. 
 e.g. [ :anObject | anObject name , ' (' , anObject class name , ')' ]
 iconBlock: returns the icon for the object and so on and so forth.
 
 So the way I have implemented it right now is something like
 
 ~
 childrenFor: anObject
   | col name |
   name := anObject class asString asSymbol.
   col := OrderedCollection new.
   name = #DMMindMap
   ifTrue: [ ^ col ].
   (anObject isKindOf: DCDiagram)
   ifTrue: [ col addAll: anObject elements ].
   (anObject isKindOf: DCNamedElement)
   ifTrue: [ 
   name = #BormParticipant
   ifTrue: [ col addAll: anObject nodes ].
   name = #BormActivity | (name = #BormState)
   ifTrue: [ col addAll: anObject outgoing ].
   name = #BormActivity
   ifTrue: [ col addAll: anObject sent ].
   name = #BormCommunication
   ifTrue: [ col addAll: anObject dataFlows ].
   name = #BormCommunication
   ifTrue: [ 
   anObject hasConstraint
   ifTrue: [ col add: anObject 
 constraint ] ].
   name = #BormTransition
   ifTrue: [ 
   anObject hasConstraint
   ifTrue: [ col add: anObject 
 constraint ] ] ].
   ^ col
 ~
 
 Which is unreadable, unmaintainable and nonextensible. All the #Borm* symbols 
 actually refer to classes of a completely independent package and repository 
 which doesn't even need to be loaded in the system.
 
 I think what Peter means is: the object oriented way to handle different 
 types is by dispatching on each type (~ double dispatch, ~ visiting). You 
 could need this for different aspects of your application. In the end, the 
 same core mechanism will be implemented multiple times. Sadly, minor 
 semantics/structural differences will make it hard to abstract this out.
 
 Yes, double-dispatch seems like the apt solution, however it would lead to 
 NxM number of classes (as mentioned before). But the more I think about it 
 the more I feel that I will not get any better than that. 
 
 Peter



smime.p7s
Description: S/MIME cryptographic signature

[Pharo-users] loading with gofer without network

[Christophe Demarey]

Hi,

I'm trying to load a package (mcz) from the package cache with Gofer.
|gofer|
gofer := Gofer new
repository: MCCacheRepository uniqueInstance;
repository: (MCSmalltalkhubRepository owner: 'Seaside' project: 
'Grease11');
version: 'Grease-Core-JohanBrichau.94';
disableRepositoryErrors;
load

As you see, I gave 2 repositories (the first is local, the second one is 
remote) to Gofer.
If you try to execute this code without the network, it will failed! It 
shouldn't because the mcz is in the package cache.

I think it is a bad behavior of Gofer. It does not have to resolve references 
for ALL the repositories because, at the end, you only use one.
I want to change this behavior to only get the first resolved reference (no 
need to iterate more). If not acceptable, I could only ignore repositories not 
reachable (disableRepositoryErrors does not work for this use case).
Is it ok to change it?

Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] stream not working with MemoryFileSystemFile

[Christophe Demarey]

Le 1 juil. 2015 à 14:42, Torsten Bergmann a écrit :

 Christophe Demarey
 I tried to create an in-memory file as following:
 (MemoryFileSystemFile named: 'Foo') 
  writeStreamDo: [ :stream | stream nextPutAll: 'bar' ]
 
 Then I get an improper store into ByteArray. It looks like Character cannot 
 be store in a byte array.
 I do not know if the bug is on the memory file system part or other.
 Any idea?
 
 Why do you use MemoryFileSystemFile directly. Usually the FileSystem is 
 the starting point. And there it works:
 
 |fs file|
 fs := FileSystem memory.
 file := fs / 'test'.
 file writeStreamDo: [ :stream | stream nextPutAll: 'bar' ].
 fs inspect
 
 The nice thing is that you can now browse the file system and file contents
 with the new GT Inspector. 

oh, perfect.
Looks like I was not using the good api.

Thanks,
Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [Pharo-dev] TestAsserter>>assertCollection:hasSameElements:

[Christophe Demarey]

Hi Peter,

Le 24 oct. 2015 à 20:36, Peter Uhnák a écrit :

> bump? :)
> 
> On Tue, Sep 29, 2015 at 12:57 AM, Peter Uhnák  wrote:
> How practical it is to do set-based comparison in 
> TestAsserter>>assertCollection:hasSameElements: ?
> 
> For example #(1 1 2) has same elements as #(1 2) which may make sense for 
> sets, but not for bags.
> 
> The main reason I was using it is that in tests the expected collection may 
> be created by hand,
> which means it is very often an array #(...), { ... }, while models very 
> often return OrderedCollections (at least in my case).


I introduced #assertCollection:hasSameElements:  for this reason.
The problem with using sets is that I do not expect that  #(1 1 2) has same 
elements as #(1 2)

> 
> So my question is --- how to compare collections irrespective of type, and 
> possibly of order?
> 
> A) always convert the actual collection toanArray
> 
> B) change difference: behavior for non-sets (ton of work with catastrophic 
> consequences)
> 
> C) extend TAssertable with new methods like 
> 
> #assertElementsOf: actualCollection equals: expectedCollection
> #assertUnorderedElementsOf: actualCollection equals: expectedCollection

This solution has my preference

> C.a) directly in Pharo/SUnit

yes, for me, it is a very basic assertion that should be available as default.

> C.b) in independend SUnit-Extensions repository/project
> C.c) just in my project
> 
> D) stop discussing non-problems
> 
> Thanks,
> Peter
> 

Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Request to put up my under-development application for automated build

[Christophe Demarey]

Le 9 juil. 2015 à 10:17, Stephan Eggermont a écrit :

 On 09-07-15 09:47, Peter Uhnák wrote:
 As I've been through this just a week ago...
 
 1. register for CI https://ci.inria.fr/
 2. request participation in pharo-contribution
 3. _read documentation_
 4. https://ci.inria.fr/pharo-contribution/job/JobTemplate/
 (4b. get inspired by configuration of other projects if needed)
 
 Easiest is mostly to find the job most similar to yours
 and just copy that, changing the required configuration/version,
 description, and emailadress.

yes, and if you do not need to run multiple configurations, do not use a matrix 
job but rather a free-style job

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Becoming good at VM development

[Christophe Demarey]

Hi

Le 28 août 2015 à 10:11, Ichiseki a écrit :

 Hi
 hello to everyone  everybody
 I am interesting in going deeply into VM design, development, optimization,
 etc. I know a lot of Haskell, a good amount of C and some languages that use
 VM like Python, Lua and Pharo of course.
 What would be a road map (books, articles, code to read) to understand
 deeply and in detail the Pharo VM, the inter operation between the VM and
 the image. I know this could take a lot of time but I'm committed to invest
 that time.
 Please help me as literature on these topics is very scarce...
 Thank SO SO much in advacne
 ichiro 

There are some nice blogs:
http://www.mirandabanda.org/cogblog/
https://clementbera.wordpress.com/ (cog/spur/sista categories)

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Mocking ZnClient

[Christophe Demarey]

It is very easy to do by using BabyMock2 for example

Le 14 sept. 2015 à 15:31, Damien Cassou a écrit :

> Hi,
> 
> I need to write unit tests for code that fetches data on the web through
> ZnClient. Is there a way to temporarily replace all creations of
> ZnClient instances by instances of another class that will reply with
> fixed results without fetching on the web?
> 
> I already wrote something like that for the Pharo launcher. I'm now
> looking for something that already exists outside so I don't have to
> extract this into a new library by myself.
> 
> -- 
> Damien Cassou
> http://damiencassou.seasidehosting.st
> 
> "Success is the ability to go from one failure to another without
> losing enthusiasm." --Winston Churchill
> 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] About GTSpotter matching

[Christophe Demarey]

Hi,

By the way, is it possible to have exact match now?
At least, I expect to have expect match on the top of the result list.
A simple use case,
open Spotter
search number
dive into implementors category

The exact matches are lost in the middle of hundreds of other selectors. This 
way, it is very difficult to find what you need.
It is also why I always need to open a playground to be able to search all 
implementors of a selector. It is not do-able with spotter without an exact 
match.

That said, I have to add that I really enjoy spotter and other GT tools :)

Christophe


Le 8 déc. 2015 à 01:56, Juraj Kubelka a écrit :

> 
>> 7. 12. 2015 v 11:59, Peter Uhnak :
>> 
>> On 12/07, Juraj Kubelka wrote:
>>> Hi,
>>> 
>>> # is a category filter. Try #class, #instance, etc.
>> 
>> Oh... right. I've been using this for long time, my brain just didn't
>> connect the dots.
>> 
>> In either case, once you dive in the category filter is no longer
>> applicable.
>> So normally I would do "#i selector", then dive in, and then filter it.
>> 
>>> Then I have learnt that people are not aware of [...] any other kind of 
>>> wild-characters.
>> 
>> People don't know what wild-chars are? I would understand that someone
>> might be uncomfortable with regexps, because there are many variations, but 
>> wildchars…
> 
> Well, some people asks for regular expressions, some people asks for 
> wild-characters, some people prefers other techniques. 
> In most cases people are satisfied with substring solution as it is right 
> now. In some special cases people thinks about more advance solution. 
> I believe that we should sort results according to relevance, e.g., if I 
> write open, then selectors called open should be first, then likely openOn:, 
> openWithSpec:, openVeryLongExplanation:, etc.
> 
> Cheers,
> Juraj
> 
>> 
>>> 
 On Dec 5, 2015, at 20:40, Peter Uhnak  wrote:
 
 Hi,
 
 are there some wildcards in GTSpotter matching?
 
 Currently it searches anywhere in the (method) name, which makes it hard
 for shorter names, because it will match a lot of junk.
 
 I've also discovered (by accident), that I can use '>>#selector' to
 anchor the start of the selection. ('#selector' for some reason doesn't
 work).
 But I would like to also search by a simple ? (any character), * (any
 characters) wildcard. Is that possible?
 
 Additionally constraining it from the end would be also nice.
 For example I want to look through #default methods, however 90% of the
 matches will be junk, so I would like to write '#default$' and it will
 not match '#defaultIcon', etc.
 
 Is this possible?
 
 Thanks,
 -- 
 Peter
 
>>> 
>>> 
>> 
>> -- 
>> Peter
>> 
> 
> 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] When a metaclass is initialised by the system ?

[Christophe Demarey]

yes, we miss a package initialize method ...

