December 2018 - Pharo-users

Re: [Pharo-users] GPS / Longitude / Latitude library

by Richard O'Keefe

As Dershowitz and Reingold showed in their book "Calendrical Calculations", converting from UTC to several calendars needs latitude, longitude, and elevation. ISO 6709 is the relevant standard: https://en.wikipedia.org/wiki/ISO_6709 On Tue, 4 Dec 2018 at 23:35, CÃ©drick BÃ©ler <cdrick65(a)gmail.com> wrote: > Interesting. > > Iâm interested with such tools too. > > I ask students to do gps point collection from mobile (Cordova) and > process them in pharo to detect if they enter known places. > > I wonder if a proper gps point class would be useful. Iâd like to record > also the precision. > > Cheers, > > Cedrick > > > > > Le 4 dÃ©c. 2018 Ã 11:15, Alistair Grant <akgrant0710(a)gmail.com> a Ã©crit : > > > > Hi Sven, > > > > Thanks. This is just for me playing around with my photo collection, > > i.e. find all photos from a particular location, i.e. within a > > specified distance of a point. Also maybe add GPS coordinates to some > > photos. > > > > Thanks again, > > Alistair > > > >> On Tue, 4 Dec 2018 at 11:04, Sven Van Caekenberghe <sven(a)stfx.eu> > wrote: > >> > >> Hi Alistair, > >> > >>> On 4 Dec 2018, at 10:21, Alistair Grant <akgrant0710(a)gmail.com> wrote: > >>> > >>> Hi, > >>> > >>> Does anyone know of a library for processing GPS coordinates? > >>> > >>> What I'm looking for are things like: > >>> > >>> - Parsing from and printing to various string formats (HMS, NESW, > decimal) > >>> - Distance between two points > >>> - etc. > >>> > >>> Thanks, > >>> Alistair > >> > >> We've got some elementary stuff based on WGS84 coordinates as points. > For example, > >> > >> T3GeoTools distanceBetween: 5.33732(a)50.926 and: 5.49705(a)50.82733. > >> T3GeoTools bearingFrom: 5.33732(a)50.926 to: 5.49705(a)50.82733. > >> T3GeoTools destinationFrom: 5.33732(a)50.926 bearing: 45 distance: 2500. > >> T3GeoTools centroidOf: { 5.48230(a)50.82249. 5.49523(a)50.81288. > 5.50138(a)50.82008. 5.50228(a)50.82595. 5.49265(a)50.82560. 5.48230(a)50.82249 }. > >> T3GeoTools is: 5.33732(a)50.92601 inside: { 5.48230(a)50.82249. > 5.49523(a)50.81288. 5.50138(a)50.82008. 5.50228(a)50.82595. 5.49265(a)50.82560. > 5.48230(a)50.82249 }. > >> > >> This is not open source, but it is not rocket science either (just > implementations of public algorithms). > >> > >> We have less need for special conversions, we do them case by case when > they occur. > >> > >> What are you planning to do ? > >> > >> Sven > >> > >> > > > >

