- Pharo-users - lists.pharo.org

Re: [Pharo-users] complex json parsing.

by Ben Coman

On Wed, 13 Mar 2019 at 02:10, Roelof Wobben <r.wobben(a)home.nl> wrote: > Op 12-3-2019 om 18:46 schreef Ben Coman: > > Then in Playground do... > collectionUrl := ' > https://www.rijksmuseum.nl/api/nl/collection?key=14OGzuak&format=json&type=… > '. > json := NeoJSONReader fromString: (ZnEasy get: collectionUrl) contents. > paintings := Paintings fromJSON: json. > paintings inspect. > > and then debug trace through that whole thing. > Is that what you were trying to do? Or have I got the wrong end of the > stick? > > cheers -ben > > > There is a one to one relationship between the objectNumber and a Painting > > I do not use the imageUrl found on this url because it's to big . > okay. So I presume you want "artObject/webImage/url" ? { "artObject": { "webImage": { "url": " http://lh6.ggpht.com/wwx2vAS9DzFmmyeZefPjMtmCNOdjD80gvkXJcylloy40SiZOhdLHVd… " > > So what I try now is first put all the objectNumber in a collection > and use that collection on a second url to fetch a image of the right size > Okay. I understand, but I am *still* coaching you to *not* do that. Rather than gathering a collection of dumb numbers and then creating new Painting objects from those, (which is awkward for joining the new data with the previous data) work with the Painting objects you already have. Add a new instance variables "webImageUrl" and "webImage" Then either Painting >> getMoreData |entity| ... webImageUrl := (artObjectJson at: 'webImage') at: 'url'. entity := (ZnEasy get: webImageUrl) entity. webImage := ImageReadWriter formFromStream: entity readStream. which may lock the UI for a moment downloading them all, or download lazily (slight variation on my previous imageEntity which doubles up on storage)... Painting >> getMoreData ... webImageUrl := (artObjectJson at: 'webImage') at: 'url'. Painting >> webImage |entity| webImageUrl ifNil: [^nil]. webImage ifNil: [ entity := (ZnEasy get: webImageUrl) entity. webImage := ImageReadWriter formFromStream: entity readStream]. ^ webImage Painting >> gtInspectorJpegIn: composite <gtInspectorPresentationOrder: 0> composite morph title: 'Painting'; display: [ self webImage ] > I hope im clear what I try to do in this step > yes it is. Please take my coaching and kill your collection of numbers. Keeping them keeps you stuck in old ways of thinking. So just kill it :) Once you've done that, alternatives will open up for you and you'll be on the path to a new way of thinking about programming. :) :) > so numbers should be a collection of objectNumbers > no. > and paintings a collection of painting with the info I need > yes, and objectNumber is part of the info of a painting. > > so to recap > > this one must give me a collection of objectNumbers : ' > https://www.rijksmuseum.nl/api/nl/collection?key=14OGzuak&format=json&type=… > '. > no. it gives you a collection of paintings, of which objectNumber is one attribute. > > and the objectNumbers I can use for this url : > > https://www.rijksmuseum.nl/api/nl/collection/<one objectNumber/tiles?key=14OGzuak&format=json > > and I do not store the objectNumber in Paintings because I do not use it to display it in a seaside project. > > Regardless of whether you display it or not, objectNumber is supplied by the server "as an attribute belonging to painting" so store it inside your Painting objects. Then rather than Seaside asking a collection of dumb numbers for each JPEG to display it asks a collection of smart Painting objects for their JPEG to display (since they know their own objectNumber, they can get the JPEG themselves). cheers -ben On Wed, 13 Mar 2019 at 02:10, Roelof Wobben <r.wobben(a)home.nl> wrote: > Op 12-3-2019 om 18:46 schreef Ben Coman: > > On Wed, 13 Mar 2019 at 00:43, Roelof Wobben <r.wobben(a)home.nl> wrote: > > I did try to make this idea work but now im complete lost. > > > > Right now the collection that should contain the objectNumbers is total > empty > > anyone who can see where I went wrong. > > > > I included my code so far. > > > A few things... > !Paintings class methodsFor: 'instance creation' stamp: 'RoelofWobben > 3/12/2019 17:26'! > getImagesDataFromJSON: json > | instance | > instance := Painting new. > (json at: #values) > do: [ :artObjectJson | > | painting | > painting := Painting new. > painting imageUrl: (json at: #name). > instance addPainting: painting ]. > ^ instance! ! > > You are adding a painting to a painting, which semantically makes no sense. > i.e. where #addPainting: is sent to "instance" , that is holding a > Painting. > > I would not expect your Painting class to have a method addPainting: > Paintings should have that method. > > btw, Stop using "instance" as a variable name. > Its as bad as variables named "a", "i", "t", "z" - it provides no semantic > context. > > -------------------- > Object subclass: #Paintings > instanceVariableNames: 'paintings numbers' > > Paintings class > instanceVariableNames: 'numbers paintings'! > > You've defined these variables on both the instance-side and class-side. > I'll hazard a guess that you should delete the class-side definition. > > ---------------------------------- > > TBApplicationRootComponent initialize!Object subclass: #Painting > instanceVariableNames: 'imageUrl' > > Why are you *not* storing objectNumber inside each Painting? > What is the relationship between paintings and objectNumbers? > Is it one-to-many or one-to-one? > > ------------------- > !Paintings methodsFor: 'instance creation' stamp: 'RoelofWobben 3/11/2019 > 08:05'! > initialize > super initialize. > numbers := OrderedCollection new! ! > > Why do you *not* initialize variable "paintings" to "OrderedCollection > new" ? > Aren't you dealing with a collection of paintings? > > ---------------- > There is lots more confusing me here, so before going further, > in a fresh new image, can you load NeoJSON and the attached fileout. > > Then in Playground do... > collectionUrl := ' > https://www.rijksmuseum.nl/api/nl/collection?key=14OGzuak&format=json&type=… > '. > json := NeoJSONReader fromString: (ZnEasy get: collectionUrl) contents. > paintings := Paintings fromJSON: json. > paintings inspect. > > and then debug trace through that whole thing. > Is that what you were trying to do? Or have I got the wrong end of the > stick? > > cheers -ben > > > > Oke, I try to tell what I want > > Right now I only want to fetch the objectNumber is a collection. > > so I can use that in a later step. > > There is a one to one relationship between the objectNumber and a Painting > > I do not use the imageUrl found on this url because it's to big . > > So what I try now is first put all the objectNumber in a collection > and use that collection on a second url to fetch a image of the right size > > I hope im clear what I try to do in this step > > so numbers should be a collection of objectNumbers and paintings a > collection of painting with the info I need > > so to recap > > this one must give me a collection of objectNumbers : ' > https://www.rijksmuseum.nl/api/nl/collection?key=14OGzuak&format=json&type=… > '. > > and the objectNumbers I can use for this url : > > https://www.rijksmuseum.nl/api/nl/collection/<one objectNumber/tiles?key=14OGzuak&format=json > > and I do not store the objectNumber in Paintings because I do not use it to display it in a seaside project. > > Roelof > > >