Le 8 déc. 2015 à 16:35, Mariano Martinez Peck a écrit :

> Dimitris,
> 
> Relying in class side initialize is not very cool. Mostly, because it's hard 
> to know EXACTLY when Monticello will send it. For sure it's the first time 
> you load that class. But then it could be if such method changes again. But 
> maybe too if you clear some Monticello caches...
> Also, you may have dependencies (of order execution) with other class side 
> initialize.  So...another approaches you can take (depending on your needs) 
> is lazy class var getters with ifNil:   or  if you have a 
> ConfigurationOfYourApp, you can define #postLoadDoIts in which you can 
> perform the whole initialization of your app. Or ...from the #postLoadDoIts 
> simply delegate to a class MyAppInitialize.
> 
> I do a bit of all of them hahaha.
> 
> Cheers,
> 
> On Mon, Dec 7, 2015 at 9:11 PM, Dimitris Chloupis  
> wrote:
> ok that means  I cannot rely on it and I will have to initialize my class 
> variables in specific methods. No problem, ifNil here I come :) 
> 
> On Tue, Dec 8, 2015 at 12:24 AM Sven Van Caekenberghe  wrote:
> Hi Dimitris,
> 
> > On 07 Dec 2015, at 23:12, Dimitris Chloupis  wrote:
> >
> > I have read that a metaclass (class side of the class) is initialised when 
> > its loaded to the image by monticello, Are there any other cases when the 
> > initialize method of the the metaclass is being called ?
> 
> Correct: a class #initialize is called when its code is loaded, but only if 
> it is not yet present or different in source code from what is already 
> present (this last point will bite you one day ;-).
> 
> Else, the #initialize should be called manually (for example to reload some 
> caches, or reset some system state). This sometimes happens in #postLoads or 
> install scripts.
> 
> I am not aware of any automatic invocations.
> 
> Sven
> 
> 
> 
> 
> 
> 
> 
> -- 
> Mariano
> http://marianopeck.wordpress.com



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] [Pharo-dev] Pharo-contribution down

[Christophe Demarey]

yes, a problem on the CI infrastructure.
IT team told me it will be fixed tomorrow

Le 2 janv. 2016 à 18:51, Esteban Lorenzano a écrit :

> everything is down
> I think we need to wait ‘till monday to fix it :(
> 
> Esteban
> 
>> On 02 Jan 2016, at 18:39, Damien Cassou  wrote:
>> 
>> The Jenkins for pharo-contribution seems to be down: builds never
>> finish.
>> 
>> https://ci.inria.fr/pharo-contribution
>> 
>> -- 
>> Damien Cassou
>> http://damiencassou.seasidehosting.st
>> 
>> "Success is the ability to go from one failure to another without
>> losing enthusiasm." --Winston Churchill
>> 
> 
> 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versionner and package dependencies

[Christophe Demarey]

Hi Cyril,

Le 30 déc. 2015 à 10:17, Ferlicot D. Cyril a écrit :

> Hi,
> 
> I wanted to use versionner to manage some existing configurations but in
> one of these configuration there is some dependencies on packages that
> have no project.
> 
> For example in the baseline there is:
> 
>   spec package: 'Glamour-EyeSee-Presentations'
>   with: [ spec requires: #('EyeSee').
>   spec repository: 
> 'http://smalltalkhub.com/mc/Moose/Glamour/main' ].
> 
> But in versionner I can only add a project as dependency.
> 
> Is there a way to do such a thing or is it a future feature of versionner?

It is not something we want in Versionner.
An external project must have a configuration to refer to it. If not, you will 
take random packages without knowing if they have themselves dependencies.
What you should do is to create a project for that and then ask Versionner to 
only load the package you need (right-click on the project dependency).

Christophe

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Versionner and package dependencies

[Christophe Demarey]

Le 4 janv. 2016 à 10:52, Ferlicot D. Cyril a écrit :

> Le 04/01/2016 10:44, Christophe Demarey a écrit :
>> Hi Cyril,
>> 
>> It is not something we want in Versionner.
>> An external project *must* have a configuration to refer to it. If not,
>> you will take random packages without knowing if they have themselves
>> dependencies.
>> What you should do is to create a project for that and then ask
>> Versionner to only load the package you need (right-click on the project
>> dependency).
>> 
>> Christophe
> 
> Hi,
> 
> So we have to create ourself a configuration for each project that
> doesn't have a configuration/group that match our needs if we want to
> use versionner if I understand right. Ok, thank you.

only if they do not have a configuration. If you need some packages or groups, 
you can specify them in the "loads" directive. 


> 
> And is there a way to execute pre/post scripts (#postLoadDoIt: in
> ConfigurationOfXXX) with versionner ? If not is it something you want to
> add to versionner or this will not be add ?

It is Metacello that execute these scripts. Versionner only allows you to edit 
configurations. The load options in Versionner use Metacello, so they are 
executed.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] CI build stuck?

[Christophe Demarey]

Le 23 déc. 2015 à 00:32, Johan Fabry a écrit :

> Hi all,
> 
> I noticed that since some days the CI builds for LRP are stuck, they take 
> forever and don’t do anything. I thought it was aCI issue and the dev team 
> was aware of this and it might clear itself up after some time but apparently 
> it’s not the case. So this is to give a heads up, just in case …
> 
> URL of the project is 
> https://ci.inria.fr/pharo-contribution/job/LiveRobotProgramming/
> 
> Console output of current build (Started 21 hr ago): 
> Started by timer
> Building remotely on 
> rmod-mac302.lille.inria.fr
> (32 mac) in workspace 
> /Users/jenkins-pharo/Documents/ci.inria.fr/pharo-contribution/workspace/LiveRobotProgramming
> 
> Deleting project workspace... done
> 
> Triggering 
> 50,development,vm
> 
> Configuration 
> 50,development,vm is still in the queue: Waiting for next available executor

The message is clear: there is no executor (slave) to run the job.
Indeed, all slaves were down.
I restarted them. It should be better now.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Basic versioning of GitHub repositories from within Pharo

[Christophe Demarey]

Hi,

Le 29 nov. 2015 à 20:00, Dimitris Chloupis a écrit :

> And there lies the trap.
> 
> If you end up making something that works with everything, you will create 
> something that just works with everything instead of something that works 
> very well with one thing.

could be. But we do not want to plug everything. Just having the door opened 
for the next cool VCS.
Even if is only used for git it is good to have a Pharo API that is not the GIT 
API. We do not want git calls mixed everywhere. We want a clean and clear API 
for versioning.
This API also prevents you from git API changes. You will only have one place 
to update and so on.

> Right now Git is by very far the undisputed king of version control and has 
> completely dethroned SVN. 

King because widely used but other are also as valuable if not more.

> Ironically the things that make git integration in pharo so hard is all the 
> thing that are there to decouple it from git , like filetree metadata , or 
> continuous support for mcz and monticello. 

I do not agree on this point. There is no VCS interoperability layer in Pharo. 
Tools are directly manipulating Monticello.

> In the end you cant have your cake and eat it too. We will have to choose 
> between very good integration with git or average / mediocre integration with 
> multiple backends. 

again, I  cannot agree. It is a bad idea to have calls to an external tool / 
library wide spread in the image. We need our own API. It is particularly true 
for git that has an API mixing daily tasks with other power admin commands 
(rewrite of the history with possible code loss).
I do not think we need full support of git in the image, i.e. supporting all 
commands and options, but only what is needed to do correct versioning from 
Pharo (special git commands can still be done with command-line).

> 
> On Sun, Nov 29, 2015 at 8:15 PM stepharo  wrote:
> What I would like for Pharo is to avoid to be bound to a given back-end
> for its versionning.
> This master is a step in the right direction
> 
> http://www.hpi.uni-potsdam.de/fileadmin/hpi/source/Technische_Berichte/HPI_54.pdf
> 
> Stef
> 
> 



smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] Understanding the role of the sources file

[Christophe Demarey]

Le 13 janv. 2016 à 10:57, Dimitris Chloupis a écrit :

> I was adding a short description to the UPBE about sources file , I always 
> thought that the sources file is the file that contains the source code of 
> the image because the image file itself stores only the bytecode. 
> 
> However its just came to my attention that the sources file does not contain 
> code that is recently installed in the image.
> 
> So how exactly the sources file works and what it is ? 


If I understood well:
- the sources file has all the source code of the released pharo version (when 
we generate this new source file PharoVXX.sources).
- the changes file has the source code since the release.

smime.p7s
Description: S/MIME cryptographic signature

Re: [Pharo-users] One comment to new Pharo5 debugger

[Christophe Demarey]

Hi Doru,

I think most people understand the great value of moldable tools. Maybe not all 
will use moldability but is a big plus to have. So, keep pushing!

I just think we miss time to make the GT Debugger dealt view  as productive as 
the spec debugger. There are users that do not use shortcuts (at least for 
debugger) and in this case, the over use of the mouse to reach step buttons is 
killing the productivity.
To me, this issue could be addressed easily but could take time to implement in 
a generic / moldable way.

I agree with Gabriel that buttons act on the stack (in fact also, on the source 
code pane that represents the AST) but what we mainly look at is the source 
code. Maybe buttons could be moved to the source code pane?

I also have another suggestion: I very often have a debugger window with a lot 
of variables not visible by default : only 4 visible, need to scroll. Couldn’t 
we split the variable pane in 2 part to display more variables? To open the 
inspector, you could either use the second column or open a third one. WDYT?

Cheers,
Christophe


> Le 4 juin 2016 à 08:45, Tudor Girba  a écrit :
> 
> Hi Sabine (and others that asked this question),
> 
> Thanks for the feedback. Sorry for the delayed reply, but I am traveling 
> these days and I have limited online time. Also, this email took some time to 
> write (I started it about a week ago).
> 
> I will answer the buttons issue. The reason they are in a different place 
> than where they used to be is related to new constraints that come with the 
> new debugger. While the default interface looks like a regular debugger, the 
> main feature of the GTDebugger is that it can be customized to match a 
> specific domain or library.
> 
> Let me provide an example of a scenario of debugging PetitParser:
> 
>   PPArithmeticParser parse: '1+2*3^(34-2)’
> 
> The default debugger looks like this:
> 
> 
> 
> But, because we have PetitParser code on the stack, we can switch to a custom 
> debugger:
> 
> 
> 
> This one also features the input stream next to the code. Also, note that the 
> stack received 3 new button actions. But, this is not all. Next to the 
> source, there are other tabs, and we can switch to one of those:
> 
> 
> 
> This one shows the source as well, but in a graphical form.
> 
> So, if we step back and talk about the buttons, you will see that if we would 
> have put the buttons on the source tabs, you have had no way to advance in 
> this state. We could have also duplicated behavior and put the buttons on the 
> new presentation as well, but the space is much smaller and since we now have 
> large buttons instead of only icons we needed to deal with that as well. We 
> encountered similar situations in other debuggers (not all, though).
> 
> But, from a more point of view, the actions act on the stack. So, if we take 
> an object-oriented approach for designing the UI, it makes sense to place 
> those actions next to the context they act on. All these reasons made us 
> choose to put them next to the stack and not next to the source.
> 
> We certainly understand that this is a departure from the classic debugger 
> and perhaps this is not the best way of doing things, but that was the 
> rationale we went through. I would be happy to hear other suggestions, but 
> please understand that we need to take into account the full spectrum of 
> constraints when choosing solutions.
> 
> In any case, the idea of this debugger is that you can change it in the way 
> you want and customize it specifically for your needs. We still need to 
> document how to do that, but if someone wants to start creating a custom 
> debugger we could use the experience to drive the documentation.
> 
> Does this explain the situation?
> 
> Cheers,
> Doru
> 
> 
>> On May 25, 2016, at 9:43 PM, Sabine Manaa > > wrote:
>> 
>> Hi,
>> 
>> today I moved to Pharo5. It is great to get feedback and solutions for
>> problems so quickly. It is a pleasure to work with Pharo and to have the
>> community. I am grateful for having this. 
>> 
>> I have one comment. I don't want to complain, I can use it but I was
>> wondering about the new position and size of the debugger buttons.
>> 
>> There is a longer distance now from code text pane to the buttons - I see
>> this as disadvantage to earlier versions of Pharo from the user experience.
>> The buttons are a lot of smaller - same disadvantage - both if you use the
>> mouse.
>> 
>> Ok, after seeing that, I thought "Sabine, you should use keyboard shortcuts
>> here, too" (I use much of them but interesting - til now not in the
>> debugger).
>> 
>> Then I saw the shortcuts for
>> Proceed cmd+r
>> Restart cmd-shift-a
>> Into cmd-e
>> over cmd-shift-o
>> through cmd-shift-t
>> 
>> especially into, over and through have so different key combinations, two
>> with shift, one without shift. When I debug, I often use for example
>> 

Re: [Pharo-users] library for comparing semver versions

[Christophe Demarey]

Hi Peter,

> Le 27 mai 2016 à 14:21, Peter Uhnak  a écrit :
> 
> Hi,
> 
> is there any library available that can compare versions according to
> SemVer? http://semver.org/

You can check the SemanticVersion class from CargoPackageManager-Minimal at 
http://smalltalkhub.com/#!/~demarey/CargoPackageManager.