Dec. 5, 2018

Re: [Pharo-users] can I avoid the nested loops

by Richard O'Keefe

First, this is a computational geometry problem, as is the second part. If this were a serious problem or had a huge amount of data, we would be looking for a good data structure for holding a collection of rectangles. We might consider quad-trees, k-d-trees, R-trees, ... With such a data structure, we might hope for O(n.log n) time where n is the number of claims. Your approach is O(sum(claim areas)**2). Since the sum of the claim areas is about half a million square inches, O(sum(claim areas)) would not be a problem, but squaring that most definitely is. The nested loops are not really a problem. What bothers me a lot is your choice of data structures. Representing a point by (a asString , '*' , b asString) is, um, not a good idea. And I don't care if it is C, C++, C#, Java, Objective C, Eiffel, Ada, Fortran, Javascript, Ruby, Lisp, Smalltalk, or almost anything but AWK: STRINGS ARE WRONG. Strings have their uses in input/output. Nietzsche wrote: âI teach you the Superman. Man is a thing to be overcome â¦ What is the ape to man? A jest or a thing of shame. So shall man be to the Superman â a rope over the abyss.â STRINGS ARE A THING TO BE OVERCOME. A string where you need high performance is a jest or a thing of shame. Alan J. Perlis wrote: 'The string is a stark data structure and everywhere it is passed there is much duplication of process. It is a perfect vehicle for hiding information.' If you need to process information fast, get it *out* of string form as early as possible, and back into string format for output as late as possible. In this case, several programming languages already have a built-in data structure for points. In Smalltalk, it is called Point, and (a @ b) will make one. When you need to represent a point, why not use a Point? When you need to represent a rectangle, why not use a Rectangle? Moreover, there is something you do which is *scary*. You do a LINEAR SEARCH in the OrderedCollection "area". Again, this really has very little to do with the programming language. You could use a Vector or an ArrayList in Java and the problem would be the same. A few linear searches, or a lot of searches in a small sequence, no worries. A lot of searches in a big sequence, OUCH. Considering both parts of the problem, you need to be able to distinguish three different states for any given cell: - no rectangle covers this cell - one rectangle covers this cell - two or more rectangles cover this cell What you want is some sort of two-dimensional sparse array mapping x,y coordinates to counts. One good answer is a hash table. Use map : Dictionary[Integer, Dictionary[Integer, Integer]] x y count read the claims as a Dictionary[Integer, Rectangle] id left@top extent: (width-1)@(height-1) create an empty Dictionary, map. for each rectangle in claims for each x from rectangle left to rectangle right row := map at: x ifAbsentPut: [Dictionary new]. for each y from rectangle top to rectangle bottom increment (row at: y). let total be the sum over the map of "use #inject:into:" the number of cells in the row "use #count:" where the value in the cell is more than one.

Dec. 5, 2018

Re: [Pharo-users] GPS / Longitude / Latitude library

by Pierce Ng

On Tue, Dec 04, 2018 at 10:21:20AM +0100, Alistair Grant wrote: > Does anyone know of a library for processing GPS coordinates? > > What I'm looking for are things like: > > - Parsing from and printing to various string formats (HMS, NESW, decimal) > - Distance between two points > - etc. SQLite-based possibilities: - Spatialite: https://www.gaia-gis.it/fossil/libspatialite/index - R*Tree: https://sqlite.org/rtree.html - GeoPoly: https://sqlite.org/geopoly.html

