- Pharo-users - lists.pharo.org

Re: [Pharo-users] About Zinc http components

by Esteban A. Maringolo

2014-10-07 4:59 GMT-03:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > Hi Alain, > > Any network protocol between two processes has overhead, even the one you would write yourself. It does make sense to use something that already exists, uses open standards and is used and battle tested by others. > > There is a fundamental difference between HTTP and WebSockets: HTTP is request/response where the initiative lies with the client, WebSockets is fully bidirectional once the connection is setup. There is also less overhead with WebSockets. In both cases you should take care to do your data conversions on/off the wire as efficient as possible. > > I would say: go for it. +1 for each of the paragraphs. Plus in the not so far future HTTP2 [1] might be promoted to a standard, meanwhile SPDY [2] is already usable and deployed, so the connection cost and the protocol transport might become cheaper along faster underlying communication channels. Best regards! [1] https://tools.ietf.org/html/draft-ietf-httpbis-http2-14 [2] http://tools.ietf.org/html/draft-mbelshe-httpbis-spdy-00

Oct. 7, 2014

Re: [Pharo-users] Smalltalk based unikernels

by Rafael Luque

Understood. I'll take a look to the sources then. Thank you! 2014-10-07 12:03 GMT+02:00 Marcus Denker <marcus.denker(a)inria.fr>: > > On 07 Oct 2014, at 12:01, Rafael Luque <rafael.luque.leiva(a)gmail.com> > wrote: > > Thank you Stef, > > Do you know any tutorial or documentation I can follow to get a PharoNOS? > > > Keep in mind that this was an experiment... (kind of a research prototype). > > I don't think that there is a Tutorial. > > Rafa > > 2014-10-07 9:17 GMT+02:00 stepharo <stepharo(a)free.fr>: > >> >> Hello, >>> >>> I'm playing with the unikernels idea [1]. Specifically I'm using Mirage >>> OS [2] to probe a minimal HTTP server running on AWS directly over Xen, >>> without any OS between the application and the hypervisor. >>> >> >> Yes yes yes >> we would love to have a Pharo solution for that. >> >>> >>> Mirage OS is based on OCaml. Is there any Smalltalk solution to build >>> this kind of unikernels? The idea of Smalltalk all the way down (at least >>> to the virtual devices of the hypervisor) is very appealing to me. Indeed I >>> think this idea was proposed for the first time by Dan Ingalls: >>> >>> ''An operating system is a collection of things that don't fit into a >>> language. There shouldn't be one.'' >>> >>> I have found SqueakNOS [3], but I don't know if it is active or if it >>> works with Pharo. >>> >> >> It was working. >> >> >> >>> Thank you, >>> >>> Rafa >>> >>> [1] http://queue.acm.org/detail.cfm?id=2566628 >>> [2] http://www.openmirage.org/ >>> [3] http://squeaknos.blogspot.com.es/ >>> >> >> >> > >

Oct. 7, 2014

Re: [Pharo-users] Smalltalk based unikernels

by Marcus Denker

> On 07 Oct 2014, at 12:01, Rafael Luque <rafael.luque.leiva(a)gmail.com> wrote: > > Thank you Stef, > > Do you know any tutorial or documentation I can follow to get a PharoNOS? > Keep in mind that this was an experimentâ¦ (kind of a research prototype). I donât think that there is a Tutorial. > Rafa > > 2014-10-07 9:17 GMT+02:00 stepharo <stepharo(a)free.fr <mailto:stepharo@free.fr>>: > > Hello, > > I'm playing with the unikernels idea [1]. Specifically I'm using Mirage OS [2] to probe a minimal HTTP server running on AWS directly over Xen, without any OS between the application and the hypervisor. > > Yes yes yes > we would love to have a Pharo solution for that. > > Mirage OS is based on OCaml. Is there any Smalltalk solution to build this kind of unikernels? The idea of Smalltalk all the way down (at least to the virtual devices of the hypervisor) is very appealing to me. Indeed I think this idea was proposed for the first time by Dan Ingalls: > > ''An operating system is a collection of things that don't fit into a language. There shouldn't be one.'' > > I have found SqueakNOS [3], but I don't know if it is active or if it works with Pharo. > > It was working. > > > > Thank you, > > Rafa > > [1] http://queue.acm.org/detail.cfm?id=2566628 <http://queue.acm.org/detail.cfm?id=2566628> > [2] http://www.openmirage.org/ <http://www.openmirage.org/> > [3] http://squeaknos.blogspot.com.es/ <http://squeaknos.blogspot.com.es/> > > >

