- Pharo-dev - lists.pharo.org

[Pharo-project] Reproducible Cog crash but no info to get an idea...

by Mariano Martinez Peck

Hi. We are having a crash in Fuel, which doesn't crash with the standard VM. To reproduce, take a Pharo 1.2 image and load the attached mcz. You will see that method: #materializeKeysAndValuesTo:with:size: has a "self halt" inside. If you ran the test FLCreateClassSerializationTest >> testCreateWithMethod it will crash. If you remove the "self halt" from #materializeKeysAndValuesTo:with:size: it works... I debuged it with GDB and I don't know why, I can ONLY see the last two instrunctions, which is a longAt: -8 I attach screenshot. A bt or a call printAllStacks() doesn't show me anything. If I run from a console, the only thing I get is a: Segmentation fault Smalltalk stack dump: Bus error I have no idea what can be happening. Thanks for any clue... -- Mariano http://marianopeck.wordpress.com

April 25, 2011

Re: [Pharo-project] new template for class empty comments for OB

by StÃ©phane Ducasse

Ok for me this stupid instance variables typed inferred comments is nearly totally unless. It gives the wrong message: a class comment is not something that can be generated automatically. The intention, collaborations and API are human decisions. I will add it to the default browser because this is the only to make people think that comments are not empty or just a stupid list of instance variables that we can get anyway. Stef On Apr 24, 2011, at 10:20 PM, Lukas Renggli wrote: > I think it is a pity to show this lengthy boilerplate when looking at > a class comment. For the normal browser it might be ok, but in OB with > RB loaded I actually like to see the inferenced instance variable > types at one glance when no other comment is available. I think this > behavior is more useful in most cases (i.e. when trying to understand > code). OB already picks up automatically the system defined class > comment template if the RB engine is not used. > > Lukas > > -- > Lukas Renggli > www.lukas-renggli.ch >

April 24, 2011

Re: [Pharo-project] [Vm-dev] Debugging Cog crash

by Mariano Martinez Peck

For the next time Bill, this blog post may help you: http://marianopeck.wordpress.com/2011/04/23/how-to-debug-the-vm/ Cheers Mariano On Sun, Apr 10, 2011 at 10:00 PM, Schwab,Wilhelm K <bschwab(a)anest.ufl.edu>wrote: > > Eliot, > > I modified the one-click's shell script as shown below and ran it from a > terminal. It still says only "Segmentation fault." I could be doing this > incorrectly or maybe not looking in the correct place for the output??? My > first attempt was to add -sendTrace=9 (note case), and it complained, which > is vastly better than ignoring my mistake; it also suggests that I got > -sendtrace=9 correct, or it would presumably complain again. > > Did I botch it, or is it really quitting before it reaches Smalltalk code? > If you suspect the latter, is there something else we can do to prove it, > and/or what's next? I am not opposed to building from source. I avoided it > for a while because I was afraid that I would "hack" the vm again and not > really fix my FFI oddities. I am close to having the latter sorted, and can > probably now trust myself with vm source :) > > Bill > > > > #!/bin/sh > > # path > ROOT=`dirname $0` > LINUX="$ROOT/Contents/Linux" > RESOURCES="$ROOT/Contents/Resources" > > # icon > gvfs-set-attribute \ > "$0" \ > "metadata::custom-icon" \ > "file://$RESOURCES/Squeak.png" \ > 2> /dev/null > > # execute > exec "$LINUX/squeakvm" \ > -sendtrace=9 \ > -plugins "$LINUX" \ > -encoding latin1 \ > -vm-display-X11 \ > "$RESOURCES/Pharo.image" > > > > -- Mariano http://marianopeck.wordpress.com

April 24, 2011

Re: [Pharo-project] new template for class empty comments for OB

by Lukas Renggli

I think it is a pity to show this lengthy boilerplate when looking at a class comment. For the normal browser it might be ok, but in OB with RB loaded I actually like to see the inferenced instance variable types at one glance when no other comment is available. I think this behavior is more useful in most cases (i.e. when trying to understand code). OB already picks up automatically the system defined class comment template if the RB engine is not used. Lukas -- Lukas Renggli www.lukas-renggli.ch

April 24, 2011

Re: [Pharo-project] [Metacello] Which packages should we include of metacello for Pharo 1.3?

by Mariano Martinez Peck

On Sun, Apr 24, 2011 at 5:59 PM, Dale Henrichs <dhenrich(a)vmware.com> wrote: > Mariano, > (excluding the OB-Metallo menu items) ... > > hi dale. What does it mean exactly? If right now I am loading : #('Metacello-Core' 'Tutorial' 'UI' 'Metacello-ProfStef' 'Metacello-Help'); and #('Tests') you mean I should remove the group 'UI' ? > I would hold off on including the MetacelloBrowser for a bit. The interface > is clunky at the moment and needs to be improved before we let it loose:) > > OK, perfect, you are the boss :) > Hopefully we'll get some more good feedback from some folks so that we can > focus on polishing up a subset of commands that can be packaged up for > "immediate" release ... > > great > Dale > > On Apr 24, 2011, at 7:33 AM, Mariano Martinez Peck wrote: > > Hi Dale and the rest. In Pharo 1.2 we used to include these packages/groups > from Metacello: #('Metacello-Core' 'Tutorial' 'UI' 'Metacello-ProfStef' > 'Metacello-Help'); and #('Tests') > > Is this still valid? because I decided them a couple of months ago...so > maybe they are not valid anymore. > For example, do we want the MetacelloBrowser or things like that ? > and the MetacelloToolbox ? > > Thanks > > -- > Mariano > http://marianopeck.wordpress.com > > > -- Mariano http://marianopeck.wordpress.com

