February 2019 - Pharo-users

Re: [Pharo-users] About Iceberg

by Tim Mackinnon

Yes I agree - when there is so much discussion and debate going on, its easy to lose sight of the hard work and determination that went into getting us to this brave new world. I too want to shout a big thank you for the tooling and also the support that goes along with that. I love been able to think a bit more polyglot, and use tools/languages more easily side by side - although of course I want to hold on to what makes Smalltalk special (which is tons of stuff). I particularly love being able to feel like its easier to contribute - updating readmeâs and docs is trivial in a web browser now - just correct them and submit a PR. And on the receiving side - GitHub makes it easy to discuss the fix, alter it, or simply approve it. Equally - modern build tools easily detect the change, pull the code and rebuild and package it (and cheap scalable infrastructure). Its also getting easier and easier to submit code fixes too - and the VCS skills you learn doing this are transferable beyond Smalltalk - so its a big win win. So yes guys - thanks for hanging in for us! Tim > On 19 Feb 2019, at 08:50, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > Hi, > > This is a thank you note about Iceberg. > > I have been moving all my external and internal Pharo code to git/tonel/7 and on multiple occasions I have been pleasantly surprised about the functionality and performance of Iceberg. Basically, it just works. > > Finally, Pharo code lives in standard open source and commercial repositories (git, GitHub, Bitbucket, ...), without losing anything. > > I know that it took years to get here and that lots of code and community battles had to be fought. So thank you, to the whole team, you did a great job ! > > Sven > > -- > Sven Van Caekenberghe > Proudly supporting Pharo > http://pharo.org > http://association.pharo.org > http://consortium.pharo.org > > > > >

Feb. 19, 2019

Re: [Pharo-users] Pier3 & content management

by Albrecht Baur

thanks for the info - I will give it a try next weeks. On 19.02.19 12:41, Esteban Lorenzano wrote: > Hello, > > Not many people are using it, but it should be not much problem make > it run in Pharo 7 (and with current version of Seaside). > > Esteban > >> On 19 Feb 2019, at 12:03, Albrecht Baur via Pharo-users >> <pharo-users(a)lists.pharo.org <mailto:pharo-users@lists.pharo.org>> wrote: >> >> >> *From: *Albrecht Baur <baur.albrecht(a)googlemail.com >> <mailto:baur.albrecht@googlemail.com>> >> *Subject: **Pier3 & content management* >> *Date: *19 February 2019 at 12:03:27 CET >> *To: *Any question about pharo is welcome >> <pharo-users(a)lists.pharo.org <mailto:pharo-users@lists.pharo.org>> >> >> >> Hey, >> >> what is the status of Pier3 + Pier3AddOns ? >> >> -> Seems to me that it is discontinued since Pharo 5, right ? >> >> I'm asking because I look around to see if there is a project usable >> as a starting point of a small web content management module for >> seaside (nothing too complicated / fancy). >> >> Is there something besides Pier that goes into this direction ? >> >> Thanks for hints, >> Albrecht >> >> >

Feb. 19, 2019

Re: [Pharo-users] How do you avoid loading master code which is indirectly referenced by a version in the baseline?

by phil＠highoctane.be

Same boat here. Phil On Tue, Feb 19, 2019, 13:05 Tim Mackinnon <tim(a)testit.works wrote: > Hi Sabine - you raise an important point, and I am interested in us > getting better answers to this too. Hopefully Dale seeâs this and is > thinking about this in Rowen. Repeatable loading is an important enterprise > feature. > > Tim > > > On 18 Feb 2019, at 15:46, Sabine Manaa <manaa.sabine(a)gmail.com> wrote: > > > > For test and deployment, I have to be sure that I know exactly which > versions > > I load and I have to be sure to be able to load exactly the same code > again. > > I do not speak from my projects but from the projects I use. > > > > I found myself asking people again and again to set tags in their > projects > > and being surprised again and again by changed code even though I am > trying > > to load versions (with versions I mean NOT #stable or #master but like > > v2.2.2). > > > > The reason for this problem I then have, is that often in baselines with > > version numbers (e.g. v2.2.2) other projects are referenced but as > master. > > So in consequence, I load version but get master from another project > which > > is referenced indirectly. > > > > For me, that this is not good at all. I was wondering that this seems > not to > > be a problem for others. > > > > I found out that I can put a lock in my baselines preLoadDoIt for those > > indirectly referenced projects. > > > > Examlpe Artefact v1.0.1 > > > > > https://github.com/pharo-contributions/Artefact/blob/v1.0.1/src/BaselineOfA… > > > > loads github://zweidenker/Units/src (which is master). > > > > I found a solution for this: > > > > In the preLoadDoIt of my baseline I lock Units to a certain commit. > > > > preLoadDoIt > > Metacello new > > baseline: 'Units'; > > repository: 'github://zweidenker/Units:98d5a3d/src'; > > lock. > > > > This solves my problem in this example. > > > > Disadvantage is that I have to analyze everything which is loaded from > the > > projects I use and create a lock for it. > > > > Questions: > > 1.) Wouldn't it make sense to have some mechanism, perhaps in Rowan then, > > where I can say that a project which has a number can only reference > other > > projects with numbers? This would avoid the whole problem > > 2.) How do others solve this problem? > > > > addendum: I do not speak about project versions which are in development, > > here #stable or master are great features. But sometime when a project > has a > > certain state, and it gets a version, then it should not change..imho > > > > > > > > -- > > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > > > > > >

Feb. 19, 2019

Re: [Pharo-users] How do you avoid loading master code which is indirectly referenced by a version in the baseline?

by Tim Mackinnon

Hi Sabine - you raise an important point, and I am interested in us getting better answers to this too. Hopefully Dale seeâs this and is thinking about this in Rowen. Repeatable loading is an important enterprise feature. Tim > On 18 Feb 2019, at 15:46, Sabine Manaa <manaa.sabine(a)gmail.com> wrote: > > For test and deployment, I have to be sure that I know exactly which versions > I load and I have to be sure to be able to load exactly the same code again. > I do not speak from my projects but from the projects I use. > > I found myself asking people again and again to set tags in their projects > and being surprised again and again by changed code even though I am trying > to load versions (with versions I mean NOT #stable or #master but like > v2.2.2). > > The reason for this problem I then have, is that often in baselines with > version numbers (e.g. v2.2.2) other projects are referenced but as master. > So in consequence, I load version but get master from another project which > is referenced indirectly. > > For me, that this is not good at all. I was wondering that this seems not to > be a problem for others. > > I found out that I can put a lock in my baselines preLoadDoIt for those > indirectly referenced projects. > > Examlpe Artefact v1.0.1 > > https://github.com/pharo-contributions/Artefact/blob/v1.0.1/src/BaselineOfA… > > loads github://zweidenker/Units/src (which is master). > > I found a solution for this: > > In the preLoadDoIt of my baseline I lock Units to a certain commit. > > preLoadDoIt > Metacello new > baseline: 'Units'; > repository: 'github://zweidenker/Units:98d5a3d/src'; > lock. > > This solves my problem in this example. > > Disadvantage is that I have to analyze everything which is loaded from the > projects I use and create a lock for it. > > Questions: > 1.) Wouldn't it make sense to have some mechanism, perhaps in Rowan then, > where I can say that a project which has a number can only reference other > projects with numbers? This would avoid the whole problem > 2.) How do others solve this problem? > > addendum: I do not speak about project versions which are in development, > here #stable or master are great features. But sometime when a project has a > certain state, and it gets a version, then it should not change..imho > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html >

Feb. 19, 2019

Re: [Pharo-users] OSProcess / OSSubprocess / LibC uniqueInstance

by Serge Stinckwich

Thank you Thierry. On Tue, Feb 19, 2019 at 12:48 PM Thierry Goubier <thierry.goubier(a)gmail.com> wrote: > Hi Serge, Albrecht, > > loading OSProcess directly with the following expression works: > > Metacello new > configuration: 'OSProcess'; > repository: 'http://www.squeaksource.com/MetacelloRepository'; > load > > There is still some work to be done on OSProcess adaptation to Pharo > 7, but they are in rarely used parts. > > Regards, > > Thierry > > Le mar. 19 fÃ©vr. 2019 Ã 12:19, Albrecht Baur via Pharo-users > <pharo-users(a)lists.pharo.org> a Ã©crit : > > > > I need to execute OS commands from within a x64 pharo 7.01 image on > centos7 (for example: 'zip -r 123.zip zipme/'). > > > > I found these 3 tools: OSProcess, OSSubprocess, LibC uniqueInstance > > > > But I did not find any docu on which one to use for which use case. > > > > OSProcess is in the catalog browser but fails loading because of > ByteArray>>ifNotNilDo: > > -> So OSProcess is not to be used with p7 ? > > > > OSSubprocess loads via catalog browser and the following snippet works > (even though 64 bit): > > > > OSSUnixSubprocess new > > command: 'zip'; > > arguments: #('-r' '123.zip' 'zipme'); > > redirectStdout; > > runAndWaitOnExitDo: [ :process :outString | > > outString inspect > > ] > > > > So is it ok to use OSSubprocess with 64 bit even though documented as 32 > bit only ? (here: > https://github.com/pharo-contributions/OSSubprocess#installation) > > > > LibC uniqueInstance system: 'zip -r 123.zip zipme/' > > ... works, but I don't know if its the right tool for this. > > > > Is there somewhere more info on which tool to use for which use case ? > > > > Thanks in advance! > > > > Albrecht > > > > PS.: I need to use zip as os process only because of this issue: > https://github.com/pharo-project/pharo/issues/2584 > > -- Serge Stinckwic âhâ Int. Research Unit on Modelling/Simulation of Complex Systems (UMMISCO) âSorbonne University (SU) French National Research Institute for Sustainable Development (IRD)â U âniversity of YaoundÃ© Iâ, Cameroun "Programs must be written for people to read, and only incidentally for machines to execute." https://twitter.com/SergeStinckwich â