Oct. 7, 2014

Re: [Pharo-users] Smalltalk based unikernels

by Rafael Luque

Thank you Stef, Do you know any tutorial or documentation I can follow to get a PharoNOS? Rafa 2014-10-07 9:17 GMT+02:00 stepharo <stepharo(a)free.fr>: > > Hello, >> >> I'm playing with the unikernels idea [1]. Specifically I'm using Mirage >> OS [2] to probe a minimal HTTP server running on AWS directly over Xen, >> without any OS between the application and the hypervisor. >> > > Yes yes yes > we would love to have a Pharo solution for that. > >> >> Mirage OS is based on OCaml. Is there any Smalltalk solution to build >> this kind of unikernels? The idea of Smalltalk all the way down (at least >> to the virtual devices of the hypervisor) is very appealing to me. Indeed I >> think this idea was proposed for the first time by Dan Ingalls: >> >> ''An operating system is a collection of things that don't fit into a >> language. There shouldn't be one.'' >> >> I have found SqueakNOS [3], but I don't know if it is active or if it >> works with Pharo. >> > > It was working. > > > >> Thank you, >> >> Rafa >> >> [1] http://queue.acm.org/detail.cfm?id=2566628 >> [2] http://www.openmirage.org/ >> [3] http://squeaknos.blogspot.com.es/ >> > > >

Oct. 7, 2014

Re: [Pharo-users] About Zinc http components

by Sven Van Caekenberghe

Hi Alain, Any network protocol between two processes has overhead, even the one you would write yourself. It does make sense to use something that already exists, uses open standards and is used and battle tested by others. There is a fundamental difference between HTTP and WebSockets: HTTP is request/response where the initiative lies with the client, WebSockets is fully bidirectional once the connection is setup. There is also less overhead with WebSockets. In both cases you should take care to do your data conversions on/off the wire as efficient as possible. I would say: go for it. Sven On 07 Oct 2014, at 01:10, Alain Rastoul <alf.mmm.cat(a)gmail.com> wrote: > Hi, > I played a bit with ZnServer and other zinc components and have > a question I can't answer myself (googling a bit didn't help neither), > and I'm seeking for advice: does it makes sense to use a ZnServer/ZnWebSocket > as a mechanism to transfer data between two pharo processes - in my case 8k ByteArray blocks ? > Or is it a total non sense ? is it reliable ? > First tests looks good : 18k blocks / second (145 Mb/s) on a laptop > core i5 2.6Ghz. > > All comments and suggestions welcome > > Thanks in advance, > > Alain

Oct. 7, 2014

Re: [Pharo-users] Smalltalk based unikernels

by stepharo

> Hello, > > I'm playing with the unikernels idea [1]. Specifically I'm using > Mirage OS [2] to probe a minimal HTTP server running on AWS directly > over Xen, without any OS between the application and the hypervisor. Yes yes yes we would love to have a Pharo solution for that. > > Mirage OS is based on OCaml. Is there any Smalltalk solution to build > this kind of unikernels? The idea of Smalltalk all the way down (at > least to the virtual devices of the hypervisor) is very appealing to > me. Indeed I think this idea was proposed for the first time by Dan > Ingalls: > > ''An operating system is a collection of things that don't fit into a > language. There shouldn't be one.'' > > I have found SqueakNOS [3], but I don't know if it is active or if it > works with Pharo. It was working. > > Thank you, > > Rafa > > [1] http://queue.acm.org/detail.cfm?id=2566628 > [2] http://www.openmirage.org/ > [3] http://squeaknos.blogspot.com.es/

Oct. 7, 2014

About Zinc http components

by Alain Rastoul

Hi, I played a bit with ZnServer and other zinc components and have a question I can't answer myself (googling a bit didn't help neither), and I'm seeking for advice: does it makes sense to use a ZnServer/ZnWebSocket as a mechanism to transfer data between two pharo processes - in my case 8k ByteArray blocks ? Or is it a total non sense ? is it reliable ? First tests looks good : 18k blocks / second (145 Mb/s) on a laptop core i5 2.6Ghz. All comments and suggestions welcome Thanks in advance, Alain

Oct. 6, 2014

Re: [Pharo-users] BLOG: Block Translators - parsing magic

by Udo Schneider

On 06.10.14 11:46, Damien Cassou wrote: > no such guide exist. Do your best and we will do another pass on it afterwards. Ok - good to know. I checked in a version which is IMHO reasonably "clean". I'll check again in two weeks or so - it's sometimes surprising how different something looks when you didn't take a look at it for a while. CU, Udo