March 13, 2019

Re: [Pharo-users] Hmmm sending at:put: to an undefined object (nil) gives a confusing error message

by Richard O'Keefe

Putting public methods in Object that it cannot honestly support makes #respondsTo: pretty unreliable. It is an obsolete practice, because having Traits means that those methods can be mixed in with (part of) a single line. On Wed, 13 Mar 2019 at 03:20, Ben Coman <btc(a)openinworld.com> wrote: > On Tue, 12 Mar 2019 at 21:56, Marcus Denker <marcus.denker(a)inria.fr> > wrote: > > > > Hi, > > > > Yes, I do not like itâ¦ I think it is there so that you can just make a > variable subclass and at: / at:put: will work without > > having to re-implement themâ¦ > > So we should consider... > * how often are variable subclasses created ? by novices or experts ? > * how much work is it to reimplement those methods ? how difficult to > know what needs to be done ? > * what is the balance > > cheers -ben > > > > > > Marcus > > > > On 12 Mar 2019, at 04:55, Richard O'Keefe <raoknz(a)gmail.com> wrote: > > > > I understand #basicAt:[put:] being in Object, but I never understood > #at:[put:] being there. > > GNU Smalltalk: > > st> nil at: 1 > > Object: nil error: Invalid value nil: object not indexable > > In my Smalltalk, you get a DNU. > > In Squeak you get a debugger window with title > > Error: instances of UndefinedObject are not indexable > > The same for Boolean, Character, and Integer, > > except Squeak where (20 factorial at: 1) answers 0 (oh dear oh dear oh > dear). > > In VW you get a "Subscript out of bounds" error, which is disappointing, > > as do true and $a, but a LargePositiveinteger gets #shouldNotImplement. > > > > > > On Tue, 12 Mar 2019 at 15:28, Tim Mackinnon <tim(a)testit.works> wrote: > >> > >> If you forget to initialise a variable which you thought was a > Dictionary - you get a confusing error message: "Error: only integers > should be used as indicesâ if you try an at:put: > >> > >> This is a consequence of having at:put: defined on Object (which is a > bit nasty) > >> > >> Should UndefinedObject at least override this and signal something a > bit more obvious? This is a common and easily done thing and we don't > support it very well. > >> > >> I just did it in some code, and was scratching my head initially until > I read the stack properly and realised it was something much simpler than I > thought I had done. > >> > >> Any thoughts on this? > > > > > >