The comparison is not fully complete for the suffix part but can be added

Christophe

Re: [Pharo-users] One comment to new Pharo5 debugger

[Christophe Demarey]

Hi Sabine,

this point was also « discussed » (no answer) here: 
http://forum.world.st/gtdebugger-debug-actions-buttons-td4874316.html 

Yes, the new buttons are counter-productive and there is even no setting to 
have them usable …
You can still with to the SpecDebugger through settings


> Le 25 mai 2016 à 22:49, Peter Uhnák  a écrit :
> 
> Hi,
> 
> yes, this question was raised before, but there are some disagreements over 
> how it should be, so it may take some time 
> http://forum.world.st/GTDebugger-shortcuts-usability-td4890316.html 
> 
> 
> You can also use a startup script to reassign them, such as this one. 
> https://github.com/peteruhnak/pharo-scripts/blob/master/config/5.0/uniformDebugger.st
>  
> 
> But a startup script is a temporary solution.
> 
> On Wed, May 25, 2016 at 9:43 PM, Sabine Manaa  > wrote:
> Hi,
> 
> today I moved to Pharo5. It is great to get feedback and solutions for
> problems so quickly. It is a pleasure to work with Pharo and to have the
> community. I am grateful for having this.
> 
> I have one comment. I don't want to complain, I can use it but I was
> wondering about the new position and size of the debugger buttons.
> 
> There is a longer distance now from code text pane to the buttons - I see
> this as disadvantage to earlier versions of Pharo from the user experience.
> The buttons are a lot of smaller - same disadvantage - both if you use the
> mouse.
> 
> Ok, after seeing that, I thought "Sabine, you should use keyboard shortcuts
> here, too" (I use much of them but interesting - til now not in the
> debugger).
> 
> Then I saw the shortcuts for
> Proceed cmd+r
> Restart cmd-shift-a
> Into cmd-e
> over cmd-shift-o
> through cmd-shift-t
> 
> especially into, over and through have so different key combinations, two
> with shift, one without shift. When I debug, I often use for example
> into-into-over-over-over-throuhg-through--into-into etc...and with the
> combinations cmd-e and cmd-shit-e, it is more "finger-work". It would be
> better to have all those shortcuts with or all shortcuts without shift. And
> the letters not so far away at the keyboard (e-o-t)
> 
> This is only my opinion and my first impression after one day with Pharo5. I
> am sure, after a few days, I forgot that this was uncomfortable because I
> have the shortcuts coming automatically into my fingers.
> 
> Regards
> Sabine
> 
> 
> 
> 
> 
> 
> --
> View this message in context: 
> http://forum.world.st/One-comment-to-new-Pharo5-debugger-tp4897390.html 
> 
> Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
> 
> 

Re: [Pharo-users] File Policy

[Christophe Demarey]

> Le 22 juin 2016 à 14:05, Valentin Ryckewaert  
> a écrit :
> 
> Hello everyone,
> 
> i'm currently working on making Pharo able to manage if it writes on his 
> files.I created a class which can define it, then  I'm already able, for 
> example, to define if Pharo writes on pharo.changes or not.
> 
> The meaning of this is to make Pharo able to execute a script but without 
> being modified by it, a script shouldn't let any trace of its execution on 
> the image or on Pharo files.
> 
> My problem concern PharoDebug.log, I would like to know what you prefer Pharo 
> to do with it in the case we configured it to not write on his files 
> (changes, source,image).
> Should Pharo writes the logs in the terminal ? in a tmp file ? or just in 
> PharoDebug.log ?

If you want to use Pharo as a script interpreter, then just do like other 
interpreter: report results and errors on stdout / stderr.
Then, you can redirect the output to also have it in a file if you want. I 
would not use PharoDebug.log or any other file.

Christophe

Re: [Pharo-users] #inform: Changed in 5.0?

[Christophe Demarey]

Hi Sean,

> Le 17 juin 2016 à 17:10, Sean P. DeNigris  a écrit :
> 
> During startup, I do `UIManager default inform: aString`. It used to show a
> Growl morph, but now opens a debugger. Did something change? If so, why and
> how do I circumvent the debugger? Thanks.

With the introduction of the session manager, error handling during startup 
actions was revisited.
You can check StartupUIManager>>#handleError:log: for errors coming before 
Morphic is initialized.
Once Morphic is initialized, you get a default UI manager (MorphicUIManager if 
in UI mode). It is managed by UIManagerSessionHandler class.

If your startup code is executed with StartupUIManager, it will add a deferred 
startup action to execute the inform, else it fallback to MorphicUIManager 
inform.
It should do the behavior you expect.

Could you give me more info (the stacktrace would be useful)? 

Cheers,
Christophe

Re: [Pharo-users] Pharo Launcher

[Christophe Demarey]

> Le 17 juin 2016 à 12:00, Serge Stinckwich <serge.stinckw...@gmail.com> a 
> écrit :
> 
> On Fri, Jun 17, 2016 at 10:22 AM, Christophe Demarey
> <christophe.dema...@inria.fr> wrote:
>> Hi,
>> 
>> I just updated Pharo Launcher. A new version 0.2.11 is available.
>> Here are the change log:
> 
> Great ! Thank you for the update.
> 
>> move Pharo Launcher to the latest stable pharo image: pharo 5.0. It implies
>> That Pharo Launcher is now shipped with Spur VM.
>> allow an easy download of the new beta image: pharo 6.0
>> the Launcher now comes with 2 VMs:
>> 
>> Spur VM (pharo 5.0 and Pharo 6.0)
>> Non-Spur VM (to be able to run old images)
> 
> I dl the mac version. I was not able to run non-spur VM images: the
> launcher crash or do nothing ...

Indeed, I think it’s a side effect of latest changes I did after testing on 
Windows.
I take a look at it now

Re: [Pharo-users] Pharo Launcher

[Christophe Demarey]

Now fixed. I replaced the DMG on files.pharo.org.
Thanks for the report

> Le 17 juin 2016 à 13:34, Christophe Demarey <christophe.dema...@inria.fr> a 
> écrit :
> 
>> 
>> Le 17 juin 2016 à 12:00, Serge Stinckwich <serge.stinckw...@gmail.com> a 
>> écrit :
>> 
>> On Fri, Jun 17, 2016 at 10:22 AM, Christophe Demarey
>> <christophe.dema...@inria.fr> wrote:
>>> Hi,
>>> 
>>> I just updated Pharo Launcher. A new version 0.2.11 is available.
>>> Here are the change log:
>> 
>> Great ! Thank you for the update.
>> 
>>> move Pharo Launcher to the latest stable pharo image: pharo 5.0. It implies
>>> That Pharo Launcher is now shipped with Spur VM.
>>> allow an easy download of the new beta image: pharo 6.0
>>> the Launcher now comes with 2 VMs:
>>> 
>>> Spur VM (pharo 5.0 and Pharo 6.0)
>>> Non-Spur VM (to be able to run old images)
>> 
>> I dl the mac version. I was not able to run non-spur VM images: the
>> launcher crash or do nothing ...
> 
> Indeed, I think it’s a side effect of latest changes I did after testing on 
> Windows.
> I take a look at it now

[Pharo-users] Smalltalkhub scheduled maintenance next monday (February, 1st)

[Christophe Demarey]

Hello,

Smalltalk will be down next monday (february, 1st) from 12:30 to 13:30 (french 
local time).
IT will add storage to the root partition.

Regards,
Christophe

[Pharo-users] Re : Re: [Pharo-dev] help needed for testing the order in spotter

[Christophe Demarey]

Hi Doru,

I don't see a better way to have beta-testers than having your code on the 
pharo alpha image ;)
You will have very soon feedback and then you can iterate.

Cheers,
Christophe

Re: [Pharo-users] Smalltalkhub scheduled maintenance today (February, 22nd)

[Christophe Demarey]

Smalltalkhub is now back on line!

Le 22 févr. 2016 à 10:59, Christophe Demarey a écrit :

> Hello,
> 
> Smalltalk will be down today (february, 22nd) from 12:30 to 13:30 (french 
> local time).
> IT will add storage to the root partition (last maintenance finally did not 
> take place).
> 
> Regards,
> Christophe

Re: [Pharo-users] Pb with the CI @ INRIA?

[Christophe Demarey]

Thanks for the link.

The ugly java stack trace is because the junit result publisher jenkins plugin 
tries to parse a file that is not well generated.
It happens when the image crashes during tests: the xml output does not contain 
end tags.
As Aliaksei sait, probably dut to the cairo lib missing. I will check.


Le 26 févr. 2016 à 15:14, Aliaksei Syrel a écrit :

> It is because image crashed without saving logs. More likely slave used to 
> build has no cairo installed.
> 
> Cheers,
> Alex
> 
> On Fri, Feb 26, 2016 at 2:47 PM, Christophe Demarey 
> <christophe.dema...@inria.fr> wrote:
> Hi Alex,
> 
> Could you provide the link to the job providing this output?
> 
> Le 26 févr. 2016 à 09:57, Alexandre Bergel a écrit :
> 
> > Hi!
> >
> > We regularly get this error:
> >
> > Failed to read test report file 
> > /builds/workspace/roassal2/Roassal2GT-Test.xml
> > org.dom4j.DocumentException: Error on line 27 of document 
> > file:///builds/workspace/roassal2/Roassal2GT-Test.xml : XML document 
> > structures must start and end within the same entity. Nested exception: XML 
> > document structures must start and end within the same entity.
> >   at org.dom4j.io.SAXReader.read(SAXReader.java:482)
> >   at org.dom4j.io.SAXReader.read(SAXReader.java:264)
> >   at hudson.tasks.junit.SuiteResult.parse(SuiteResult.java:123)
> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:286)
> >   at 
> > hudson.tasks.junit.TestResult.parsePossiblyEmpty(TestResult.java:232)
> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:167)
> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:150)
> >   at hudson.tasks.junit.TestResult.(TestResult.java:126)
> >   at 
> > hudson.tasks.junit.JUnitParser$ParseResultCallable.invoke(JUnitParser.java:132)
> >   at 
> > hudson.tasks.junit.JUnitParser$ParseResultCallable.invoke(JUnitParser.java:107)
> >   at hudson.FilePath$FileCallableWrapper.call(FilePath.java:2691)
> >   at hudson.remoting.UserRequest.perform(UserRequest.java:120)
> >   at hudson.remoting.UserRequest.perform(UserRequest.java:48)
> >   at hudson.remoting.Request$2.run(Request.java:326)
> >
> > Any idea why?
> >
> > Alexandre
> > --
> > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> > Alexandre Bergel  http://www.bergel.eu
> > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
> >
> >
> >
> >
> 
> 
> 

Re: [Pharo-users] Pb with the CI @ INRIA?