Dec. 5, 2018

Re: [Pharo-users] can I avoid the nested loops

by phil＠highoctane.be

A claim is a claim ID + a rectangle. "If the Elves all proceed with their own plans, none of them will have enough fabric. How many square inches of fabric are within two or more claims?" <--- actual question. units are inches. So, question translates to: "What is the total area of rectangle intersections?" Now you can see it as rectangles or as all measurements are integers, are cells in a matrix (rasterizing the geometry). That's the approach I took and it worked well enough. HTH Phil On Tue, Dec 4, 2018 at 11:05 PM Roelof Wobben <r.wobben(a)home.nl> wrote: > Thanks I found that out already. > > Now trying to understand what they mean with this exactly > > If the Elves all proceed with their own plans, none of them will have enough fabric. How many square inches of fabric are within two or more claims? > > > They give some 13000 plans > > > Roelof > > > > > > Op 4-12-2018 om 22:55 schreef Ben Coman: > > > > On Wed, 5 Dec 2018 at 00:45, Roelof Wobben <r.wobben(a)home.nl> wrote: > >> so like this ? >> >> so first the intersect between rectangle1 and rectangle2 >> store it >> and then the intersection between the stored one and rectangle3 >> and so on >> then its time to look how I must define a rectangle when I know the >> starting point and the length and width >> >> then it time to figure out how I can define a Rectangle with the starting >> point the length and the width >> > > Use Spotter to bring up the System Browser on the Rectangle class. > Then right-click >> Class-refs > to look at how the class is used directly to create new instances. > > cheers -ben > > > >> >> Roelof >> >> >> >> Op 4-12-2018 om 17:27 schreef phil(a)highoctane.be: >> >> Use Rectangle class and intersect: method? >> >> >> >> >> On Tue, Dec 4, 2018 at 5:07 PM Roelof Wobben <r.wobben(a)home.nl> wrote: >> >>> Hello, >>> >>> For adventofCode I had to write some code that calculates the overlap >>> that multiple rectangles have. >>> >>> Fo far I have it working for 1 rectangle but I do not like the code >>> because of the nested loops. >>> >>> Can I make this more the pharo way >>> >>> Code so far : >>> >>> | item coordinates area overlap regexp line beginX endX beginY endY| >>> area := OrderedCollection new. >>> overlap := OrderedCollection new. >>> regexp := '\#(\d+) @ (\d+),(\d+)\: (\d+)x(\d+)' asRegex. >>> line := '#1 @ 1,3: 4x4'. >>> (regexp matches: line) >>> ifTrue: [ >>> beginX := (regexp subexpression: 2) asInteger + 1. >>> endX := (regexp subexpression: 5) asInteger + beginX . >>> beginY := (regexp subexpression: 4) asInteger + 1. >>> endY := (regexp subexpression: 6) asInteger + beginY. >>> beginX to: (endX - 1) do: [:a | >>> beginY to: (endY -1) do: [:b | >>> item := a asString , '*' , b asString . >>> (area includes: item) >>> ifTrue: [ overlap add: item ] >>> ifFalse: [ area add: item ]]]] . >>> ^ area. >>> >>> Thanks, >>> >>> Roelof >>> >>> >>> >> >

Dec. 4, 2018

Re: [Pharo-users] can I avoid the nested loops

by Roelof Wobben

Dec. 4, 2018

Re: [Pharo-users] Pharo 7 Monticello browser and github repositories

by Ben Coman

On Wed, 5 Dec 2018 at 00:45, Sanjay Minni <sm(a)planage.com> wrote: > I get this problem when trying to install a package (VoyageMongo) thru > Catalog Browser > > Pharo 7.0.0rc1 on Win 10 32 bit fresh image->Catalogue Browser > Select VoyageMongo -><right click>-> install stable version > > why does the installation just abort with the message: > <http://forum.world.st/file/t368721/CatalogBrowserVoyageMongoDies01.jpg> > should it not come up a debugger walk back or at least with more > information. > it only comes up with a debugger if the error was unhandled. General philosophy is its better to handle them - particular in a end-user app but on the flip side as developers we get very used to getting insights with the debugger. @esteban, Probably a button to also open a debugger would be useful for Iceberg, which is a "developer" tool. > Iceberg status is: > <http://forum.world.st/file/t368721/Iceberg01.jpg> > Is this due to the other error post you made? After trying the hack I suggested, open the "Repository" on dialog of voyage, and re-check out the branch, and/or another branch. cheers -ben P.S. Thanks for giving Iceberg a shot. As it gets used in more real world and varied environments, it helps us iron out the edge cases. > > Ben Coman wrote > > On Mon, 3 Dec 2018 at 01:48, Sanjay Minni via Pharo-users < > > > pharo-users@.pharo > > >> wrote: > > > >> How to add new github repositories in Monticello (Browser) in Pharo 7 > >> (e.g to include the repository for Voyage) > >> > >> Monticello browser does show github repositories - how to edit / add new > >> ones ? > >> > > > > Please try Iceberg. Your feedback as a fresh Iceberg user would be > > useful. > > Sorry I don't have any links handy, but it would be interesting to see > > what > > you can work out intuitively. > > > > One thing though, if you are using Windows you'll need to "Use custom SSH > > keys..." > > (if you are using ssh based protocol) > > > > cheers -ben > > > > > > ----- > cheers, > Sanjay > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > >

