- Pharo-users - lists.pharo.org

Re: [Pharo-users] Collection with memory?

by Denis Kudriashov

2015-03-23 9:57 GMT+03:00 stepharo <stepharo(a)free.fr>: > Le 22/3/15 23:01, Sven Van Caekenberghe a Ã©crit : > >> On 22 Mar 2015, at 22:40, stepharo <stepharo(a)free.fr> wrote: >>> >>> >>> >>> Le 22/3/15 22:05, Sven Van Caekenberghe a Ã©crit : >>> >>>> On 22 Mar 2015, at 21:44, Denis Kudriashov <dionisiydk(a)gmail.com> >>>>> wrote: >>>>> >>>>> Hi >>>>> >>>>> It is another perfect task for XStreams: >>>>> >>>>> r := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') reading selecting: [:each >>>>> | each first = $a]. >>>>> r get > a12 >>>>> r get > a13 >>>>> >>>> Beautiful ! >>>> >>> Yes but with >>> r next. >>> r next >>> >> The choice for different selector names was intentional, by design. To >> avoid confusion, because #next and #get are not 100% identical >> (semantically). This is an important point. >> > No I asked martin personally. This is the real reason. There were also > experimenting to offer an API closer to other language. > So that we have > r get. > r next: 4 > But in original docs from https://code.google.com/p/xtreams there are only #get and #read: messages.

March 23, 2015

Re: [Pharo-users] Collection with memory?

by Sven Van Caekenberghe

> On 23 Mar 2015, at 07:57, stepharo <stepharo(a)free.fr> wrote: > > > > Le 22/3/15 23:01, Sven Van Caekenberghe a Ã©crit : >>> On 22 Mar 2015, at 22:40, stepharo <stepharo(a)free.fr> wrote: >>> >>> >>> >>> Le 22/3/15 22:05, Sven Van Caekenberghe a Ã©crit : >>>>> On 22 Mar 2015, at 21:44, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: >>>>> >>>>> Hi >>>>> >>>>> It is another perfect task for XStreams: >>>>> >>>>> r := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') reading selecting: [:each | each first = $a]. >>>>> r get > a12 >>>>> r get > a13 >>>> Beautiful ! >>> Yes but with >>> r next. >>> r next >> The choice for different selector names was intentional, by design. To avoid confusion, because #next and #get are not 100% identical (semantically). This is an important point. > No I asked martin personally. This is the real reason. There were also experimenting to offer an API closer to other language. > So that we have > r get. > r next: 4 IIRC, one difference between #get and #next is that get cannot return nil and will throw an Incomplete exception when EOF next can return nil and might throws an Exception Changing the meaning of an existing selector will cause many problems and discussions. People will think they do not have to change anything and they will be wrong. > superb inconsistencet >> >>> I was reading the XTreams API recently but this is forbidden for me to open another project before finishing >>> what I started. Now if somebody else would like to help pushing Xtreams in Pharo 50. >>> :) >> Well, it does/did load OK in Pharo 3 and probably will in 4 too. >> >> https://ci.inria.fr/pharo-contribution/job/Xtreams/ > Yes loading is just a part of the story :) > We should remove the old ones. > > And we should revise the API because > I **hate ++, -=, += Yeah, but is not because it looks a bit like C++ that it is no longer Smalltalk, these are still normal messages. But yes, some API discussions are possible. >> >>>> Thx Denis. >>>> >>>>> 2015-03-22 20:47 GMT+03:00 stepharo <stepharo(a)free.fr>: >>>>> Hi >>>>> >>>>> I need a collection that does the following: >>>>> >>>>> col := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') asMyNewCollection. >>>>> col findFirst: [:each | each first = $a] >>>>> > a12 >>>>> col findFirst: [:each | each first = $a] >>>>> > a13 >>>>> >>>>> Do you know a collection that would work? >>>>> So I have the impression that such behavoir could be defined with a kind of methods wrapper. >>>>> >>>>> Stef >> >> > >

March 23, 2015

Re: [Pharo-users] Collection with memory?