[Christophe Demarey]

For this job, the image crashes because of an OutOfMemory as you can check 
here: https://ci.inria.fr/moose/job/roassal2/2253/console

Le 26 févr. 2016 à 16:01, Christophe Demarey a écrit :

> Thanks for the link.
> 
> The ugly java stack trace is because the junit result publisher jenkins 
> plugin tries to parse a file that is not well generated.
> It happens when the image crashes during tests: the xml output does not 
> contain end tags.
> As Aliaksei sait, probably dut to the cairo lib missing. I will check.
> 
> 
> Le 26 févr. 2016 à 15:14, Aliaksei Syrel a écrit :
> 
>> It is because image crashed without saving logs. More likely slave used to 
>> build has no cairo installed.
>> 
>> Cheers,
>> Alex
>> 
>> On Fri, Feb 26, 2016 at 2:47 PM, Christophe Demarey 
>> <christophe.dema...@inria.fr> wrote:
>> Hi Alex,
>> 
>> Could you provide the link to the job providing this output?
>> 
>> Le 26 févr. 2016 à 09:57, Alexandre Bergel a écrit :
>> 
>> > Hi!
>> >
>> > We regularly get this error:
>> >
>> > Failed to read test report file 
>> > /builds/workspace/roassal2/Roassal2GT-Test.xml
>> > org.dom4j.DocumentException: Error on line 27 of document 
>> > file:///builds/workspace/roassal2/Roassal2GT-Test.xml : XML document 
>> > structures must start and end within the same entity. Nested exception: 
>> > XML document structures must start and end within the same entity.
>> >   at org.dom4j.io.SAXReader.read(SAXReader.java:482)
>> >   at org.dom4j.io.SAXReader.read(SAXReader.java:264)
>> >   at hudson.tasks.junit.SuiteResult.parse(SuiteResult.java:123)
>> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:286)
>> >   at 
>> > hudson.tasks.junit.TestResult.parsePossiblyEmpty(TestResult.java:232)
>> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:167)
>> >   at hudson.tasks.junit.TestResult.parse(TestResult.java:150)
>> >   at hudson.tasks.junit.TestResult.(TestResult.java:126)
>> >   at 
>> > hudson.tasks.junit.JUnitParser$ParseResultCallable.invoke(JUnitParser.java:132)
>> >   at 
>> > hudson.tasks.junit.JUnitParser$ParseResultCallable.invoke(JUnitParser.java:107)
>> >   at hudson.FilePath$FileCallableWrapper.call(FilePath.java:2691)
>> >   at hudson.remoting.UserRequest.perform(UserRequest.java:120)
>> >   at hudson.remoting.UserRequest.perform(UserRequest.java:48)
>> >   at hudson.remoting.Request$2.run(Request.java:326)
>> >
>> > Any idea why?
>> >
>> > Alexandre
>> > --
>> > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>> > Alexandre Bergel  http://www.bergel.eu
>> > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>> >
>> >
>> >
>> >
>> 
>> 
>> 
> 

[Pharo-users] Smalltalkhub scheduled maintenance today (February, 22nd)

[Christophe Demarey]

Hello,

Smalltalk will be down today (february, 22nd) from 12:30 to 13:30 (french local 
time).
IT will add storage to the root partition (last maintenance finally did not 
take place).

Regards,
Christophe

Re: [Pharo-users] Moose-Algos-Graph Library in Pharo 5

[Christophe Demarey]

Hi,

Moose-Algos-Graph contains very classical algorithmes applying on graphes 
(Breadth-first search, dijkstra, tarjan, topological sorting, etc).
It is now in Pharo 5 because it is used by the dependency analyser.
As Alexandre said, there are almost no documentation and it is not so easy to 
combine algorithms (run an algo on the output of another algo).
You can look at Moose-Tests-Algos-Graph package (to be loaded) and 
DADependencyChecker>>#shortestPathToPackageIntroducingDependency:startingFrom: 
for some examples.

Christophe


> Le 24 avr. 2016 à 20:55, Alexandre Bergel  a écrit :
> 
> Hi Volkert!
> 
> This package contains some classical graph algorithms we use in software 
> analysis (e.g., clustering, strong component).
> Unfortunately, there is no much documentation about it. If you are using 
> these algorithms, then it would be great to write some documentation.
> 
> In addition to this package, and as you know, Roassal offers many layout 
> algorithms.
> 
> Regarding,
> Alexandre
> 
>> On Apr 24, 2016, at 6:39 AM, Volkert  wrote:
>> 
>> Dear all,
>> 
>> if have found the "Moose-Algos-Graph" package in Pharo 5. What is the idea 
>> behind this package? Is this "the" official Pharo Graph Library?
>> Do we have some examples how to use it?
>> 
>> LG,
>> Volkert
>> 
> 
> -- 
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel  http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
> 
> 
> 
> 

Re: [Pharo-users] Moose-Algos-Graph Library in Pharo 5

[Christophe Demarey]

Metacello new
configuration: 'MooseAlgos';
smalltalkhubUser: 'Moose' project: 'MooseAlgos';
version: #bleedingEdge;
load: 'Moose-Tests-Algos-Graph'

> Le 25 avr. 2016 à 18:52, Volkert <volk...@komponentenwerkstatt.de> a écrit :
> 
> Christophe, Alexandre, i will give it a try  How can i load 
> "Moose-Tests-Algos-Graph" into Pharo 5.0?
> 
> Thanks, Volkert
> 
> On 25.04.2016 10:23, Christophe Demarey wrote:
>> Hi,
>> 
>> Moose-Algos-Graph contains very classical algorithmes applying on graphes 
>> (Breadth-first search, dijkstra, tarjan, topological sorting, etc).
>> It is now in Pharo 5 because it is used by the dependency analyser.
>> As Alexandre said, there are almost no documentation and it is not so easy 
>> to combine algorithms (run an algo on the output of another algo).
>> You can look at Moose-Tests-Algos-Graph package (to be loaded) and 
>> DADependencyChecker>>#shortestPathToPackageIntroducingDependency:startingFrom:
>>  for some examples.
>> 
>> Christophe
>> 
>> 
>>> Le 24 avr. 2016 à 20:55, Alexandre Bergel <alexandre.ber...@me.com> a écrit 
>>> :
>>> 
>>> Hi Volkert!
>>> 
>>> This package contains some classical graph algorithms we use in software 
>>> analysis (e.g., clustering, strong component).
>>> Unfortunately, there is no much documentation about it. If you are using 
>>> these algorithms, then it would be great to write some documentation.
>>> 
>>> In addition to this package, and as you know, Roassal offers many layout 
>>> algorithms.
>>> 
>>> Regarding,
>>> Alexandre
>>> 
>>>> On Apr 24, 2016, at 6:39 AM, Volkert <volk...@komponentenwerkstatt.de> 
>>>> wrote:
>>>> 
>>>> Dear all,
>>>> 
>>>> if have found the "Moose-Algos-Graph" package in Pharo 5. What is the idea 
>>>> behind this package? Is this "the" official Pharo Graph Library?
>>>> Do we have some examples how to use it?
>>>> 
>>>> LG,
>>>> Volkert
>>>> 
>>> -- 
>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>> Alexandre Bergel  http://www.bergel.eu
>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>> 
>>> 
>>> 
>>> 
>> 
> 
> 

Re: [Pharo-users] Image not starting

[Christophe Demarey]

I would try to run pharo through strace (http://linux.die.net/man/1/strace 
).
This way, you will probably see a system call trying to write to a file in 
read-only mode.

Hope this helps.
Christophe

> Le 12 juil. 2016 à 15:33, Valentin Ryckewaert  
> a écrit :
> 
> Hello everyone,
> 
> I'm currently working on the project file policy and I'm having a problem 
> that I don't understand
> 
> I import my code, configure my image as readOnly (Pharo won't try to write on 
> changesFiles,source,image), set the folder of the image in readonly and 
> try to open the image again and it's not possible, I don't get any stacktrace 
> so it's hard to understand the problem.
> 
> PS : If we set back the folder in write mode, it's possible again to open the 
> image again.
> 
> Here is the code to reproduce the bug, could someone give me a clue please?
> 
> sudo rm -r testPharo
> mkdir testPharo
> cd testPharo
> 
> echo "Metacello new 
>   smalltalkhubUser: 'Uryon' project: 'Coral';
>   configuration: 'Coral';
>   version: #development;
>   onConflictUseIncoming;
>   load.
>   SmalltalkImage configureSystemAsReadonlyAndQuit:true
> " >> load.st 
> 
> mkdir coral
> cd coral
> curl get.pharo.org/alpha  | bash 
> curl get.pharo.org/vm60  | bash 
> pharo-vm/pharo Pharo.image ../load.st  
> cd ..
> chmod -R -w coral
> coral/pharo-vm/pharo coral/Pharo.image
> 

Re: [Pharo-users] Pharo Launcher

[Christophe Demarey]

Hi Tim,

> Le 27 juin 2016 à 22:12, Tim Mackinnon <tim@testit.works> a écrit :
> 
> Hi guys - I finally got some free time to take a look at Pharo-5 and wanted 
> Launcher back as well - I just downloaded it on my mac and I’m wondering if 
> something is wrong?
> 
> The UI doesn’t look right (compared to the older 4.x version) - the top panel 
> is white, and there is a grey run button next to the green one (the triangle) 
> that seems to be missing transparency - so it all looks quite ugly. Have I 
> downloaded something that is wrong?

No, is the current state of Pharo Launcher UI. I did not touch to this part 
but, yes, could be improved.

> I also notice, that when I go to create a new image from a template - the 
> name prompt isn’t as good as the old version which used to include the parent 
> node in the suggested name so you got something like “Pharo 6.0 (beta) 
> 123455” as a descriptive name - it now just seems to offer “1234456” as the 
> name now?

I did this change after a small survey on pharo launcher users that were, like 
me, bothered with previous proposed names: in most cases, they only want the 
name of their project. That could be discussed but now, in your example, you 
get 60121 that is still descriptive. Here, we should find the best proposition 
for most users, will not for to all.

> I don’t understand the how the paths work now - it seems that my old images 
> launch fine, but a 5.x image doesn’t? Do I set the spur path for this - or is 
> it the other way around (vm-path should be the new vm)? The floating help 
> isn’t very descriptive on this.

Pharo launcher is now shipped with both vm (pre-spur and spur). The cog-spur vm 
now powered PharoLauncher and should be used to run spur images. If not, maybe 
you have a wrong setting for it. If you remove settings for both VM it should 
work. Maybe I should remove these settings since both VMs are now shipped with 
the launcher app. I do not know if it is relevant to some of you to use another 
VM than the default ones.