March 12, 2019

Re: [Pharo-users] complex json parsing.

by Roelof Wobben

March 12, 2019

TelePharo in Tonel

by Torsten Bergmann

Hi Denis, would it be possible to convert TelePharo to Tonel? https://github.com/pharo-ide/TelePharo Currently one gets load problem on Windows due to long file names and Tonel would solve that. Thanks T.

March 12, 2019

Re: [Pharo-users] doWithIndex: or withIndexDo: ?

by Richard O'Keefe

Neither. No two ways about it. #keysAndValuesDo: However, seq with: seq keys do: aBlock and seq withIndexDo: aBlock do the same thing, so #withIndexDo: makes some sense. This is not my opinion alone. I'm typing this on a machine that has Squeak but not Pharo. In Squeak we find doWithIndex: elementAndIndexBlock "Use the new version with consistent naming" ^ self withIndexDo: elementAndIndexBlock "New" means since 1997. On Wed, 13 Mar 2019 at 04:27, Oleksandr Zaytsev <olk.zaytsev(a)gmail.com> wrote: > Hello, > > I am working on a new collection and this question caught my interest. > Between doWithIndex: or withIndexDo:, which do you find more logical and > why? > > Oleks >

March 12, 2019

Re: [Pharo-users] complex json parsing.

by Roelof Wobben

March 12, 2019

Re: [Pharo-users] complex json parsing.

by Ben Coman

On Wed, 13 Mar 2019 at 00:43, Roelof Wobben <r.wobben(a)home.nl> wrote: > I did try to make this idea work but now im complete lost. > > Right now the collection that should contain the objectNumbers is total empty > anyone who can see where I went wrong. > > I included my code so far. A few things... !Paintings class methodsFor: 'instance creation' stamp: 'RoelofWobben 3/12/2019 17:26'! getImagesDataFromJSON: json | instance | instance := Painting new. (json at: #values) do: [ :artObjectJson | | painting | painting := Painting new. painting imageUrl: (json at: #name). instance addPainting: painting ]. ^ instance! ! You are adding a painting to a painting, which semantically makes no sense. i.e. where #addPainting: is sent to "instance" , that is holding a Painting. I would not expect your Painting class to have a method addPainting: Paintings should have that method. btw, Stop using "instance" as a variable name. Its as bad as variables named "a", "i", "t", "z" - it provides no semantic context. -------------------- Object subclass: #Paintings instanceVariableNames: 'paintings numbers' Paintings class instanceVariableNames: 'numbers paintings'! You've defined these variables on both the instance-side and class-side. I'll hazard a guess that you should delete the class-side definition. ---------------------------------- TBApplicationRootComponent initialize!Object subclass: #Painting instanceVariableNames: 'imageUrl' Why are you *not* storing objectNumber inside each Painting? What is the relationship between paintings and objectNumbers? Is it one-to-many or one-to-one? ------------------- !Paintings methodsFor: 'instance creation' stamp: 'RoelofWobben 3/11/2019 08:05'! initialize super initialize. numbers := OrderedCollection new! ! Why do you *not* initialize variable "paintings" to "OrderedCollection new" ? Aren't you dealing with a collection of paintings? ---------------- There is lots more confusing me here, so before going further, in a fresh new image, can you load NeoJSON and the attached fileout. Then in Playground do... collectionUrl := ' https://www.rijksmuseum.nl/api/nl/collection?key=14OGzuak&format=json&type=… '. json := NeoJSONReader fromString: (ZnEasy get: collectionUrl) contents. paintings := Paintings fromJSON: json. paintings inspect. and then debug trace through that whole thing. Is that what you were trying to do? Or have I got the wrong end of the stick? cheers -ben