by stepharo

Le 22/3/15 23:10, Alain Rastoul a Ã©crit : > Le 22/03/2015 22:39, stepharo a Ã©crit : >> This is like the Iterator Design pattern and this is interesting. > Yes, usually in dotnet IQueryable is implemented by (or on top of) a > IEnumerable. It is one of the key elements of Linq that makes Linq > very user friendly for developers >> Why don't you give a try to see how it grows? > I did a small try . it works, and there was not too much errors in my > suggestion . sometimes I surprise myself ... ;-) > I think it could be cool not only for collections, but with Glorp for > example (translate predicates to sql, some kind of "Linq for > smalltalk"). I think I'll dig a bit in that direction > May be time for some explorations with Glorp. Excellent! Improve our lives :)

March 23, 2015

Re: [Pharo-users] Collection with memory?

by stepharo

Le 22/3/15 23:01, Sven Van Caekenberghe a Ã©crit : >> On 22 Mar 2015, at 22:40, stepharo <stepharo(a)free.fr> wrote: >> >> >> >> Le 22/3/15 22:05, Sven Van Caekenberghe a Ã©crit : >>>> On 22 Mar 2015, at 21:44, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: >>>> >>>> Hi >>>> >>>> It is another perfect task for XStreams: >>>> >>>> r := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') reading selecting: [:each | each first = $a]. >>>> r get > a12 >>>> r get > a13 >>> Beautiful ! >> Yes but with >> r next. >> r next > The choice for different selector names was intentional, by design. To avoid confusion, because #next and #get are not 100% identical (semantically). This is an important point. No I asked martin personally. This is the real reason. There were also experimenting to offer an API closer to other language. So that we have r get. r next: 4 superb inconsistencet > >> I was reading the XTreams API recently but this is forbidden for me to open another project before finishing >> what I started. Now if somebody else would like to help pushing Xtreams in Pharo 50. >> :) > Well, it does/did load OK in Pharo 3 and probably will in 4 too. > > https://ci.inria.fr/pharo-contribution/job/Xtreams/ Yes loading is just a part of the story :) We should remove the old ones. And we should revise the API because I **hate ++, -=, += > >>> Thx Denis. >>> >>>> 2015-03-22 20:47 GMT+03:00 stepharo <stepharo(a)free.fr>: >>>> Hi >>>> >>>> I need a collection that does the following: >>>> >>>> col := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') asMyNewCollection. >>>> col findFirst: [:each | each first = $a] >>>> > a12 >>>> col findFirst: [:each | each first = $a] >>>> > a13 >>>> >>>> Do you know a collection that would work? >>>> So I have the impression that such behavoir could be defined with a kind of methods wrapper. >>>> >>>> Stef > >

March 23, 2015

Re: [Pharo-users] A Pillar language package for Atom editor?

by HernÃ¡n Morales Durand

Hi Doru, This happens with the following files: https://ci.inria.fr/pharo-contribution/job/PharoForTheEnterprise/lastSucces… https://ci.inria.fr/pharo-contribution/job/PharoForTheEnterprise/lastSucces… https://ci.inria.fr/pharo-contribution/job/PharoForTheEnterprise/lastSucces… https://ci.inria.fr/pharo-contribution/job/PharoForTheEnterprise/lastSucces… https://ci.inria.fr/pharo-contribution/job/PharoForTheEnterprise/lastSucces… HernÃ¡n 2015-03-22 10:05 GMT-03:00 Tudor Girba <tudor(a)tudorgirba.com>: > Could you send me the file you are using? Given that the parser is just an > approximate one, it can be that your specific constructs are not supported. > > Cheers, > Doru > > On Fri, Mar 20, 2015 at 8:59 PM, HernÃ¡n Morales Durand < > hernan.morales(a)gmail.com> wrote: > >> Ok, alternatively, is someone generating Pillar chapters in Windows? >> >> My problem is GT-InspectorExtensions-Pillar highlights the syntax but not >> properly (see attachment). Could I be using a wrong syntax? >> >> I also figured out a command line to execute pdflatex passing a modified >> PFTE.tex, but latex errors are cryptic (and I don't want to follow there). >> >> HernÃ¡n >> >> >> >> >> 2015-03-17 17:57 GMT-03:00 HernÃ¡n Morales Durand < >> hernan.morales(a)gmail.com>: >> >> Is this working? >>> How to preview a .pier file? >>> >>> https://github.com/pillar-markup/language-pillar >>> >>> Cheers, >>> >>> HernÃ¡n >>> >>> >> > > > -- > www.tudorgirba.com > > "Every thing has its own flow" >