Thanks for the feedback.
Christophe

> 
> Thanks,
> 
> Tim
> 
> 
> 
> Sent from my iPhone
>> On 17 Jun 2016, at 13:10, Christophe Demarey <christophe.dema...@inria.fr> 
>> wrote:
>> 
>> Now fixed. I replaced the DMG on files.pharo.org.
>> Thanks for the report
>> 
>>>> Le 17 juin 2016 à 13:34, Christophe Demarey <christophe.dema...@inria.fr> 
>>>> a écrit :
>>>> 
>>>> 
>>>> Le 17 juin 2016 à 12:00, Serge Stinckwich <serge.stinckw...@gmail.com> a 
>>>> écrit :
>>>> 
>>>> On Fri, Jun 17, 2016 at 10:22 AM, Christophe Demarey
>>>> <christophe.dema...@inria.fr> wrote:
>>>>> Hi,
>>>>> 
>>>>> I just updated Pharo Launcher. A new version 0.2.11 is available.
>>>>> Here are the change log:
>>>> 
>>>> Great ! Thank you for the update.
>>>> 
>>>>> move Pharo Launcher to the latest stable pharo image: pharo 5.0. It 
>>>>> implies
>>>>> That Pharo Launcher is now shipped with Spur VM.
>>>>> allow an easy download of the new beta image: pharo 6.0
>>>>> the Launcher now comes with 2 VMs:
>>>>> 
>>>>> Spur VM (pharo 5.0 and Pharo 6.0)
>>>>> Non-Spur VM (to be able to run old images)
>>>> 
>>>> I dl the mac version. I was not able to run non-spur VM images: the
>>>> launcher crash or do nothing ...
>>> 
>>> Indeed, I think it’s a side effect of latest changes I did after testing on 
>>> Windows.
>>> I take a look at it now
> 
> 
> 

Re: [Pharo-users] Code loading/unloading order

[Christophe Demarey]

> Le 26 janv. 2017 à 12:25, webwarrior  a écrit :
> 
> Thanks.
> 
> Reason why I ask is because I ran into some issues with slots that depended
> on other classes in the package (when creating slot, like SomeSlot new
> valueClass: SomeOtherClass). I had to move slots into separate package to
> make it work.

probably because dependency computation done by Monticello does not take into 
account slots.
would be a good addition.

Re: [Pharo-users] Issue with Iceberg

[Christophe Demarey]

curl get.pharo.org/vmLatest60 | bash

> Le 17 févr. 2017 à 14:48, Alexandre Bergel  a écrit :
> 
> Getting the latest unstable VM is a bit mysterious. Where to get it?
> Is it available here:
> http://files.pharo.org/vm/pharo/mac/
> ?
> No idea
> 
> The latest VM on the files.pharo.org is tagged as 6.0-pre. I guess this is 
> not the latest unstable VM...
> 
> Alexandre
> 
>> On Feb 15, 2017, at 1:32 PM, Juraj Kubelka  wrote:
>> 
>> You likely need the latest (unstable) VM because the stable one has 
>> corrupted git library.
>> 
>> Cheers,
>> Juraj
>> 
>>> El 15-02-2017, a las 12:24, Alexandre Bergel  
>>> escribió:
>>> 
>>> Hi!
>>> 
>>> I have "Error: External module not found” when trying to clone a repository.
>>> I have the VM 7.0 (built on Oct 7, 2016) and Pharo 60371
>>> https://github.com/npasserini/iceberg/issues/281
>>> 
>>> Help appreciated.
>>> 
>>> Cheers,
>>> Alexandre
>>> 
>>> -- 
>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>> Alexandre Bergel  http://www.bergel.eu
>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>> 
>>> 
>>> 
>>> 
>> 
>> 
> 
> -- 
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel  http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
> 
> 
> 
> 

[Pharo-users] STON encoding of slashes

[Christophe Demarey]

Hi,

I just noticed that STON encoding of forward slashes changed.

STON toString: 'g...@github.com:foo/bar.git’ => ''g...@github.com:foo\/bar.git’'

It used to be ''g...@github.com:foo/bar.git’’.

Is it on purpose?

Thanks,
Christophe

Re: [Pharo-users] Iceberg and git workflow

[Christophe Demarey]

Hi Holger,

I think your scenario is not covered by Iceberg. It could be a nice feature but 
not one of the most important. Nicolas will answer but I think he focuses on 
most common scenarios.
In my case, what I miss is to be able to choose what I want to commit. Nicolas 
already planned to add this.

Christophe

> Le 12 sept. 2016 à 16:15, Holger Freyther  a écrit :
> 
> Hi,
> 
> I am a heavy git user with languages like C, C++, Python, Ruby and even GNU 
> Smalltalk and I hope iceberg will bring the same powerful experience to Pharo.
> 
> Last Friday I started to add a bigger refactoring for a new feature to my 
> software and didn't finish. Sadly today an issue in the code was found and I 
> would like to fix this before fixing my code. I use this as opportunity to 
> ask if Iceberg has some answers for that.
> 
> 
> With a non-Pharo project I would do:
> 
> a.) If current HEAD is same as origin/master
> 
> $ git stash (stash away my not finished changes)
> $ vi code.c fix..
> $ git commit -a -c "subject
> 
> long explanation of fix
> reference to bug"
> $ git stash apply (and go back to working on my feature)
> 
> 
> b.) E.g. if I finished n-commits but I am not fully done
> 
> # store my work
> $ git commit -a -m "Work In Progress hack.."
> $ git checkout -b new-feature-branch
> 
> # go back to master
> $ git checkout master 
> $ git reset --hard origin/master (to restore)
> 
> # work on the fix
> $ vi code.c fix..
> $ git commit -a -c "fix..."
> 
> # go back and continue on my fix
> $ git checkout new-feature-branch
> $ git rebase origin/master
> $ git reset HEAD^1
> .. continue to work
> 
> 
> 

Re: [Pharo-users] Iceberg and git workflow

[Christophe Demarey]

> Le 12 sept. 2016 à 22:20, Stephan Eggermont <step...@stack.nl> a écrit :
> 
> On 12/09/16 21:51, Christophe Demarey wrote:
>> I think your scenario is not covered by Iceberg. It could be a nice feature 
>> but not one of the most important.
> 
> Why do you think so?

In my mind, the top priorities are to have a UI allowing to easily:
- clone a repository
- sync with a remote
- a good diff and merge support
- an easy way to understand the project history (among all branches): log, 
visual graph of branches and versions
- choose what you want to commit (even if we do not use git index)
- easy switch to another tag, branch
(- and use of libgit instead of gitfiletree: important for windows users)

I see git stash usage less important than this one. But maybe it is a 
top-priority feature for some people heavily relying on it.
Anyway, the best way is to play/use Iceberg and give feedback to Nicolas about 
what is good, what could be improved / added. We just need to take care that 
Nicolas has limited time and, at some point, we need to take decisions.

Re: [Pharo-users] loading petitParser

[Christophe Demarey]

yes because the current way to load projects is:

Metacello new
  configuration: 'PetitParser';
  smalltalkhubUser: 'Moose' project: 'PetitParser';
  load.

if you want to load a specific group, you should use the #load: message 
(current groups are Core, Tests, Parser and  ParserIDE). Default will load all.

> Le 28 mars 2017 à 10:02, Stephane Ducasse  a écrit :
> 
> Hi 
> 
> in the deep into pharo book it is written
> 
> Gofer new
> 
> smalltalkhubUser: 'Moose' project: 'PetitParser'; package: 
> 'ConfigurationOfPetitParser';
> 
> load.
> 
> (Smalltalk at: #ConfigurationOfPetitParser) perform: #loadDefault
> 
> 
> 
> But this is not correct. There is no loadDefault. 
> 
> Should I change it? 
> 
> 
> 
> Stef
> 

Re: [Pharo-users] Pharo launcher on Linux Mint

[Christophe Demarey]

Hi Sebastian,

> Le 30 juil. 2017 à 18:12, Sebastian Heidbrink via Pharo-users 
>  a écrit :
> 
> 
> De: Sebastian Heidbrink 
> Objet: Pharo launcher on Linux Mint
> Date: 30 juillet 2017 18:12:07 UTC+2
> À: Any question about pharo is welcome 
> Hi!
> 
> I used Pharo launcher on Windows and Ubuntu before and never had any troubles.
> 
> Now, I am trying to used it on Linux Mint without success.
> 
> Once I start the freshly created image the appropriate VM is downloaded but 
> the image is not started.
> 
Which version do you use?
> A PharoDebug.log is created in the image folder complaining that it can't 
> find the PharoVXX.sources files in the "pharo/bin/lib/pharo/5.0-201707201942" 
> folder.
> 
> After I copy the missing file into that folder the image is still not opened 
> and no further logs are created or updated.
> 
I had this problem when updating the launcher because the downloaded zip file 
is unzipped by Pharo itself but ZipArchive does not preserve permissions 
attributes.
To fix that, I added PhLVirtualMachineManager>>#ensureIsExecutable:that is 
called right after the Launcher fetch a new VM. It does an OsProcess call to 
add executable permissions.
Has Pharo executable write permissions?

Christophe

[Pharo-users] get output of a forked process on windows

[Christophe Demarey]

Hi,

I would like to evaluate a Smalltalk expression in a forked process (i.e. 
something like OSProcess thisOSProcess waitForCommandOutput: 'pharo Pharo.image 
eval "SystemVersion current dottedMajorMinor"').
It is doable on linux/mac with OSProcess but not on Windows. Is there another 
way to do that on Windows? I would prefer to avoid to write to a file.

Thanks,
Christophe

Re: [Pharo-users] get output of a forked process on windows

[Christophe Demarey]

PharoConsole.exe is only available for image versions > 50.
I need a solution that works on older image versions.

Thanks for the advice,
Christophe

> Le 18 juil. 2017 à 14:26, Guillermo Polito <guillermopol...@gmail.com> a 
> écrit :
> 
> Have you checked the Console version of the windows VM?
> 
> On Tue, Jul 18, 2017 at 10:48 AM, Christophe Demarey 
> <christophe.dema...@inria.fr <mailto:christophe.dema...@inria.fr>> wrote:
> Hi,
> 
> I would like to evaluate a Smalltalk expression in a forked process (i.e. 
> something like OSProcess thisOSProcess waitForCommandOutput: 'pharo 
> Pharo.image eval "SystemVersion current dottedMajorMinor"').
> It is doable on linux/mac with OSProcess but not on Windows. Is there another 
> way to do that on Windows? I would prefer to avoid to write to a file.
> 
> Thanks,
> Christophe
> 
> 
> 
> -- 
>
> Guille Polito
> 
> Research Engineer
> French National Center for Scientific Research - http://www.cnrs.fr 
> <http://www.cnrs.fr/>
> 
> 
> Web: http://guillep.github.io <http://guillep.github.io/>
> Phone: +33 06 52 70 66 13

Re: [Pharo-users] get output of a forked process on windows

[Christophe Demarey]

Hi Thierry,

> Le 18 juil. 2017 à 15:40, Thierry Goubier  a écrit 
> :
> 
> Hi Christophe,
> 
> You have to use ProcessWrapper.
> 
> Metacello new
>   configuration: 'ProcessWrapper';
>   repository: 'http://smalltalkhub.com/mc/Pharo/MetaRepoForPharo40/main 
> ';
>   load

Thanks, it is working with the given example from 
http://smalltalkhub.com/#!/~hernan/ProcessWrapper/.

