March 2017 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Debugger: is there a way to skip a message?

by Max Leske

What do you mean by "skip"? Not execute? Then not that I'm aware of. But you can: 1. step into 2. return value from method (right click on the current frame) With a return value of nil that will have the same effect. HTH, Max > On 10 Mar 2017, at 21:20, stepharong <stepharong(a)free.fr> wrote: > > Hi > > I need to skip message while debugging and I thought that it was possible. > May be I do not see it. > > Stef > > -- > Using Opera's mail client: http://www.opera.com/mail/ >

March 13, 2017

Re: [Pharo-dev] Growing large images: the case of Moose models

by ClÃ©ment Bera

Hi again, I summarised my experiment with Moose and 2 other experiments in a "Tuning the Pharo garbage collector" blog post here: https://clementbera.wordpress.com/2017/03/12/tuning-the-pharo-garbage-colle… I may integrate some new methods in the VirtualMachine class in Pharo in the next few months to make VM tuning easier to deal with. I will update the scripts in the blog post accordingly if I do so. Best, On Sat, Mar 4, 2017 at 12:49 AM, stepharong <stepharong(a)free.fr> wrote: > tx clement > > this is indeed really important > > On Fri, 03 Mar 2017 11:56:05 +0100, ClÃ©ment Bera <bera.clement(a)gmail.com> > wrote: > > Hello everyone, > > This morning I investigated with Vincent Blondeau a problem reported by > the Moose community a while ago: loading Moose model is slower in Spur > (Pharo 5+) than in pre-Spur (Pharo 4 and older). In general, this problem > was present for anyone growing images to a significant size. > > To investigate the problem, we loaded a 200Mb[3] Moose model on a 250Mb > image, growing the image to 450Mb. Loading such a model takes 2 minutes in > Spur and 1m30s in pre-Spur VMs. > > Using the stable Pharo VM, the analysis results were the following: > - total time spent to load the Model: 2 minutes > - time spent in full GC: 1 minute (4 fullGCs) > - time spent in scavenges[1]: 15 seconds > On the 2 minutes spent, we have 50% of the time spent in full GCs, 12.5% > in scavenges, 37.5% executing code. > > We then used the latest VM that features the new compactor (VM from > beginning of March 2017 and over). The full GC execution time went down > from 1 minute to 2 seconds. > > In addition, we increased the size of Eden[2] from 4Mb to 12Mb. Time spent > in scavenges decreased from 15 seconds to 5 seconds. > > Overall, loading the model is now taking ~50 seconds instead of 2 minutes. > > To increase Eden size, one needs to run a script similar to: > > | currentEdenSize desiredEdenSize | > currentEdenSize := Smalltalk vm parameterAt: 44. > desiredEdenSize := currentEdenSize * 4. > Smalltalk vm parameterAt: 45 put: desiredEdenSize. > > *And then restart the image.* > > I hope this report can be useful for some of you. I will try to make a > blog post out of it, detailing other GC settings one can change from the > image to improve performance. > > Best, > > Clement > > [1] A scavenge is basically the garbage collection of only young objects > [2] Eden is basically the space where objects are initially allocated. > [3] All numbers in the report are order of magnitudes and not precise > numbers > > > > > > > -- > Using Opera's mail client: http://www.opera.com/mail/ >