Oct. 6, 2014

Smalltalk based unikernels

by Rafael Luque

Hello, I'm playing with the unikernels idea [1]. Specifically I'm using Mirage OS [2] to probe a minimal HTTP server running on AWS directly over Xen, without any OS between the application and the hypervisor. Mirage OS is based on OCaml. Is there any Smalltalk solution to build this kind of unikernels? The idea of Smalltalk all the way down (at least to the virtual devices of the hypervisor) is very appealing to me. Indeed I think this idea was proposed for the first time by Dan Ingalls: ''An operating system is a collection of things that don't fit into a language. There shouldn't be one.'' I have found SqueakNOS [3], but I don't know if it is active or if it works with Pharo. Thank you, Rafa [1] http://queue.acm.org/detail.cfm?id=2566628 [2] http://www.openmirage.org/ [3] http://squeaknos.blogspot.com.es/

Oct. 6, 2014

Re: [Pharo-users] NeoCSVReader skipEmptyLines?

by Esteban A. Maringolo

Excellent! And your implementation is better than simply skipping empty lines :) Thank you Sven! Esteban A. Maringolo 2014-10-06 12:52 GMT-03:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > In #bleedingEdge: > > === > Name: Neo-CSV-Core-SvenVanCaekenberghe.21 > Author: SvenVanCaekenberghe > Time: 6 October 2014, 5:49:01.691696 pm > UUID: a6faeef3-aaa7-460f-a815-8e77b62f9906 > Ancestors: Neo-CSV-Core-SvenVanCaekenberghe.20 > > Added NeoCSVReader>>#addIgnoredFields: > Added NeoCSVReader>>#select:[thenDo:] to skip empty or other records > Added NeoCSVReaderTests>>#testSkippingEmptyRecords > === > Name: Neo-CSV-Tests-SvenVanCaekenberghe.18 > Author: SvenVanCaekenberghe > Time: 6 October 2014, 5:49:36.206743 pm > UUID: 95d87278-4398-41d3-96ed-83705d5488ce > Ancestors: Neo-CSV-Tests-SvenVanCaekenberghe.17 > > Added NeoCSVReader>>#addIgnoredFields: > Added NeoCSVReader>>#select:[thenDo:] to skip empty or other records > Added NeoCSVReaderTests>>#testSkippingEmptyRecords > === > > Here is the test: > > testSkippingEmptyRecords > | input output | > input := '1,2,3\\4,5,6\,,\7,8,9' withCRs. > output := (NeoCSVReader on: input readStream) > select: [ :each | each notEmpty and: [ (each allSatisfy: #isNil) not ] ]. > self assert: output equals: #(#('1' '2' '3') #('4' '5' '6') #('7' '8' '9')). > output := (NeoCSVReader on: input readStream) > emptyFieldValue: ''; > select: [ :each | each notEmpty and: [ (each allSatisfy: #isEmpty) not ] ]. > self assert: output equals: #(#('1' '2' '3') #('4' '5' '6') #('7' '8' '9')) > > Regards, > > Sven > > On 05 Oct 2014, at 02:39, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: > >> 2014-10-04 10:47 GMT-03:00 Sven Van Caekenberghe <sven(a)stfx.eu>: >>> Esteban, >>> >>> On 04 Oct 2014, at 02:42, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: >>> >>>> Sven, >>>> >>>> Is it possible to skip blank/empty lines in NeoCSVReader? >>>> >>>> So if a line is empty or contains ",,,,,,,," it would be skipped from >>>> the iteration. >>>> >>>> Does it already provider such feature? (I couldn't find any setting >>>> related with this). >>>> >>>> Esteban A. Maringolo >>> >>> Could you elaborate a bit on the use case ? >> >> Yes, users upload a CSV file with the proper columns (sometimes they >> don't, but that's another story), and because they export/copy the CSV >> from Excel most of >> the times there are *lots* (might be hundreds) of empty lines at the >> end, it is, lines >> with separators only, or even blank lines. >> >>> My first reaction would be that you have to deal with these yourself afterwards. >> >> When dealing with "array based" records, it is, with no #recordClass >> defined it would be easier. Because you can check whether all the >> values of the array are empty or nil. So on each iteration (#next) you >> could return the value array if it isn't empty or send #next again >> until you get a record with any data. >> >> When working with a record class I think you can't do this, because it >> seems you instantiate the record before iterating through each field >> of the record. >> >>> What would be the definition of an empty record ? What about empty quoted fields ? >> >> An empty record of a four columns CSV file would look like: >> ,,, >> or (if some columns are quoted) >> ,"","", >> >> I don't see how a record like that could be of any use. >> >> Best regards! >> >> Esteban A. Maringolo >> >> >>> PS: I got your other contribution, I think it is OK to add, thanks. >> >> Great! > >

Oct. 6, 2014