Best regards,
Christophe

[Pharo-users] Re : Re: STON Question

[Christophe Demarey]

+1
It is exactly what i've done with Cargo. Define your own specific object for 
serialization / deserialization. It is the simplest way to have a human 
readable conf file (e.g. urls, pathes default serialization is not what you 
expect)

Christophe
- Sven Van Caekenberghe  a écrit :
> STON cannot be used to write partial, manual output because it needs to see 
> the full object graph (like FUEL does).
> 
> If I would want to use STON for prettier looking config files, I would make 
> an object just for that external representation, converting it where 
> necessary to better internal objects (like from String to FileReference).
> 
> > On 5 Jun 2017, at 08:58, Stephane Ducasse  wrote:
> > 
> > Sven I know what Ston is :) 
> > 
> > Now Pillar dev used Ston as input and it works. Then with some hooks they 
> > let the user convert the strings
> > into specific objects and this is ok. 
> > Now I want to add export in Ston and for configurations I need to control 
> > for the application and not as a ston object the value saved.
> > For example a file reference should not be just a file reference in Ston 
> > but 
> > a string relative the the baseDirectory of the configuration. 
> > 
> > So I thought that I can build a hook and control the call to Ston (convert 
> > the object before to the correct values) and export. 
> > It should be working now doing this I found myself redoing the output of 
> > Ston. 
> > So I will do it. Then I thought that may be I could reuse this logic. 
> > 
> > Stef
> > 
> > 
> > 
> > On Sun, Jun 4, 2017 at 11:49 PM, Sven Van Caekenberghe  wrote:
> > Stef,
> > 
> > STON is like FUEL, it can write any object to a stream and read it back. 
> > Although you can customise how this is done, only one way of doing so is 
> > supported (i.e. you cannot change the format for each application).
> > 
> > People have used STON for various applications now, including for 
> > configuration stuff (because it is a textual format that is easy to read).
> > 
> > I would suggest you just try to see what your config looks like once you 
> > have the objects in Pharo.
> > 
> >   STON toStringPretty: myConfig.
> > 
> > Then we can see how to tune the representation.
> > 
> > Sven
> > 
> > > On 4 Jun 2017, at 22:58, Stephane Ducasse  wrote:
> > >
> > > Hi sven
> > >
> > > I'm working on a new configuration frameworks similar to the one use 
> > > pillar.conf.
> > >
> > > So basically I have
> > >
> > > ston := '{
> > >"newLine":#unix,
> > >   "separateOutputFiles":true
> > >   }'.
> > > and I want to recreate it.
> > >
> > > So I did
> > >
> > > Object >> toConfigurationString
> > >
> > >   ^ STON toString: self
> > >
> > > SimpleConfiguration >> exportStream
> > >
> > >   ^ String streamContents: [ :str |
> > >   str << '{' .
> > >   self propertiesKeysAndValuesDo: [ :key :value|
> > >   str << key toConfigurationString.
> > >   str << ':'.
> > >   str << value toConfigurationString.
> > >   str << ','.
> > >   str lf.
> > >   ].
> > >   str << '}'.
> > >   str contents
> > >   ]
> > >
> > > The idea is that for special configuration values people should be able 
> > > to define
> > > how to go from the object to configuration.
> > > For example for some FileReference I want to have just the path up to the 
> > > baseDirectory.
> > >
> > > Now I wonder if I could not better reuse ston.
> > > I saw the stonOn:
> > > but I did not look further because I was wondering how the writer where 
> > > passed.
> > >
> > > Any feedback is welcome.
> > >
> > > Stef (now bed time).
> > >
> > 
> > 
> > 
> 
> 

Re: [Pharo-users] (no subject)

[Christophe Demarey]

Hi, 

I published an update of the Launcher yesterday fixing some issues to run Pharo 
images from Linux. You can find latest-versions (0.2.13) here: 
http://files.pharo.org/platform/launcher/ 
It already propose to download Pharo 70 images. The VM is now downloaded 
automatically if the adequate one is not available. 
Let us know if everything is fine. 

Regards, 
Christophe 

- Mail original -

> De: "Vityou ." 
> À: pharo-users@lists.pharo.org
> Envoyé: Dimanche 13 Août 2017 10:31:06
> Objet: [Pharo-users] (no subject)

> I recently discovered the pharo launcher. However, it only gives me options
> to use outdated images, and I'm not sure about the vm. Is there any way to
> change this?

Re: [Pharo-users] (no subject)

[Christophe Demarey]

> Le 17 sept. 2017 à 05:21, Sean P. DeNigris  a écrit :
> 
> demarey wrote
>> Let us know if everything is fine. 
> 
> Was this issue [1] fixed:
>> I think I found it. It seems to be a bug in the way Launcher/Pharo unzips
>> the downloaded VM. 
>> The symlink libgit2.dylib becomes unlinked and appears as just a regular
>> file. If I delete the 
>> Pharo.app and unzip 70.zip (which btw maybe should have been deleted
>> during cleanup after 
>> the dl) to ./vms/70, the error disappears! 
> 
> [1] From
> http://forum.world.st/Launcher-Error-w-Pharo-7-tt4959777.html#a4960417

Not solved. 
Indeed, it is a problem of the Zip library that does not unzip properly (no 
preservation of permissions, does not handle symlinks).

It is a workaround but maybe I will add a strategy to use the unzip exe if I 
can find it on the system. That should solve most problems.
WDYT?

Re: [Pharo-users] (no subject)

[Christophe Demarey]

Hi Sean, Stephan,

I just updated the PharoLauncher and it should now work fine for 64-bits images.
I just had time to test it roughly but you can give it a try 
(https://ci.inria.fr/pharo/view/Launcher/job/Launcher/149/).

Christophe.


> Le 20 sept. 2017 à 10:52, stephan  a écrit :
> 
> On 20-09-17 08:26, Ben Coman wrote:
> 
>> Must be something environmental in lookup paths.  Can you reproduce or 
>> suggest further investigation I can do?
> 
> If I understand it correctly it is a problem with the unzipping. Not handling 
> symlinks and preserving permissions. Is that a plugin problem?
> 
> Stephan
> 
> 

Re: [Pharo-users] (no subject)

[Christophe Demarey]

Hi Ben, 

Could you test it with the bleedingEdge version of the launcher to see if the 
problem is solved?

Thanks,
Christophe.

> Le 20 sept. 2017 à 08:26, Ben Coman <b...@openinworld.com> a écrit :
> 
> 
> 
> On Wed, Sep 20, 2017 at 1:12 PM, Ben Coman <b...@openinworld.com 
> <mailto:b...@openinworld.com>> wrote:
> On Thu, Sep 14, 2017 at 8:25 PM, Christophe Demarey 
> <christophe.dema...@inria.fr <mailto:christophe.dema...@inria.fr>> wrote:
> > Hi,
> >
> > I published an update of the Launcher yesterday fixing some issues to run
> > Pharo images from Linux. You can find latest-versions (0.2.13) here:
> > http://files.pharo.org/platform/launcher/ 
> > <http://files.pharo.org/platform/launcher/>
> > It already propose to download Pharo 70 images. The VM is now downloaded
> > automatically if the adequate one is not available.
> > Let us know if everything is fine.
> 
> Thanks Christophe for updating PharoLauncher.  I'm trying it with Pharo 7 
> images for the first time.
> I hit a problem using Iceberg in downloaded images. 
> 
> $ unzip Pharo-linux-0.2.13.zip
> $ cd pharo
> $ unzip ../PharoLauncher-user-stable-2017.09.14.zip
> $ ./pharo
> + PharoLauncher > Settings > Enable development environment
> + World > Tools > Iceberg  
> ==>  iceberg opens okay
> + PharoLauncher > Templates > Pharo 7.0(beta) > latest-32  
> '7latest32'
> + '7latest32' 
> + World > Tools > Iceberg  
> ==>  "Error: EXternal module not found"
> 
> $ uname -a 
> ==> Linux 3.16.0-4-686-pae #1 SMP Debian 3.16.7-ckt25-2 (2016-04-08) i686 
> GNU/Linux
> 
> 
> but this works okay...
> $ curl get.pharo.org/70+vm <http://get.pharo.org/70+vm> | bash
> $ ./pharo-ui
> + World > Tools > Iceberg  
> ==>  iceberg opens okay
> 
> 
> while this PharoLauncher alternative fails...
> $  curl get.pharo.org <http://get.pharo.org/> | bash
> $ ./pharo-ui
> + World > Tools > Catalog > PharoLauncher 
> + World > PharoLauncher 
> ==> Doesn't show Pharo 7 templates
> + World > Monticello > ...PharoLauncher/main 
> + ConfigurationOfPharoLauncher-ChristopheDemarey.56 
> + World > Playground > "ConfigurationOfPharoLauncher load" 
> + World > PharoLauncher > Settings > Hard reset persistant state
> ==> Pharo 7 templates showing
> + '7latest32'   "i.e. the previously downloaded image"
> + World > Tools > Iceberg  
> ==>  "Error: External module not found"
> 
> + World > Monticello... load latest mczs...
> * PharoLauncher-Core-ChristopheDemarey.123
> * PharoLauncher-Spec-ChristopheDemarey.51
> + '7latest32'   "i.e. the previously downloaded image"
> ==> "Fetching VM to run Pharo 70 images" 
> + World > Tools > Iceberg  
> ==>  "Error: External module not found"
> 
> 
> Version comparison...
> 
> direct get.pharo.org/70+vm <http://get.pharo.org/70+vm>
>   /home/ben/Pharo/adhoc/Pharo7/Pharo.image
> Pharo7.0alpha.build.132.sha.4ea2f39a9f43185d31b844be5ad33b677f43bf17
>   /home/ben/Pharo/adhoc/Pharo7/pharo-vm/lib/pharo/5.0-201708271955/pharo
> CoInterpreter VMMaker.oscog-eem.2265 uuid: 
> 76b62109-629a-4c39-9641-67b53321df9a Aug 27 2017
> 
> via PharoLauncher... 
>   /home/ben/Pharo/images/7latest32/7latest32.image
> Pharo7.0alpha.build.132.sha.4ea2f39a9f43185d31b844be5ad33b677f43bf17  
>   /home/ben/Pharo/vms/70-x86/lib/pharo/5.0-201708271955/pharo
> CoInterpreter VMMaker.oscog-eem.2265 uuid: 
> 76b62109-629a-4c39-9641-67b53321df9a Aug 27 2017
> 
> Must be something environmental in lookup paths.  Can you reproduce or 
> suggest further investigation I can do?
> 
> cheers -ben
> 
> 
> 
> 
> 

Re: [Pharo-users] (no subject)

[Christophe Demarey]

> Le 20 sept. 2017 à 23:36, Stephan Eggermont  a écrit :
> 
> On 17/09/17 05:21, Sean P. DeNigris wrote:
>> demarey wrote
>>> Let us know if everything is fine.
>> Was this issue [1] fixed:
>>> I think I found it. It seems to be a bug in the way Launcher/Pharo unzips
>>> the downloaded VM.
>>> The symlink libgit2.dylib becomes unlinked and appears as just a regular
>>> file. If I delete the
>>> Pharo.app and unzip 70.zip (which btw maybe should have been deleted
>>> during cleanup after
>>> the dl) to ./vms/70, the error disappears!
> 
> With todays bleedingEdge PharoLauncher on mac I was able to download and 
> start 60510. When trying to load GToolKit
> 
> Iceberg enableMetacelloIntegration: true.
> Metacello new
>   baseline: 'GToolkit';
>   repository: 'github://feenkcom/gtoolkit/src';
>   load.
> 
> I ran into some problem with libgit2 that I could solve by doing the 
> unzipping of the downloaded vm zipfile with the Archive Utility and replacing 
> the vm (and re-adding the sources).