Feb. 19, 2019

Re: [Pharo-users] How to load a single package in a Metacello baseline

by Tim Mackinnon

Yeah - I have a similar issue with ExternalWebBrowser, which I load as: spec baseline: 'ExternalWebBrowser' with: [ spec repository: 'github://seandenigris/Pharo-Web-Browser:master/repository' ]; import: 'ExternalWebBrowser'. That drags in BabyMock and huge raft of tests (which Iâm grateful were written), but Iâd prefer not to muddy the waters when presenting Exercism. I think what we have to do is submit pull requests to these projects such that they create better baselines with Groups. Eg. The Exercism baseline does this so that students load core, but developers get extra tools to generate exercises - eg. spec package: 'Exercism' with: [ ]; package: 'ExercismDev' with: [ spec requires: #('Exercism' 'ExercismTools')]; package: 'ExercismTools' with: [ spec requires: #('OSProcess' 'ExternalWebBrowser') ]; group: 'default' with: #('ExercismTools'); group: 'dev' with: #('ExercismDevâ)] Then you can specify in your load request which group to use (I donât have the syntax handy for baselines, but when you load via iceberg, the Metacelle menu has a second option where you specify the group to use, so Iâm sure it can be done in code too). Iâd be interested if there is some way to do it without groups though, so that you donât have to rely on the package authors to do this (similar in thought to what Sabine is complaining about by not being able to better control version dependencies). This whole area is becoming more and more important as we grow up and what to have more controllable load characteristics. Tim > On 19 Feb 2019, at 11:13, N. Bouraqadi <bouraqadi(a)gmail.com> wrote: > > Hi, > > I'm updating the the baseline of PharoJS to use the zinc websockets from the github repo https://github.com/svenvc/zinc <https://github.com/svenvc/zinc>. I want to load package Zinc-WebSocket-Core. So, in BaselineOfPharoJS>>#baseline: I've defined > spec baseline: 'ZincHTTPComponents' with: [ > spec > loads: #('Zinc-WebSocket-Core'); > repository: 'github://svenvc/zinc <github://svenvc/zinc>' ]. > > However, for some reason I get Zinc-WebSocket-Tests loaded. Any hint how to avoid it?(edited) > Looking at BaselineOfZincHTTPComponents>>#baseline:, package WebSocket-Core does not seem to depend on tests > spec package: 'Zinc-WebSocket-Core' with: [ spec requires: 'Zinc-HTTP' ] > > Thanks, > Noury

Feb. 19, 2019

Re: [Pharo-users] OSProcess / OSSubprocess / LibC uniqueInstance

by Thierry Goubier

Hi Serge, Albrecht, loading OSProcess directly with the following expression works: Metacello new configuration: 'OSProcess'; repository: 'http://www.squeaksource.com/MetacelloRepository'; load There is still some work to be done on OSProcess adaptation to Pharo 7, but they are in rarely used parts. Regards, Thierry Le mar. 19 fÃ©vr. 2019 Ã 12:19, Albrecht Baur via Pharo-users <pharo-users(a)lists.pharo.org> a Ã©crit : > > I need to execute OS commands from within a x64 pharo 7.01 image on centos7 (for example: 'zip -r 123.zip zipme/'). > > I found these 3 tools: OSProcess, OSSubprocess, LibC uniqueInstance > > But I did not find any docu on which one to use for which use case. > > OSProcess is in the catalog browser but fails loading because of ByteArray>>ifNotNilDo: > -> So OSProcess is not to be used with p7 ? > > OSSubprocess loads via catalog browser and the following snippet works (even though 64 bit): > > OSSUnixSubprocess new > command: 'zip'; > arguments: #('-r' '123.zip' 'zipme'); > redirectStdout; > runAndWaitOnExitDo: [ :process :outString | > outString inspect > ] > > So is it ok to use OSSubprocess with 64 bit even though documented as 32 bit only ? (here: https://github.com/pharo-contributions/OSSubprocess#installation) > > LibC uniqueInstance system: 'zip -r 123.zip zipme/' > ... works, but I don't know if its the right tool for this. > > Is there somewhere more info on which tool to use for which use case ? > > Thanks in advance! > > Albrecht > > PS.: I need to use zip as os process only because of this issue: https://github.com/pharo-project/pharo/issues/2584