April 24, 2011

Re: [Pharo-project] About testing that ensure works

by Igor Stasenko

On 24 April 2011 15:13, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > Ok I got it I ws confused by SUnit interference > :)) > testDo > Â Â Â Â "self run: #testDo" > > Â Â Â Â self do: [ Error signal]. > Â Â Â Â self assert: x = 3 > > > > self run: #testDo -> red tests > > self debug: #testDo + proceed -> green tests > Yes, sunit interrupts execution at the moment when error was triggered, before unwinding the stack. -- Best regards, Igor Stasenko AKA sig.

April 24, 2011

Re: [Pharo-project] Cog doesn't load anymore in PharoCore1.3 because of the removal of AbstractLauncher

by Mariano Martinez Peck

If Cog really needs it we should create a separate repostiroy like PharoCogSupport and put the class there and add it as a dependency of Cog in ConfigurationOfCog. Another hacky way is to add a #preLoadDoit: in ConfigurationOfCog that does a #compile: of AbstractLauncher Cheers Mariano On Sun, Apr 24, 2011 at 7:37 PM, Mariano Martinez Peck < marianopeck(a)gmail.com> wrote: > AbstractLauncher was removed in: > http://code.google.com/p/pharo/issues/detail?id=4002 > > Cog has a subclass from AbstractLauncher called CommandLineLauncher, hence, > it cannot be loaded anymore > > CommandLineLauncher has no class comment and no sender. So I have no idea > what it does, nor if it can be removed from cog, nor how to solve this > problem in Pharo. > > Cheers > > -- > Mariano > http://marianopeck.wordpress.com > > -- Mariano http://marianopeck.wordpress.com

April 24, 2011

[Pharo-project] new template for class empty comments for OB

by StÃ©phane Ducasse

Hi lukas could you add the following in OB for missing template http://code.google.com/p/pharo/issues/detail?id=4090 templateOn: aStream aStream nextPutAll: theClass name; nextPutAll: ' has not been documented yet. The class comment should describe the purpose of the class, its collaborations and its variables. '. aStream nextPutAll: 'We encourage you to fill up the following template.'; cr;cr. aStream nextPutAll: 'Class intention';cr ; tab. aStream nextPutAll: 'I''m doing the following ...';cr;cr. aStream nextPutAll: 'Class collaborations'; cr;tab. aStream nextPutAll: 'I usually do not work in isolation but with the help of ...'; cr ;cr . aStream nextPutAll: 'Class main API'; cr; tab. aStream nextPutAll: 'My main public API is ...'. aStream nextPutAll: 'My subclasses may want to override such specific hooks ...'; cr. aStream nextPutAll: 'Implementation notes';cr. self templateLabel: 'Instance Variables:' class: theClass variables: theClass instVarNames on: aStream. self templateLabel: 'Class Instance Variables:' class: theClass class variables: theClass class instVarNames on: aStream I did the same for the default browser. Stef

April 24, 2011

[Pharo-project] Cog doesn't load anymore in PharoCore1.3 because of the removal of AbstractLauncher

by Mariano Martinez Peck

AbstractLauncher was removed in: http://code.google.com/p/pharo/issues/detail?id=4002 Cog has a subclass from AbstractLauncher called CommandLineLauncher, hence, it cannot be loaded anymore CommandLineLauncher has no class comment and no sender. So I have no idea what it does, nor if it can be removed from cog, nor how to solve this problem in Pharo. Cheers -- Mariano http://marianopeck.wordpress.com

April 24, 2011

Re: [Pharo-project] What is a well documented class?

by Michael Forster

On Sun, Apr 24, 2011 at 11:17 AM, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: >>> Â - each public method belongs to a method category named public* >>> Â - each private method belongs to a method category named private* >> >> Others here may disagree -- and the dreary of legislating and >> enforcing conventions aside -- but I strive (usually with success) to >> move all privately-used methods to one or more delegates. Â One class's >> private method is just another's public??? > > A private method may have to access instances variables. Moving the method to another class may be difficult sometimes. All the more reason to move it! A private method accessing instance variables is just global-variable-oriented spaghetti programming disguised by a class. Such a method is just begging to be moved to a delegate and to accept the instance var values as arguments. >>> Â - other methods are considered as "package visible", meaning that they belong to a category that does not begins with 'private' or 'public' >> >> Again, I sympathise with facing the visibility problem, but I don't >> trust that conventions will be upheld by either end, so I tire of >> following them. Â Replace Pharo/Squeak categories with Newspeak >> modules; replace Pharo/Squeak protocols with traits (stateless, >> please); I'll be the first in line. > > I do not think this should be enforced. It is easy to infer method visibility with a set of well defined scenarios, for example the one that comes in unit tests. You can't infer reliably from such shaky premises, whether they are manifest as conventions or unit tests. Cheers, Mike

April 24, 2011