I tried to reproduce the problem with the same configuration without success. 
It works well.
I guess you already downloaded the required VM with a previous version of the 
launcher (using Pharo unzipping). If you delete this VM and use the launcher to 
run the image, it should download the VM and uses your system unzip command to 
extract it.

Re: [Pharo-users] (no subject)

[Christophe Demarey]

> Le 21 sept. 2017 à 13:25, Stephane Ducasse  a écrit :
> 
> Christophe
> 
> which version should we take?

For now, the bleeding edge. I will publish a stable as soon as it gets tested 
enough.

Re: [Pharo-users] (no subject)

[Christophe Demarey]

> Le 14 sept. 2017 à 14:55, stephan <step...@stack.nl> a écrit :
> 
> On 14-09-17 14:25, Christophe Demarey wrote:
>> Let us know if everything is fine.
> 
> Downloaded latest 7 64:
> Cannot open iceberg, libgit2 can not be found

Thanks for the report.
I will check that

Re: [Pharo-users] (no subject)

[Christophe Demarey]

Hi Serge,

> Le 14 sept. 2017 à 14:59, Serge Stinckwich <serge.stinckw...@gmail.com> a 
> écrit :
> 
> On Thu, Sep 14, 2017 at 2:25 PM, Christophe Demarey 
> <christophe.dema...@inria.fr <mailto:christophe.dema...@inria.fr>> wrote:
> Hi,
> 
> I published an update of the Launcher yesterday fixing some issues to run 
> Pharo images from Linux. You can find latest-versions (0.2.13) here: 
> http://files.pharo.org/platform/launcher/ 
> <http://files.pharo.org/platform/launcher/>
> It already propose to download Pharo 70 images. The VM is now downloaded 
> automatically if the adequate one is not available.
> Let us know if everything is fine.
> 
> 
> 
> ​Hi Christophe,
> 
> thank you for your support on PharoLauncher !
> ​I dl the last version for Mac OS X: Pharo_0.2.13.dmg
> This is really nice to be able to download the VM depending on the VM.

Yes. without that feature, the launcher started to be useless.

> I have only two issues: 
> - some icons do no show up when PharoLauncher is launch. They only appear 
> when on mouse hover.

yes, a regression in Pharo 6 :(

> - for some of my images, I have an error: Instances of UndefinedObject are 
> not indexable.

Could you send me the stacktrace and / or the images if possible?

Re: [Pharo-users] (no subject)

[Christophe Demarey]

Hi Stephan,

> Le 14 sept. 2017 à 14:49, stephan <step...@stack.nl> a écrit :
> 
> On 14-09-17 14:25, Christophe Demarey wrote:
>> I published an update of the Launcher yesterday fixing some issues to run 
>> Pharo images from Linux. You can find latest-versions (0.2.13) here: 
>> http://files.pharo.org/platform/launcher/
>> It already propose to download Pharo 70 images. The VM is now downloaded 
>> automatically if the adequate one is not available.
>> Let us know if everything is fine.
> 
> - not downloading the sources;

What do you mean?
For Pharo images < 70, sources file for the current version and the previous 
version are downloaded with the VM.

> - sources now need to be next to the image;

For Pharo 7 images, no sources file are dowloaded since they come with the 
image.

> - when you upgrade the vm, you need to replace all those sources files
> - default now dark theme, and not easy to switch

I think default theme for the Launcher should be the same as the Pharo theme.
Would a (Launcher) preference help?

> - icons only shown on hover

yes, Pharo 6 regression.
I had the problem on other UIs. Will see if I can find a fix quickly.

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Gabriel,

> Le 6 oct. 2017 à 15:48, Gabriel Cotelli  a écrit :
> 
> I've tried it on Linux, and after the download has finished it cannot launch 
> a Pharo 6.1 image. Find attached the PharoDebug.log
> 
> Seems it's searching for the sources file in a wrong location.
> Pharo cannot locate the sources file named 
> /home/gcotelli/Pharo/pharo-launcher/bin/lib/pharo/5.0-201707201942/PharoV60.sources.

Strange indeed!
I’m not able to reproduce it on my ubuntu vm.

It’s surprising that sources file is search in the folder of the VM shipped 
with the launcher. Images are run with VM downloaded in the VM store 
($HOME/Pharo/vms) and sources file is search in the vm directory.
Could you tell me if you run the launcher from a terminal? Could you send me 
your environment variables to see if I can find something relevant there?

[Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi all,

I just released a new version of the PharoLauncher 1.0.1.
Here are major improvements since last version:
Pharo launcher is now distributed with a Pharo 60 image with a 32-bits threaded 
VM (to avoid hangs when using OSProcess).
Add a VM management utility able to find (and download) the appropriate VM for 
a given image. So, the launcher is NOT shipped anymore with an extra VM. It 
will fetch required VM when needed.
Pharo Launcher files (images, VMs) are now considered as user Documents (easier 
to find). Consequently, they are now stored in the standard documents folder, 
i.e.
$HOME/Documents/Pharo on OS X,
$HOME/My Documents/Pharo on windows,
$HOME/Pharo on Linux (some linux distributions provide a document folder but 
some others not, so we put it in the HOME directory in all cases).
In this folder, you will find your images and virtual machines needed to run 
images.
Use of ProcessWrapper to run images on Windows.
Adds download of sources file for Pharo images > 60.
Left panel is now used for official images and Jenkins jobs templates and right 
panel is used for local images.
New mooc entry with the image ready to start the Pharo Mooc.
New deprecated category to find old official Pharo images.
Use system unzip when possible (Pharo unzip does not manage file permissions 
nor symbolic links).
You can get platform bundles from files.pharo.org:
Os X: http://files.pharo.org/platform/launcher/Pharo_1.0.1.dmg
Windows: http://files.pharo.org/platform/launcher/pharo_installer-1.0.1.exe
Linux: http://files.pharo.org/platform/launcher/Pharo-linux-1.0.1.zip

Enjoy,
Christophe.

For those wanted more info on how the adequate VM is computed, here is the 
process:
determine the image format version
find (and optionnaly fetch) a compatible VM (a VM able to run this image 
format) and run the image to get its Pharo version number
find (and optionnaly fetch) the appropriate VM for this specific Pharo image 
version. Sources files are also downloaded with the VM when applicable.
run the image with the appropriate VM

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Vitor, 

> Le 10 oct. 2017 à 19:00, Vitor Medina Cruz  a écrit :
> 
> « Program » folder is the default location where to install apps on Windows.
> If a user does not have admin rights, he can simply install it under its home 
> directory.
> 
>> Also there is the UAC Virtualization fiasco that complicates installing 
>> upgrades.
>> http://lists.pharo.org/pipermail/pharo-dev_lists.pharo.org/2014-January/091645.html
>>  
>> 
> 
>> Could we *please* install Windows PharoLauncher into user folders?
> 
> 
> Users can choose to install Pharo Launcher where they wish.
> We could choose another default location but not sure it is best option.
> 
> Nope, I don’t have any option of changing Pharo launcher installation. I 
> double click the exe and receive a message telling me I need administrative 
> rights. Can this be changed? It’s a deal breaker for those using Windows.
> 
For Windows, Pharo is shipped with an installer. Its only goal is to ask you 
where to install Pharo. So yes, the installation directory is configurable.
Just to be sure, here the link to the installer: 
http://files.pharo.org/platform/launcher/pharo_installer-1.0.1.exe

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

> Le 11 oct. 2017 à 12:10, stephan <step...@stack.nl> a écrit :
> 
> On 06-10-17 13:26, Christophe Demarey wrote:
>> For those wanted more info on how the adequate VM is computed, here is the 
>> process:
>> determine the image format version
>> find (and optionnaly fetch) a compatible VM (a VM able to run this image 
>> format) and run the image to get its Pharo version number
>> find (and optionnaly fetch) the appropriate VM for this specific Pharo image 
>> version. Sources files are also downloaded with the VM when applicable.
>> run the image with the appropriate VM
> 
> Thanks, Christophe.
> 
> It might be useful to regularly check for newer vms, at least the stable 
> ones. VMs are supposed to be backwards compatible.

Right.
I added an issue: https://github.com/pharo-project/pharo-launcher/issues/42

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Vitor,

> Le 11 oct. 2017 à 13:44, Vitor Medina Cruz  a écrit :
> 
> Hi Christophe,
> 
> Yes, but:
> 1- I can’t execute it if I don’t have administrative rights
> 
ok. That the point I missed.
I think I will add a zip file with PharoLauncher for windows for user like you 
not having admin rights.
>  — I have tried with the provided link. Before I get to the installation 
> directory choice, windows prompt me an error telling me I must be an admin. 
> It is probably some windows policy for all exe files, I know it is possible 
> to bypass this as some applications do (Intellij for example), but it is 
> usually easier to provide a zip;
> 
> 
> 
> 
> 2- 7-Zip can open this executable as it were a zip file, that’s how I manage 
> to install Pharo Launcher.

I need to think on the best way to handle that.

Regards,
Christophe.

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Ben,

> Le 13 oct. 2017 à 06:20, Ben Coman <b...@openinworld.com> a écrit :
> 
> On Thu, Oct 12, 2017 at 3:43 PM, Christophe Demarey 
> <christophe.dema...@inria.fr <mailto:christophe.dema...@inria.fr>> wrote:
> Hi Vitor,
> 
>> Le 11 oct. 2017 à 13:44, Vitor Medina Cruz <vitormc...@gmail.com 
>> <mailto:vitormc...@gmail.com>> a écrit :
>> 
>> Hi Christophe,
>> 
>> Yes, but:
>> 1- I can’t execute it if I don’t have administrative rights
>> 
> 
> ok. That the point I missed.
> I think I will add a zip file with PharoLauncher for windows for user like 
> you not having admin rights.
> 
> Might this be a "portable" app, where downloaded images and VMs are stored 
> next to the PharoLancher application, configured by a startup-setting file?
> 
> 
> Alternatively...
> searching around I found this...
> https://stackoverflow.com/questions/18990418/nsis-require-admin-permission 
> <https://stackoverflow.com/questions/18990418/nsis-require-admin-permission>
> https://github.com/NSIS-Dev/Documentation/blob/master/Reference/RequestExecutionLevel.md
>  
> <https://github.com/NSIS-Dev/Documentation/blob/master/Reference/RequestExecutionLevel.md>

Good catch. probably by using RequestExecutionLevel = highest should solve the 
problem.

