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

Re: [Pharo-dev] RES: Problem with FileSystem in WIndows

by Sven Van Caekenberghe

> On 30 Apr 2017, at 19:31, Stephane Ducasse <stepharo.self(a)gmail.com> wrote: > > Can you open a bug entry with clear examples? > We should improve this part. > After having the interpretation of strings sucks. I do not like it I prefer > > FileLocator C / 'temp' / 'e8720bb4-b90a-0d00-9b1f-008709e5552b.txt' Indeed. Also, you should never have to write the platform test, an important idea of FileSystem is to prevent that. Have you read the chapter on FileSystem from the Deep Into Pharo book ? Maybe some Windows users can help you better, but I don't recall others having trouble using FileSystem on Windows. > On Sun, Apr 30, 2017 at 6:54 PM, Casimiro de Almeida Barreto <casimiro.barreto(a)gmail.com> wrote: > Hello, > > -----Mensagem original----- > De: Pharo-dev [mailto:pharo-dev-bounces@lists.pharo.org] Em nome de Sven Van Caekenberghe > Enviada em: domingo, 30 de abril de 2017 05:30 > Para: Pharo Development List <pharo-dev(a)lists.pharo.org> > Assunto: Re: [Pharo-dev] Problem with FileSystem in WIndows > > > > On 30 Apr 2017, at 09:48, Max Leske <maxleske(a)gmail.com> wrote: > > > > Hi Casimiro, > > > > have you tried using normal slashes in the path name? FileSystem should convert those to system path delimiters automatically. So: > > It seems that FileSystem does not convert / into \ . > ((FileSystem root) / 'c:\temp\tempfile.txt') writeStream wonÂ´t work but... > FileStream fileNamed: 'C:\temp\tempfile.txt' works ok. > > So... I could make it work for getting a stream both for Unix and Windows: > > OSPlatform currentPlatformName = 'unix' > ifTrue: [ > [ str := (fs / fn) writeStream ] on: Exception do: [ ^ nil ] ] > ifFalse: [ > [ str := FileStream fileNamed: fn ] on: Exception do: [ ^ nil ] ]. > > And to delete: > > OSPlatform currentPlatformName = 'unix' > ifTrue: [ > (FileSystem root / aFileName) delete ] > ifFalse: [ > rgx := RxMatcher forString: '^[A-Z]\:'. > drive := rgx matchesIn: aFileName. > drive := (drive at: 1) copyWithRegex: '\:' matchesReplacedWith: ''. > ((FileLocator driveNamed: drive) / (aFileName copyWithRegex: '^[A-Z]\:' matchesReplacedWith: '')) delete ] > > Not good, nor elegant nor anything close to good transparent code but... works :'( > > > > > C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt > > > > And possibly you'll need a leading slash, as "C:" is a reference to the root of the file system: > > > > /C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt > > There is also the following option: > > FileLocator C / 'temp' / 'e8720bb4-b90a-0d00-9b1f-008709e5552b.txt' > > > Cheers, > > Max > > > >> On 30 Apr 2017, at 03:15, Casimiro de Almeida Barreto <casimiro.barreto(a)gmail.com> wrote: > >> > >> Good night all. > >> > >> IÂ´ve had trouble in windows concerning FileSystem. > >> > >> ((FileSystem disk root) / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ) delete > >> > >> Fails as if the file didnÂ´t exist. Message: > >> > >> FileDoesNotExist: Path / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ > >> > >> And I wonder whatÂ´s wrong. Because it works in linux and MacOS. > >> > >> Trasncript show: ((FileSYstem disk root) / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ) > >> > >> Returns: > >> > >> File @ C:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txt\ > >> > >> And I wonder where the last \ came from and if is it that is messing all and how to fix things. > >> > >> Best regards, > >> > >> Casimiro Barreto > >> > >> Livre de vÃrus. www.avast.com. > > > > > > > --- > Este email foi escaneado pelo Avast antivÃrus. > https://www.avast.com/antivirus > > >