March 12, 2019

Re: [Pharo-users] Pharo 7.0 image size

by Trussardi Dario Romano

Ciao, About Hilaire: > Beside the browser windows, did you close all Transcript windows? Yes. > Hi, > > I am confused by what you are reporting. I apologize, i'm trying to understand better. > > If you add 10.000 entries to the Transcript, then I don't see how that results in 1000s of those Strings being alive (and certainly not after a GC). > > I tried the following (both with the Transcript window open and closed): > > Transcript. > Transcript open. > Transcript clear. > > 1 to: 1e4 do: [ :each | > Transcript crShow: DateAndTime now asString , ' MARKTEST ' , each asString ]. > > 3 timesRepeat: [ Smalltalk garbageCollect ]. > > [ | count | > count := 0. > String allSubInstancesDo: [ :each | > (each includesSubstring: 'MARKTEST') ifTrue: [ count := count + 1 ] ]. > count ] value. I did some tests following the code you reported. The thing I noticed is that the image size after some entry, from 144MB is splashed at 272MB. +10 MB at each step: 1 to: 1e4 do: [ :each | Transcript crShow: DateAndTime now asString , ' MARKTEST ' , each asString ]. > > The count never went above 100. I am using Pharo 8 but 7 would be similar. > > If you are retaining so many of the messages that you wrote to the Transcript, then you must be doing something else. I haven't been able to document it for now, but I had the feeling that the following code creates problems: testImageSizeCreateTranscriptEntryB " Create some transcript entry " | count | count := 0. [ 1 to: 1e4 do:[ :each | [ Transcript crShow: DateAndTime now asString, ' DARIOTEST ', each asString. 1 / ( 0 to: 10 ) atRandom ] on: Error do:[ :ex | count := count +1. ex retry ] ] ]value. ^ count Could it be that the retry creates, has created problems? Thanks, Dario > > That being said, we do have an issue with the WriteStream held by the stream instance variable inside ThreadSafeTranscript. It is being reset from time to time and when #clear is called, but that does not bring down the size of the underlying collection below its 'one time peak'. > > Sven > >> On 8 Mar 2019, at 11:59, Trussardi Dario Romano <dario.trussardi(a)tiscali.it> wrote: >> >> Ciao, >> >> this morning i started with an image of about 240 MB. >> >> I have 3 System Browser open ( for a long time ). >> >> I close all the System Browser and save the image. >> >> The image size after the save is about 131 MB. >> >> The new SpaceTally printSpaceAnalysis report: >> >> Class code space # instances inst space percent inst average size >> Array 4115 561766 37353480 25.60 66.49 >> ByteString 2942 249761 31313136 21.50 125.37 >> >> >> The Transcript entry reference basically does not change: >> >> ByteString allInstances select:[ :i | i includesSubstring: 'NETWORK' ] >> >> where 'NETWORK' is a string used in the methods to add Transcript entry Transcript show: ...... NETWORK' ......' . >> >> >> The system found 2958 ByteString instances.... ( relative to the NETWORK ) >> >> The system found 22655 ByteString instances.... ( relative to the CASH ) >> >> >> Why are there many references in the system to the ByteString instances, created in the methods to report the data in Transcript ? >> >> Someone to considerations ? >> >> Thanks, >> >> Dario >> >>> Ciao, >>> >>> I thought that using Transcript as a report to analyze the operation of the code was a good thing - solution. >>> >>> Unfortunately, however, the size of the image continues to increase ( and I think it's due to the use of the transcript ) >>> >>> and i can not find the solution to avoid the problem. >>> >>> I could create a file to report everything about the operations code, but then it becomes difficult to manage, or am i wrong? >>> >>> Other solutions - indications - references ? >>> >>> Thanks, >>> >>> Dario >>> >>>> Ciao, >>>> >>>> thanks. >>>> >>>> i have a Pharo 7.0 alpha build 1262. >>>> >>>> I development a Seaside application. >>>> >>>>> May be you have a lot of still open Seaside session, which if I remember >>>>> correctly are automatically shutdown after 10 min when not active. >>>>> >>>>> Looking at Seaside session instances may give clue. >>>> >>>> OK, i clear the seaside session with the relative seaside status Clear action. >>>> >>>> >>>> >>>> But the image size keeps increasing, without an apparent reason. >>>> >>>> The consideration that I can do and that in the method code i often use Transcript show: '.....NETWORK | CASH | .....' to check the correct functioning. >>>> >>>> With the SpaceTally printSpaceAnalysis i note: >>>> >>>> as of February 16th: >>>> >>>> Class code space # in stances inst space percent inst average size >>>> Array 4115 656764 43904528 23.30 66.85 >>>> ByteString 2942 233871 27527272 14.60 117.70 >>>> >>>> >>>> as of March 4th >>>> Class code space # instances inst space percent inst average size >>>> Array 4115 837017 50564552 23.40 60.41 >>>> ByteString 2942 266620 31933216 14.80 119.77 >>>> >>>> Now i do: >>>> >>>> ByteString allInstances select:[ :i | i includesSubstring: 'NETWORK' ] >>>> >>>> where 'NETWORK' is a string used in the Transcript show: ...... NETWORK' ......' report entry. >>>> >>>> >>>> The system found 3214 ByteString instances.... ( relative to the NETWORK ) >>>> >>>> The system found 22650 ByteString instances.... ( relative to the CASH ) >>>> >>>> The system found .............. >>>> >>>> >>>> Because ????? >>>> >>>> >>>> Does the system remember something about the string create for the Transcript show: entry ? >>>> >>>> I need to reset something? >>>> >>>> How can I analyze the situation? >>>> >>>> I follow some aByteString entry with the pointersTo method >>>> >>>> but for now I have not found the solution. >>>> >>>> >>>> >>>> Thanks, >>>> >>>> Dario >>>> >>>> >>> >>> >> >> > >