> Also, when PharoLauncher is installed by an administrator and later used by a 
> standard user, I guess there may be problems with Pharo needing to write 
> changes to a restricted location??   Maybe that can be dealt with by... when 
> a new user runs PharoLauncher from the start menu, copy it to their user 
> folder $LocalAppdata\Programs\PharoLauncher
> and run PharoLauncher from there where they have privilege to write? 
> https://stackoverflow.com/questions/31641818/install-application-files-to-standard-windows-user-using-nsis
>  
> <https://stackoverflow.com/questions/31641818/install-application-files-to-standard-windows-user-using-nsis>

Yes, you could have problems but I do not think that adding more complexity to 
the installer is the way to go. I would rather prefer, in preference order:
1/ Pharo Launcher do not need to write data,
2/ if files need to be written, they should go to a Pharo data folder (e.g. 
$LocalAppdata\Programs\PharoLauncher) or to the user apps setting folder.

What I would like is that a Pharo image writes data where it should

> 
> btw, where are the NSIS source files accessible? 

https://github.com/pharo-project/pharo-build-scripts/tree/master/windows-installer

> Can they be dropped in the PharoLauncher git repo?

No because they are shared and used with other Pharo build scripts.

Thanks for the information.
Cheers,
Christophe

[Pharo-users] Re : Force headless mode

[Christophe Demarey]

Hi

What about deploying a minimal image without  UI ?
But, as you probably know, headless mode does not prevent to dump the code.

Christophe
- Cyril Ferlicot  a écrit :
> Hi,
> 
> At Synectique we are working on the deployment of a web application
> and we would like to force the headless mode. We tried to just
> override the method SmalltalkImage>>#isHeadless to return true all the
> time be it is not enough because the image is still opened. It is
> probabaly the vm that open the window.
> 
> I know we can add some arguments to the command line to open the image
> as headless but we want to forbid the user to launch the image without
> headless to protect the code.
> 
> Is this possible? Do you have some pointers?
> 
> Thanks in advance.
> 
> -- 
> Cyril Ferlicot
> https://ferlicot.fr
> 
> http://www.synectique.eu
> 2 rue Jacques Prévert 01,
> 59650 Villeneuve d'ascq France
> 

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Nicolai,

> Le 7 oct. 2017 à 09:38, Nicolai Hess  a écrit :
> 
> Just starting pharolauncher on windows still does not work (if you put it 
> into the default "program" folder).
> On start, it tries to create a local direcitory "pharo-local" and does not 
> have the permission to do so.
> You need to start it with admin-rights.

A Pharo image needs write permission in the image folder to store image 
sessions (epicea) as well as other image data (package cache, play-cache, 
changes file, etc.)
Maybe we should disable Epicea and it could work in a read only environment?

For now, either you can start it with admin rights or you install Pharo 
Launcher in a user folder.

Regards,
Christophe

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

> Le 6 oct. 2017 à 22:43, Gabriel Cotelli  a écrit :
> 
> I've tried running it from the terminal and it works!.
> 
> If I double-click it from the file manager and choose "Run in Terminal" it 
> gives the previously commented error.

I fear I cannot do much for this problem.
I tried on ubuntu and fedora to run but I have no way to run this exe directly. 
I have no menu « Run in terminal » .
I was only able to create a launcher file (pharo.desktop). Then, by 
double-clicking on this launcher file, I was able to run the launcher and a 
pharo 6.1 image without error.
Maybe the best would be to use a launcher file (exemple as attachment).

Regards,
Christophe



pharo.launcher
Description: Binary data

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

Hi Peter,

> Le 6 oct. 2017 à 21:13, Peter Uhnák  a écrit :
> 
> +1.0.1 :-)
> 
> > Pharo Launcher files (images, VMs) are now considered as user Documents 
> > (easier to find). Consequently, they are now stored in the standard 
> > documents folder, i.e.
> > $HOME/Pharo 
> 
> This seems like a step backwards. The expected location is 
> $XDG_DATA_HOME/Pharo . Is it possible to change or configure this?

We decide to consider Pharo Launcher files (virtual machines, images) as user 
documents. That’s why we changed the location of these files on all platforms.
$XDG_DATA_HOME/Pharo is not a document location on linux and there is no 
standard Documents location. That’s why we ended to put Pharo Launcher files in 
a Pharo folder in the HOME directory.
Maybe it is not the best choice. 

If we can find an agreement on the best Pharo Launcher documents location on 
Linux, it can be changed.

You can already configure where you want to store Pharo Launcher images and vms 
(2 different settings).

Christophe

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

> Le 7 oct. 2017 à 14:26, Ben Coman  a écrit :
> 
> On Sat, Oct 7, 2017 at 3:38 PM, Nicolai Hess  > wrote:
> Just starting pharolauncher on windows still does not work (if you put it 
> into the default "program" folder).
> On start, it tries to create a local direcitory "pharo-local" and does not 
> have the permission to do so.
> You need to start it with admin-rights.
> 
> Why are we making it harder for Windows users in corporate environments by 
> requiring them to have admin-rights they often won't have, while in Linux 
> PharoLancher is installed into user folders?

« Program » folder is the default location where to install apps on Windows.
If a user does not have admin rights, he can simply install it under its home 
directory.

> Also there is the UAC Virtualization fiasco that complicates installing 
> upgrades.
> http://lists.pharo.org/pipermail/pharo-dev_lists.pharo.org/2014-January/091645.html
>  
> 
> 
> Could we *please* install Windows PharoLauncher into user folders?

Users can choose to install Pharo Launcher where they wish.
We could choose another default location but not sure it is best option.

To me, the real problem is: why does Pharo apps (and Pharo Launcher) needs to 
write data in the application directory? By example, for Pharo Launcher, we do 
not need to load any code nor to backup sessions.

Re: [Pharo-users] [Ann] PharoLauncher v1.0.1 released!

[Christophe Demarey]

> Le 9 oct. 2017 à 15:56, Peter Uhnák  a écrit :

> I guess I hate when applications are throwing their garbage into my $HOME 
> (this also goes for VirtualBox, Eclipse, and everyone else).

Perfectly understandable.

> If I am not accessing the files directly, but they are pretty much always 
> managed by some other applications, then for me they are polluting $HOME. 
> (And yes, I sometimes do launch to image by hand, or manipulate VirtualBox by 
> hand, but that doesn't change the fact that it is not the primary way to work 
> with it.)
> 
> But no worries, I rest my case and I patch it for myself. (So no need to add 
> configuration for this either.)

No need to patch, you already have settings to configure where to store images 
and virtual machines.

Regards,
Christophe

Re: [Pharo-users] I love the launcher!!!!

[Christophe Demarey]

Hi Phil,

> Le 26 nov. 2017 à 22:25, p...@highoctane.be a écrit :
> 
> So, I try it on my CentOS 7 64 bit and... as there is no 64 bit version of 
> the launcher…

Right.
Can’t you use the 32 bit version?
Anyway, I think it would be good to have a 64 bit version.  Could you open an 
issue for that please?

> 
> pharo6-64-ui PharoLauncher.image 
> This interpreter (vers. 68021) cannot read image file (vers. 6521).
> Press CR to quit...
> 
> How is one using the launcher for 64 bit stuff?
> 
> Phil
> 
> On Sun, Nov 26, 2017 at 9:45 PM, Stephane Ducasse  > wrote:
> If you don't try you will not understand what you miss.
> 
> On Sun, Nov 26, 2017 at 2:10 PM, Herby Vojčík  > wrote:
> > Stephane Ducasse wrote:
> >>
> >> Then do not try :) like that you will never get a chance to know :)
> >
> >
> > Could you please explain? I understood your sentence has high dose of homour
> > which I cannot decipher since I lack certain abilities for that.
> >
> > Thanks, Herby
> >
> >> Stef
> >>
> >>
> >> On Sun, Nov 26, 2017 at 1:48 PM, Herby Vojčík >> >  wrote:
> >>>
> >>> Stephane Ducasse wrote:
> 
>  Why don't you try? It does not bite.
> 
>  For me it works in all scenario. I have projects that i manage over
>  several weeks and others I drop day to day.
>  And I have also startup script per versions.
> >>>
> >>>
> >>> Maybe I will. The main problem was I didn't what it's good about at all -
> >>> load image for a version, runs it - what's the "wow it's great" about?
> >>> Missed part was it loads correct vm; plus makes sense when you have
> >>> hundreds
> >>> (I don't). Now it must be combined with the startup script magic to work
> >>> with long-time projects, but those are also new to me - did not know of
> >>> them
> >>> until your booklet, not using them at all yet.
> >>>
> >>> Herby
> >>>
> >>>
>  Stef
> 
>  On Fri, Nov 24, 2017 at 3:10 PM, Herby Vojčík  >   wrote:
> >
> > Thank you all, now I understand it better. Good for lots of "branches".
> >
> > However, I wonder how does it work with the rule I read somewhere:
> > "start
> > each day [of work on a project] with a new image", which means I should
> > not
> > reuse clean new image (as it needs populating from VCSes etc.) nor
> > reuse
> > existing image (I should start with the new one). Or does it combine
> > with
> > some startup-magic described in one of the recent Steph's booklets, the
> > "one
> > start per [new] image" case (but again, one should discriminate
> > projects
> > from each other)?
> >
> > Thanks, Herby
> >
> > Peter Uhnák wrote:
> >>
> >> Hi Herby,
> >>
> >> normally people use different images for their different projects,
> >> different versions, trying things, etc. Which means we end up with
> >> many
> >> locations on disk, and it can be hard to track.
> >> So PharoLauncher is a nice tool where you can download fresh image
> >> just
> >> by clicking, and you see the list of your local images and can launch
> >> them, etc.
> >>
> >> Peter
> >>
> >> On Thu, Nov 23, 2017 at 12:56 PM, Herby Vojčík >> 
> >> >>   wrote:
> >>
> >>   Stephane Ducasse wrote:
> >>
> >>   Hi
> >>
> >>   I love the PharoLauncher.
> >>
> >>
> >>   Pardon my question, I have downloaded it and looked at it, but I
> >>   don't get it. What does it do / what are the use cases (honest
> >>   question)?
> >>
> >>   Thanks, Herby
> >>
> >>
> >>   It helps me to manage my parallel development and projects.
> >>
> >>   We should put a link on the Pharo web site because
> >>
> >>   http://files.pharo.org/platform/launcher/ 
> >> 
> >>    >> >
> >>
> >>   is arcane.
> >>
> >>   Stef
> >>
> >>
> >>
> >>
> >>>
> >>
> >
> >
> 
> 
> 

[Pharo-users] files.pharo.org downloads => needs testers

[Christophe Demarey]

Hi all,

We set up 2 new servers to replace (at least for some time) the problematic 
files.pharo.org  server (slow downloads, bad CRC).
Please, could you test these new servers and give us feedback to see if 
downloads are fast and nom ore CRC errors?
* http://file-pharo.inria.fr/
* http://files2.pharo.org (hosted on the cloud)

Data might be a bit outdated (sync last friday) but is just for testing now.

According to your feedbacks, we will switch official URLs to one of these 
servers.
The situation should be back to normal soon.

Christophe

Re: [Pharo-users] Travis build failing because of bad CRC

[Christophe Demarey]

Hi,

The new server is now available to the world: http://file-pharo.inria.fr/
Could you test and tell me if you still encounter slowness and bad CRC?

Christophe


Note: latest data is not there since we still publish to files.pharo.org.