Dec. 4, 2018

Re: [Pharo-users] can I avoid the nested loops

by Ben Coman

On Wed, 5 Dec 2018 at 00:45, Roelof Wobben <r.wobben(a)home.nl> wrote: > so like this ? > > so first the intersect between rectangle1 and rectangle2 > store it > and then the intersection between the stored one and rectangle3 > and so on > then its time to look how I must define a rectangle when I know the > starting point and the length and width > > then it time to figure out how I can define a Rectangle with the starting > point the length and the width > Use Spotter to bring up the System Browser on the Rectangle class. Then right-click >> Class-refs to look at how the class is used directly to create new instances. cheers -ben > > Roelof > > > > Op 4-12-2018 om 17:27 schreef phil(a)highoctane.be: > > Use Rectangle class and intersect: method? > > > > > On Tue, Dec 4, 2018 at 5:07 PM Roelof Wobben <r.wobben(a)home.nl> wrote: > >> Hello, >> >> For adventofCode I had to write some code that calculates the overlap >> that multiple rectangles have. >> >> Fo far I have it working for 1 rectangle but I do not like the code >> because of the nested loops. >> >> Can I make this more the pharo way >> >> Code so far : >> >> | item coordinates area overlap regexp line beginX endX beginY endY| >> area := OrderedCollection new. >> overlap := OrderedCollection new. >> regexp := '\#(\d+) @ (\d+),(\d+)\: (\d+)x(\d+)' asRegex. >> line := '#1 @ 1,3: 4x4'. >> (regexp matches: line) >> ifTrue: [ >> beginX := (regexp subexpression: 2) asInteger + 1. >> endX := (regexp subexpression: 5) asInteger + beginX . >> beginY := (regexp subexpression: 4) asInteger + 1. >> endY := (regexp subexpression: 6) asInteger + beginY. >> beginX to: (endX - 1) do: [:a | >> beginY to: (endY -1) do: [:b | >> item := a asString , '*' , b asString . >> (area includes: item) >> ifTrue: [ overlap add: item ] >> ifFalse: [ area add: item ]]]] . >> ^ area. >> >> Thanks, >> >> Roelof >> >> >> >