April 30, 2017

Re: [Pharo-dev] Smalltalk Internet Browser

by Stephane Ducasse

- data manipulation - some math libraries - handling large file (yes you will not load 128 Gb file even in Java :) But so far I failed :) On Sun, Apr 30, 2017 at 7:33 PM, Stephane Ducasse <stepharo.self(a)gmail.com> wrote: > I tried to convince alex to focus on the missing pieces around roassal: > > > On Sun, Apr 30, 2017 at 6:03 PM, volkert <volkert(a)nivoba.de> wrote: > >> +1 >> >> i think the development of pharo should more focus on an "easy to use and >> adaptive live data analysis environment" which supports data science tasks >> with easy to use data retrieving, data preparation, data exploration, >> data modelling, scripting capabilities. We already have so many pieces and >> knowledge for this kind of platform around. I think only the cool story is >> missing .. And yes i know Moose, but i have the feeling this environment is >> not Moose. For me Moose is to much associated with software analysis ... >> >> Volkert >> >> >> >> Am 30.04.2017 um 17:06 schrieb Stephan Eggermont: >> >>> On 29/04/17 04:11, askoh wrote: >>> >>>> Being connected to the internet is going to be a necessity for any >>>> piece of >>>> software in the immediate future. So every Smalltalk development >>>> environment >>>> or application should have that capability as default. To push that >>>> envelop, >>>> every image should have a Smalltalk native Internet Browser. >>>> >>> >>> Web is a very bad platform from an engineering point of view. The amount >>> of accidental complexity is astounding, as is the number of useless layers >>> obfuscating this all. I don't think we are ready to waste the amount of >>> engineering needed to do something useful at the browser implementation >>> side of the web (except for things like amber, pharojs and squeakjs). >>> >>> Stephan >>> >>> >>> >>> >> >> >

April 30, 2017

Re: [Pharo-dev] Smalltalk Internet Browser

by Stephane Ducasse

I tried to convince alex to focus on the missing pieces around roassal: On Sun, Apr 30, 2017 at 6:03 PM, volkert <volkert(a)nivoba.de> wrote: > +1 > > i think the development of pharo should more focus on an "easy to use and > adaptive live data analysis environment" which supports data science tasks > with easy to use data retrieving, data preparation, data exploration, > data modelling, scripting capabilities. We already have so many pieces and > knowledge for this kind of platform around. I think only the cool story is > missing .. And yes i know Moose, but i have the feeling this environment is > not Moose. For me Moose is to much associated with software analysis ... > > Volkert > > > > Am 30.04.2017 um 17:06 schrieb Stephan Eggermont: > >> On 29/04/17 04:11, askoh wrote: >> >>> Being connected to the internet is going to be a necessity for any piece >>> of >>> software in the immediate future. So every Smalltalk development >>> environment >>> or application should have that capability as default. To push that >>> envelop, >>> every image should have a Smalltalk native Internet Browser. >>> >> >> Web is a very bad platform from an engineering point of view. The amount >> of accidental complexity is astounding, as is the number of useless layers >> obfuscating this all. I don't think we are ready to waste the amount of >> engineering needed to do something useful at the browser implementation >> side of the web (except for things like amber, pharojs and squeakjs). >> >> Stephan >> >> >> >> > >

April 30, 2017

Re: [Pharo-dev] GarageGlorp

by Stephane Ducasse

tx On Sun, Apr 30, 2017 at 5:36 PM, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: > I will do so as soon as I sit in front of my computer. > > Regards! > > El abr. 29, 2017 3:35 PM, "Stephane Ducasse" <stepharo.self(a)gmail.com> > escribiÃ³: > >> Hi >> >> for a tutorial I wanted to load Glorp in Pharo 60 as described in the >> Glorp doc. >> >> Metacello new >> smalltalkhubUser: 'DBXTalk' project: 'Garage'; >> configuration: 'GarageGlorp'; >> version: #stable; >> load. >> >> I get an error stating that >> The symbolic version stable is not defined in the configuration for the >> ... for the platform. >> >> I changed the loading instruction to >> >> Metacello new >> smalltalkhubUser: 'DBXTalk' project: 'Garage'; >> configuration: 'GarageGlorp'; >> version: '0.2'; >> load. >> >> And it seems to work. May be we should publish a symbolic version for >> pharo 60. >> I do not think that I have commit right. >> >> Stef >> >