March 12, 2019

UI Frameworks

by Bob Cowdery

Hi, What is the state of the various UI frameworks. I did play with Squeak many moons ago and used Morphic but I wanted something a bit higher up the food chain. I see in Pharo, Spec seems to be the built in framework for widgets but I don't see any graphics capability in there. There does seem to be a bunch of Graphics- classes but couldn't find any info on those. I also found Block and Brick which say they are next generation, Block being for graphics andÂ Brick for widgets. These are not yet in Pharo and again I couldn't find any info although there is a tutorial for Block but not for Brick. Having forgotten most everything I knew I didn't have much success trying to run the Brick examples. To build a reasonably complex UI today that requires both a good widget collection and decent drawing capability what would the best choice be. Thanks bob

March 12, 2019

Re: [Pharo-users] Smalltalk CI - anyone have any luck excluding tests?

by Tim Mackinnon

FYI: Just bumping this - but on investigation it looks like the Pharo support for Tags breaks in Smalltalk CI - so top level packages with similar names (e.g. Exercism and ExercismWIP) all seem to get included by the include directive of âExercismâ (with no wildcard). This is in an issue: https://github.com/hpi-swa/smalltalkCI/issues/417 It does mean that many of you might be running more tests than you need to, and it might explain strange build failures if you are trying to use exclude. At some point I will have a go at fixing this, if someone doesnât beat me to it. Tim > On 6 Mar 2019, at 09:10, Tim Mackinnon <tim(a)testit.works> wrote: > > As part of Exercism we have to show our language track can build and run tests - and they use Travis - so, pretty awesome we have that tooling. > > So it was pretty straightforward t hook in - and there are enough examples around to support the docs for the .yml and .ston file - however I hit an issue with the inclusion and exclusion of test files. I put an issue on SmalltalkCI - but its sat there for 8days (https://github.com/hpi-swa/smalltalkCI/issues/417 <https://github.com/hpi-swa/smalltalkCI/issues/417>) so Iâm wondering if anyone has experience with this? > > Essentially if I include my other packages âExercismDevâ and âExercismToolsâ (see sample below - Iâve removed them from the include list) then some of my tests run twice (I believe because the exclude of #AllExercismTests doesnât work). I also have some environmental tests that need improving to run more generically (Submit/Download) and again those donât seem to be excluded. > > Itâs like âexclude" simply doesnât work - but I would think this would be reported more widely if this was the case. So maybe I donât understand the overlap of include/exclude etc. Anyone have any thoughts? > > Tim > SmalltalkCISpec { > #loading : [ > SCIMetacelloLoadSpec { > #baseline : 'Exercism', > #directory : 'dev/src', > #load : [ 'dev' ], > > #platforms : [ #pharo ] > } > ], > > #testing : { > #include : { > #packages : [ 'Exercism' ] > }, > #exclude : { > #classes : [ #AllExercismTests, #ExercismTest, #ExercismSubmitTest, #ExercismDownloadTest ], > #packages : [ 'ExercismWIP'] > } > } > } >

March 12, 2019