March 12, 2017

Re: [Pharo-dev] perform:withArguments: and collection

by Eliot Miranda

On Sun, Mar 12, 2017 at 3:16 PM, Peter Uhnak <i.uhnak(a)gmail.com> wrote: > Hi, > > Apparently sending #perform:withArguments: and using OrderedCollection > leads to PrimitiveFailed. > > I can recast it, but maybe we can just add > > (argArray isArray) > ifFalse: [ ^ self error: 'Arguments must be instance of Array' ] > +1. The alternative is having the primitive failure code test for the argument being sequence able and doing it for you: perform: selector withArguments: argArray inSuperclass: lookupClass "NOTE: This is just like perform:withArguments:, except that the message lookup process begins, not with the receivers's class, but with the supplied superclass instead. It will fail if lookupClass cannot be found among the receiver's superclasses. Primitive. Essential. See Object documentation whatIsAPrimitive." <primitive: 100> argArray isArray ifFalse: [(argArray isCollection and: [argArray isSequenceable]) ifTrue: [^self perform: selector withArguments: argArray asArray inSuperclass: lookupClass]. ^self error: 'argArray argument must be an Array']. selector isSymbol ifFalse: [^self error: 'selector argument must be a Symbol']. selector numArgs = argArray size ifFalse: [^self error: 'incorrect number of arguments']. (self class == lookupClass or: [self class inheritsFrom: lookupClass]) ifFalse: [^self error: 'lookupClass is not in my inheritance chain']. (self class lookupMethodFor: selector ifAbsent: nil) ifNotNil: [:method| | overflowArgs | (method numArgs = argArray size and: [argArray size >= 15]) ifTrue: [overflowArgs := argArray copyFrom: 1 to: 15. overflowArgs at: 15 put: (argArray copyFrom: 15 to: overflowArgs size). ^self perform: selector withArguments: overflowArgs inSuperclass: lookupClass]]. self primitiveFailed > > to make the error less cryptic. > > (or allow OrderedCollection, but since the operation is done in VM it > might be needlessly more work.) > Right. IMO, it's a bad idea putting knowledge of OrderedCollection's representation in the VM. > Peter > > -- _,,,^..^,,,_ best, Eliot

March 12, 2017

perform:withArguments: and collection

by Peter Uhnak

Hi, Apparently sending #perform:withArguments: and using OrderedCollection leads to PrimitiveFailed. I can recast it, but maybe we can just add (argArray isArray) ifFalse: [ ^ self error: 'Arguments must be instance of Array' ] to make the error less cryptic. (or allow OrderedCollection, but since the operation is done in VM it might be needlessly more work.) Peter

March 12, 2017

Re: [Pharo-dev] Debugger: is there a way to skip a message?

by Ron Teitelbaum

Hi All, There are ways that I skip execution. Usually in loop processing. If you change the index of a loop, (in the debugger change the number and accept) it will continue on from that index point. You can also easily use run to here, to get to the top of a method, then change the method with false ifTrue: [ ... ] than accept to restart the method. You could also set a temp var like runThis := true. Wrap you code in runThis ifTrue: [ ]. then while debugging you can change the value of runThis in the debugger to skip execution. All the best, Ron On Sun, Mar 12, 2017 at 12:45 PM stepharong <stepharong(a)free.fr> wrote: > > > Oh, maybe âReturn entered valueâ is what Stef was looking for. > > If yes, you can obtain it from the contextual menu of the Stack. > > > it is not clear that it will skip the execution because it could do the > execution and return the value you type. > > > > Cheers, > Doru > > > On Mar 11, 2017, at 3:03 PM, Ben Coman <btc(a)openInWorld.com > <btc(a)openinworld.com>> wrote: > > On Sat, Mar 11, 2017 at 4:20 AM, stepharong <stepharong(a)free.fr> wrote: > > Hi > > I need to skip message while debugging and I thought that it was possible. > May be I do not see it. > > > This would be a nice feature. I never noticed if we ever had it > Two work arounds I sometimes use... > * select rest of the method and "Debug It". > * step-Into, and then "Return entered value" > > cheer -ben > > > -- > www.tudorgirba.com > www.feenk.com > > "Every thing should have the right to be different." > > > > > > > > -- > Using Opera's mail client: http://www.opera.com/mail/ >

March 12, 2017

Re: [Pharo-dev] Did we lose edit-in-place in GT Inspector in Pharo 6?