April 30, 2017

Re: [Pharo-dev] RES: Problem with FileSystem in WIndows

by Stephane Ducasse

Can you open a bug entry with clear examples? We should improve this part. After having the interpretation of strings sucks. I do not like it I prefer FileLocator C / 'temp' / 'e8720bb4-b90a-0d00-9b1f-008709e5552b.txt' On Sun, Apr 30, 2017 at 6:54 PM, Casimiro de Almeida Barreto < casimiro.barreto(a)gmail.com> wrote: > Hello, > > -----Mensagem original----- > De: Pharo-dev [mailto:pharo-dev-bounces@lists.pharo.org] Em nome de Sven > Van Caekenberghe > Enviada em: domingo, 30 de abril de 2017 05:30 > Para: Pharo Development List <pharo-dev(a)lists.pharo.org> > Assunto: Re: [Pharo-dev] Problem with FileSystem in WIndows > > > > On 30 Apr 2017, at 09:48, Max Leske <maxleske(a)gmail.com> wrote: > > > > Hi Casimiro, > > > > have you tried using normal slashes in the path name? FileSystem should > convert those to system path delimiters automatically. So: > > It seems that FileSystem does not convert / into \ . > ((FileSystem root) / 'c:\temp\tempfile.txt') writeStream wonÂ´t work but... > FileStream fileNamed: 'C:\temp\tempfile.txt' works ok. > > So... I could make it work for getting a stream both for Unix and Windows: > > OSPlatform currentPlatformName = 'unix' > ifTrue: [ > [ str := (fs / fn) writeStream ] on: Exception do: [ ^ nil > ] ] > ifFalse: [ > [ str := FileStream fileNamed: fn ] on: Exception do: [ ^ > nil ] ]. > > And to delete: > > OSPlatform currentPlatformName = 'unix' > ifTrue: [ > (FileSystem root / aFileName) delete ] > ifFalse: [ > rgx := RxMatcher forString: '^[A-Z]\:'. > drive := rgx matchesIn: aFileName. > drive := (drive at: 1) copyWithRegex: '\:' > matchesReplacedWith: ''. > ((FileLocator driveNamed: drive) / (aFileName > copyWithRegex: '^[A-Z]\:' matchesReplacedWith: '')) delete ] > > Not good, nor elegant nor anything close to good transparent code but... > works :'( > > > > > C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt > > > > And possibly you'll need a leading slash, as "C:" is a reference to the > root of the file system: > > > > /C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt > > There is also the following option: > > FileLocator C / 'temp' / 'e8720bb4-b90a-0d00-9b1f- > 008709e5552b.txt' > > > Cheers, > > Max > > > >> On 30 Apr 2017, at 03:15, Casimiro de Almeida Barreto < > casimiro.barreto(a)gmail.com> wrote: > >> > >> Good night all. > >> > >> IÂ´ve had trouble in windows concerning FileSystem. > >> > >> ((FileSystem disk root) / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ) > delete > >> > >> Fails as if the file didnÂ´t exist. Message: > >> > >> FileDoesNotExist: Path / âC:\temp\e8720bb4-b90a-0d00- > 9b1f-008709e5552b.txtâ > >> > >> And I wonder whatÂ´s wrong. Because it works in linux and MacOS. > >> > >> Trasncript show: ((FileSYstem disk root) / âC:\temp\e8720bb4-b90a-0d00- > 9b1f-008709e5552b.txtâ) > >> > >> Returns: > >> > >> File @ C:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txt\ > >> > >> And I wonder where the last \ came from and if is it that is messing > all and how to fix things. > >> > >> Best regards, > >> > >> Casimiro Barreto > >> > >> Livre de vÃrus. www.avast.com. > > > > > > > --- > Este email foi escaneado pelo Avast antivÃrus. > https://www.avast.com/antivirus > > >