Feb. 19, 2019

Re: [Pharo-users] Pier3 & content management

by Esteban Lorenzano

Hello, Not many people are using it, but it should be not much problem make it run in Pharo 7 (and with current version of Seaside). Esteban > On 19 Feb 2019, at 12:03, Albrecht Baur via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > > From: Albrecht Baur <baur.albrecht(a)googlemail.com> > Subject: Pier3 & content management > Date: 19 February 2019 at 12:03:27 CET > To: Any question about pharo is welcome <pharo-users(a)lists.pharo.org> > > > Hey, > > what is the status of Pier3 + Pier3AddOns ? > > -> Seems to me that it is discontinued since Pharo 5, right ? > > I'm asking because I look around to see if there is a project usable as a starting point of a small web content management module for seaside (nothing too complicated / fancy). > > Is there something besides Pier that goes into this direction ? > > Thanks for hints, > Albrecht > >

Feb. 19, 2019

How to load a single package in a Metacello baseline

by N. Bouraqadi

Hi, I'm updating the the baseline of PharoJS to use the zinc websockets from the github repo https://github.com/svenvc/zinc. I want to load package Zinc-WebSocket-Core. So, in BaselineOfPharoJS>>#baseline: I've defined spec baseline: 'ZincHTTPComponents' with: [ spec loads: #('Zinc-WebSocket-Core'); repository: 'github://svenvc/zinc' ]. However, for some reason I get Zinc-WebSocket-Tests loaded. Any hint how to avoid it?(edited) Looking at BaselineOfZincHTTPComponents>>#baseline:, package WebSocket-Core does not seem to depend on tests spec package: 'Zinc-WebSocket-Core' with: [ spec requires: 'Zinc-HTTP' ] Thanks, Noury

Feb. 19, 2019

Re: [Pharo-users] OSProcess / OSSubprocess / LibC uniqueInstance

by Albrecht Baur

I just tried the easy way via catalog browser. I think it is OSProcess 4.6.9 from http://www.squeaksource.com/OSProcess On 19.02.19 12:01, Serge Stinckwich wrote: > > > On Tue, Feb 19, 2019 at 11:56 AM David T. Lewis <lewis(a)mail.msen.com > <mailto:lewis@mail.msen.com>> wrote: > > On Tue, Feb 19, 2019 at 11:37:38AM +0100, Albrecht Baur via > Pharo-users wrote: > > From: Albrecht Baur <baur.albrecht(a)googlemail.com > <mailto:baur.albrecht@googlemail.com>> > > Subject: OSProcess / OSSubprocess / LibC uniqueInstance > > To: pharo-users(a)lists.pharo.org <mailto:pharo-users@lists.pharo.org> > > Date: Tue, 19 Feb 2019 11:37:38 +0100 > > > > I need to execute OS commands from within a x64 pharo 7.01 image on > > centos7 (for example: 'zip -r 123.zip zipme/'). > > > > I found these 3 tools: OSProcess, OSSubprocess, LibC uniqueInstance > > > > But I did not find any docu on which one to use for which use case. > > > >Â * OSProcess is in the catalog browser but fails loading because of > >Â Â ByteArray>>ifNotNilDo: > >Â Â -> So OSProcess is not to be used with p7 ? > > Hi Albrecht, > > I am sorry that I cannot check it right now, but it may be that > ConfigurationOfOSProcess needs to be updated. The #ifNotNilDo: problem > was fixed by Thierry Goubier in July 2017, so perhaps you are using > an out of date version. > > Hopefully someone else can give you a better answer, otherwise I will > try to follow up in a day or two. > > > I have the same problem of ifNotNilDo: when I load OSProcess in Pharo 7. > What version do you use ? > Regards, > > -- > Serge Stinckwic > h > > Int. Research Unit > Â on Modelling/Simulation of Complex Systems (UMMISCO) > Sorbonne University > Â (SU) > French National Research Institute for Sustainable Development (IRD) > U > niversity of YaoundÃ© I, Cameroun > "Programs must be written for people to read, and only incidentally > for machines to execute." > https://twitter.com/SergeStinckwich >

Feb. 19, 2019