Dec. 4, 2018

Re: [Pharo-users] Voyage

by Ben Coman

On Wed, 5 Dec 2018 at 00:25, Sanjay Minni <sm(a)planage.com> wrote: > Hi Esteban, > > I get a Debug Popup as per below - how do I debug from here > > <http://forum.world.st/file/t368721/VoyageIcebergLoadErr01.jpg> > > <http://forum.world.st/file/t368721/VoyageLoadError01.jpg> > Is this on Windows? It could be that long path names continue to bite. Could you try a hack... go in on the command line and shorten the directory names after "source" ? cheers -ben EstebanLM wrote > > There is indeed a problem because cryptography package is still using an > > old API: #ifNotNilDo: that does not exists anymore in Pharo. > > You are probably seen something like this: > > > > > > > > > > (And this is a bug) > > Anyway, while we fix this, you can replace #ifNotNilDo: for #ifNotNil:, > > then save and select âproceedâ. > > > > That will allow voyage to load properly. > > > > Cheers, > > Esteban > > > > > > > >> On 2 Dec 2018, at 12:39, Sanjay Minni < > > > sm@ > > > > wrote: > >> > >> I am unable to load Voyage-Mongo and Voyage-Memory in Pharo 7.0.0rC1 > >> > >> tried using the scipt below and also thru the Catalog Browser entries > for > >> Voyage > >> any pointers > >> > >> > >> CyrilFerlicot wrote > >>> Le 16/11/2018 Ã 02:02, horrido a Ã©crit : > >>>> I'm trying to load Voyage for MongoDB and I'm having great difficulty. > >>>> The > >>>> docs seem to be out of date. > >>>> > >>>> What are the current instructions for doing this? > >>>> > >>>> The docs also refer to "Configurations Browser" in World Menu/Tools, > >>>> but > >>>> there is no longer any such thing. > >>>> > >>>> > >>> > >>> Hi, > >>> > >>> Voyage-Mongo > >>> > >>> Metacello new > >>> repository: 'github://pharo-nosql/voyage/mc'; > >>> baseline: 'Voyage'; > >>> onConflictUseIncoming; > >>> load: 'mongo tests'. > >>> > >>> Voyage-UnQLite > >>> > >>> Metacello new > >>> repository: 'github://pharo-nosql/voyage/mc'; > >>> baseline: 'Voyage'; > >>> onConflictUseIncoming; > >>> load: 'unqlite tests'. > >>> > >>> Voyage-Memory > >>> > >>> Metacello new > >>> repository: 'github://pharo-nosql/voyage/mc'; > >>> baseline: 'Voyage'; > >>> onConflictUseIncoming; > >>> load: 'memory tests'. > >>> > >>> Should work. > >>> > >>>> > >>>> -- > >>>> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > >>>> > >>> > >>> > >>> -- > >>> Cyril Ferlicot > >>> https://ferlicot.fr > >>> > >>> > >>> > >>> signature.asc (836 bytes) > >>> <http://forum.world.st/attachment/5088985/0/signature.asc> > >> > >> > >> > >> > >> > >> ----- > >> cheers, > >> Sanjay > >> -- > >> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > >> > > > > > > > > Screenshot 2018-12-02 at 15.02.19.png (165K) > > < > http://forum.world.st/attachment/5090021/0/Screenshot%202018-12-02%20at%201… > ; > > > > > > ----- > cheers, > Sanjay > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > >

Dec. 4, 2018

Alternative to NautilusMethodSelected in Calypso

by Gustavo Santos

Hello, I'm trying to upgrade MacroRecorder to Pharo7. I used NautilusClassSelected and NautilusMethodSelected announcements to apply a macro in a given class or method right after the developer clicks on the browser. Something like AbstractNautilusPlugin subclass: #MRNautilusPlugin ... registerTo: aModel aModel announcer weak when: NautilusClassSelected send: #entitySelected: to: self; when: NautilusMethodSelected send: #entitySelected: to: self. How can I do the same with Calypso? Cheers, -- Gustavo Santos http://gustavojss.github.io/

Dec. 4, 2018

Re: [Pharo-users] Pharo 7 Monticello browser and github repositories

by Sanjay Minni

I get this problem when trying to install a package (VoyageMongo) thru Catalog Browser Pharo 7.0.0rc1 on Win 10 32 bit fresh image->Catalogue Browser Select VoyageMongo -><right click>-> install stable version why does the installation just abort with the message: <http://forum.world.st/file/t368721/CatalogBrowserVoyageMongoDies01.jpg> should it not come up a debugger walk back or at least with more information. Iceberg status is: <http://forum.world.st/file/t368721/Iceberg01.jpg> Ben Coman wrote > On Mon, 3 Dec 2018 at 01:48, Sanjay Minni via Pharo-users < > pharo-users@.pharo >> wrote: > >> How to add new github repositories in Monticello (Browser) in Pharo 7 >> (e.g to include the repository for Voyage) >> >> Monticello browser does show github repositories - how to edit / add new >> ones ? >> > > Please try Iceberg. Your feedback as a fresh Iceberg user would be > useful. > Sorry I don't have any links handy, but it would be interesting to see > what > you can work out intuitively. > > One thing though, if you are using Windows you'll need to "Use custom SSH > keys..." > (if you are using ssh based protocol) > > cheers -ben ----- cheers, Sanjay -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Dec. 4, 2018