April 30, 2017

RES: Problem with FileSystem in WIndows

by Casimiro de Almeida Barreto

Hello, -----Mensagem original----- De: Pharo-dev [mailto:pharo-dev-bounces@lists.pharo.org] Em nome de Sven Van Caekenberghe Enviada em: domingo, 30 de abril de 2017 05:30 Para: Pharo Development List <pharo-dev(a)lists.pharo.org> Assunto: Re: [Pharo-dev] Problem with FileSystem in WIndows > On 30 Apr 2017, at 09:48, Max Leske <maxleske(a)gmail.com> wrote: > > Hi Casimiro, > > have you tried using normal slashes in the path name? FileSystem should convert those to system path delimiters automatically. So: It seems that FileSystem does not convert / into \ . ((FileSystem root) / 'c:\temp\tempfile.txt') writeStream wonÂ´t work but... FileStream fileNamed: 'C:\temp\tempfile.txt' works ok. So... I could make it work for getting a stream both for Unix and Windows: OSPlatform currentPlatformName = 'unix' ifTrue: [ [ str := (fs / fn) writeStream ] on: Exception do: [ ^ nil ] ] ifFalse: [ [ str := FileStream fileNamed: fn ] on: Exception do: [ ^ nil ] ]. And to delete: OSPlatform currentPlatformName = 'unix' ifTrue: [ (FileSystem root / aFileName) delete ] ifFalse: [ rgx := RxMatcher forString: '^[A-Z]\:'. drive := rgx matchesIn: aFileName. drive := (drive at: 1) copyWithRegex: '\:' matchesReplacedWith: ''. ((FileLocator driveNamed: drive) / (aFileName copyWithRegex: '^[A-Z]\:' matchesReplacedWith: '')) delete ] Not good, nor elegant nor anything close to good transparent code but... works :'( > > C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt > > And possibly you'll need a leading slash, as "C:" is a reference to the root of the file system: > > /C:/temp/e8720bb4-b90a-0d00-9b1f-008709e5552b.txt There is also the following option: FileLocator C / 'temp' / 'e8720bb4-b90a-0d00-9b1f-008709e5552b.txt' > Cheers, > Max > >> On 30 Apr 2017, at 03:15, Casimiro de Almeida Barreto <casimiro.barreto(a)gmail.com> wrote: >> >> Good night all. >> >> IÂ´ve had trouble in windows concerning FileSystem. >> >> ((FileSystem disk root) / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ) delete >> >> Fails as if the file didnÂ´t exist. Message: >> >> FileDoesNotExist: Path / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ >> >> And I wonder whatÂ´s wrong. Because it works in linux and MacOS. >> >> Trasncript show: ((FileSYstem disk root) / âC:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txtâ) >> >> Returns: >> >> File @ C:\temp\e8720bb4-b90a-0d00-9b1f-008709e5552b.txt\ >> >> And I wonder where the last \ came from and if is it that is messing all and how to fix things. >> >> Best regards, >> >> Casimiro Barreto >> >> Livre de vÃrus. www.avast.com. > --- Este email foi escaneado pelo Avast antivÃrus. https://www.avast.com/antivirus