by Alistair Grant

On 12 March 2017 at 03:36, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > Double clicking on a value in the raw tab seems to have no effect. > > p.s I asked twice in Discord with no response. I guess I'm finally getting > old enough that I don't understand how to use new platforms ;) While I'm feeling much the same (about being old and not understanding), and can empathise, I think in this case it is simply that the chat platforms are in flux at the moment. Esteban has requested we move to Discord, but the numbers are still fairly low. Hopefully things will improve soon. I.e. don't give up on discord just yet. :-) Cheers, Alistair

March 12, 2017

Re: [Pharo-dev] Did we lose edit-in-place in GT Inspector in Pharo 6?

by Andrei Chis

On Sun, Mar 12, 2017 at 5:07 PM, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > Tudor Girba-2 wrote > > this feature is temporarily not available due to the move to FastTable > > :( > > > Tudor Girba-2 wrote > > but we are working on adding it back. > > :) > > I use this all the time to e.g. nil out an inst var to > re-lazily-initialize. > I noticed it was missing (after trying to press the mouse button harder for > a few minutes... sigh) within minutes of testing out Pharo 6. There was no > setter, so I ended up doing an ugly, and time-consumiung, > `#instVarNamed:put:`. Will it be re-added in 6 before the release? I'd say > a > lost key feature is a bug, no? > This feature was lost because of the inspector move to FastTable and the adding of automatic refresh. But we hope to add it back for Pharo 6. Just it will be enabled on Shift+Click and not double click. Cheers, Andrei > > > > ----- > Cheers, > Sean > -- > View this message in context: http://forum.world.st/Did-we- > lose-edit-in-place-in-GT-Inspector-in-Pharo-6-tp4938264p4938297.html > Sent from the Pharo Smalltalk Developers mailing list archive at > Nabble.com. > >

March 12, 2017

debugger feature I would pay for

by stepharong

Hi doru You have some features and I would love to use it for real. Here is my scenario: I have a test failing in pillar dev but not in my previous image. :( I want to run a test in an image (green test). - get the test execution stack In another image (dev), I would like to load the green test stack and say to the debugger - execute the test in the local image up to there is a change by comparison with the other. I would pay for this feature. Stef

March 12, 2017

Re: [Pharo-dev] what is the SpecLayout architecture?

by stepharong

On Sat, 11 Mar 2017 14:55:35 +0100, Peter Uhnak <i.uhnak(a)gmail.com> wrote: > Hi, > > this is probably a tall order, but is there some description of > SpecLayout architecture? > > To clarify: I am NOT asking how to use SpecLayout, I am asking how > SpecLayout works internally. > > iirc SpecInterpreter was rewritten by Johan & Stef, so maybe you guys > have some notes? No we did not. In fact the interpreter not really nicely done. Esteban rewrote it but the spec author never wanted to listen and after a while he refactored the main ugly interpret method and since then I did not look at it. And Esteban got deadly attracted by uFFI and related ugly instability. I think that spec deserves a large pass but this is not something that you do without enough time and concentration. > > Thanks, > Peter > -- Using Opera's mail client: http://www.opera.com/mail/

March 12, 2017

Re: [Pharo-dev] Debugger: is there a way to skip a message?

by stepharong

> Oh, maybe âReturn entered valueâ is what Stef was looking for. > > If yes, you can obtain it from the contextual menu of the Stack. it is not clear that it will skip the execution because it could do the execution and return the value you type. > > > > Cheers, > Doru > > >> On Mar 11, 2017, at 3:03 PM, Ben Coman <btc(a)openInWorld.com> wrote: >> >> On Sat, Mar 11, 2017 at 4:20 AM, stepharong <stepharong(a)free.fr> wrote: >>> Hi >>> >>> I need to skip message while debugging and I thought that it was >>> possible. >>> May be I do not see it. >> >> This would be a nice feature. I never noticed if we ever had it >> Two work arounds I sometimes use... >> * select rest of the method and "Debug It". >> * step-Into, and then "Return entered value" >> >> cheer -ben >> > > -- > www.tudorgirba.com > www.feenk.com > > "Every thing should have the right to be different." > > > > -- Using Opera's mail client: http://www.opera.com/mail/

March 12, 2017