April 30, 2017

Re: [Pharo-dev] Smalltalk Internet Browser

by Dimitris Chloupis

> And IMHO, electron should be killed!! It's really a bad bad idea. You have > like 10 single apps that run a full V8/node stack. Atom, Discord, > LightTable, Chrome itself...it's so RAM & CPU consuming. For me it is like > saying that a desktop application is pure front-end development. > best > The appeal of electron is not that it will be very good performance wise but that its much easier for a web dev to move his/her web app to desktop with minimum amount of effort. Of course minimum effort has the side effect of lousy performance but in the case of apps like Discord which they do not do much its ok. So electron makes sense. Whether we should support such a framework is something that is up to the Pharo devs themselves. Take a look at Ruby on Rails, it was started from a single person for his own needs. Like many of our own Pharo libraries. You make something you need, you release it and if people find it useful they will contribute , if not its still useful to you ;) Such libs/tools like Electron have zero appeal to us desktop developers mainly because even web devs do not like web development for the known problem it has. Also web devs tend to over exaggerate about the appeal of individual web frameworks , I read an article making the bold claims of massive success for Node.js as a server API , I did some research and I came in front page evidence of massive failure, servers remain firmly committed to Apache which yes its written in C. Not that I doubt for a second that an API for a dynamic language would be able to compete with the behemoth of performance that C really is on its own field, obviously servers needs top performance because of the big load they have to manage. I enjoy ninja attacking such web posts and expose them for what they really are, fake. The other day another web dev made the bold claim that web games are on the rise, took me a second to do my research and find one of the most highly regarded research surveys in game industry that clearly show not only that web games are no more than 4% but they going straight to the hell of 0%. Apparently the one ones excited to play games inside the web browser are web devs that want to make blog post with ridiculous claims. Web remains and will remain strictly inside the web browser, outside it, very few people like it and actually choose it as they are preferred way of development. The same way none would pick C/C++ for a non performance orientated app well at least most logical people, unless of course there is a very good reason for such a choice. Of course C devs do not feel the need , at least AFAIK , to exaggerate the usefulness of their own language in blog posts with zero reference to actual reliable sources. But they are no angels either. Personally I think the future of the Web which will only grow in popularity will be outside the browser and already desktop apps have taken advantage of this. We see this a lot with mobile apps that steal away the users of web browsers. How many people use the web browser to like and chat on facebook compared to the ones using the native mobile clients ? Not many.

April 30, 2017

Re: [Pharo-dev] Smalltalk Internet Browser

by volkert

+1 i think the development of pharo should more focus on an "easy to use and adaptive live data analysis environment" which supports data science tasks with easy to use data retrieving, data preparation, data exploration, data modelling, scripting capabilities. We already have so many pieces and knowledge for this kind of platform around. I think only the cool story is missing .. And yes i know Moose, but i have the feeling this environment is not Moose. For me Moose is to much associated with software analysis ... Volkert Am 30.04.2017 um 17:06 schrieb Stephan Eggermont: > On 29/04/17 04:11, askoh wrote: >> Being connected to the internet is going to be a necessity for any >> piece of >> software in the immediate future. So every Smalltalk development >> environment >> or application should have that capability as default. To push that >> envelop, >> every image should have a Smalltalk native Internet Browser. > > Web is a very bad platform from an engineering point of view. The > amount of accidental complexity is astounding, as is the number of > useless layers obfuscating this all. I don't think we are ready to > waste the amount of engineering needed to do something useful at the > browser implementation side of the web (except for things like amber, > pharojs and squeakjs). > > Stephan > > >

April 30, 2017

Re: [Pharo-dev] GarageGlorp

by Esteban A. Maringolo

I will do so as soon as I sit in front of my computer. Regards! El abr. 29, 2017 3:35 PM, "Stephane Ducasse" <stepharo.self(a)gmail.com> escribiÃ³: > Hi > > for a tutorial I wanted to load Glorp in Pharo 60 as described in the > Glorp doc. > > Metacello new > smalltalkhubUser: 'DBXTalk' project: 'Garage'; > configuration: 'GarageGlorp'; > version: #stable; > load. > > I get an error stating that > The symbolic version stable is not defined in the configuration for the > ... for the platform. > > I changed the loading instruction to > > Metacello new > smalltalkhubUser: 'DBXTalk' project: 'Garage'; > configuration: 'GarageGlorp'; > version: '0.2'; > load. > > And it seems to work. May be we should publish a symbolic version for > pharo 60. > I do not think that I have commit right. > > Stef >

April 30, 2017

Re: [Pharo-dev] Smalltalk Internet Browser

by Stephan Eggermont

On 29/04/17 04:11, askoh wrote: > Being connected to the internet is going to be a necessity for any piece of > software in the immediate future. So every Smalltalk development environment > or application should have that capability as default. To push that envelop, > every image should have a Smalltalk native Internet Browser. Web is a very bad platform from an engineering point of view. The amount of accidental complexity is astounding, as is the number of useless layers obfuscating this all. I don't think we are ready to waste the amount of engineering needed to do something useful at the browser implementation side of the web (except for things like amber, pharojs and squeakjs). Stephan

April 30, 2017