March 23, 2015

Re: [Pharo-users] How to depend on another Github repo from the baseline ?

by Dale Henrichs

Kilon, Thierry's solution is the preferred one ... I know that a lot of people are inclined to insert arbitrary Smalltalk code into specifications, but one should pretty much stick to the spcification-based solution and avoid including arbitrary Smalltalk ... including a load like you have may "work" but if any issues start showing up (and they are likely to eventually) I won't be able to help you ... If one uses git/github, it should be completely unnecessary to have a ConfigurationOf at all ... as the combination of a BaselineOf and git covers all of the bases ... Thierry is correct that a BselineOf can be used with a non-git repository, but if you take that approach you will always load the bleeding edge and there will be no way to put a stake in the ground with respect to a certain fixed version. Git gives you the SHA and tags to identify a specific version. Dale On 3/17/15 4:55 AM, Thierry Goubier wrote: > Hi Kilon, > > I think you can simply do: > > baseline: spec > <baseline> > spec > for: #pharo > do: [ > spec baseline: 'SmaCC' with: [ spec repository: > 'github://ThierryGoubier/SmaCC' ]. > spec package: 'Ephestos' with: [ spec requires: #('SmaCC') ] ] > > You can also restrict what you load from SmaCC in this way: > > baseline: spec > <baseline> > spec > for: #pharo > do: [ > spec baseline: 'SmaCC' with: [ spec repository: > 'github://ThierryGoubier/SmaCC' ]. > spec import: 'SmaCC'. > spec package: 'Ephestos' with: [ spec requires: #('SmaCC-Python') ] ] > > (i.e. by importing the baseline of SmaCC, you can choose among the > groups or the packages of SmaCC...) > > From what Dale told me, you can only import one baseline this way. But > you can add from more than one baseline with import: provides: > > baseline: spec > <baseline> > spec > for: #pharo > do: [ > spec baseline: 'SmaCC' with: [ spec repository: > 'github://ThierryGoubier/SmaCC' ]. > spec import: 'SmaCC' provides: #('SmaCC-Python'). > spec package: 'Ephestos' with: [ spec requires: #('SmaCC-Python') ] ] > > Dale, is that correct? I haven't tested those, so... > > Note that any kind of repository can host a baseline this way, > smalltalkhub, squeaksource, etc... > > version05: spec > <version: '0.5'> > spec > for: #'pharo4.x' > do: [ > spec > baseline: 'GitFileTree' > with: [ spec repository: > 'http://smalltalkhub.com/mc/ThierryGoubier/GitFileTreePharo40/main' ]; > import: 'GitFileTree' ] > > And of course a baseline can require a configuration... > > I'm still exploring all the possibilities this allows (and, more > important, how configurations may be simplified to defer effective > loading to a baseline inside the master repository, so that the > configuration very rarely has to be updated). > > Thierry > > 2015-03-17 12:11 GMT+01:00 kilon alios <kilon.alios(a)gmail.com > <mailto:kilon.alios@gmail.com>>: > > I have this very simple baseline for Ephestos > > baseline: spec > <baseline> > spec > for: #pharo > do: [ > spec package: 'Ephestos' with: [Metacello new > baseline: 'SmaCC'; > repository: 'github://ThierryGoubier/SmaCC'; > load] ]. > > The above code works fine for me. The user can install Ephestos > from configuration browser that has a configuration that loads > this baseline which in turn not only loads the latest code for the > project but also runs this block to make sure Smacc is installed > for Ephestos to use. > > The question however how else I can do this and what are the > advantages and disadvantages for those other approaches ? > >

March 22, 2015

Re: [Pharo-users] Git overriding repo load from different source

by Dale Henrichs

Peter, sorry for not answering sooner ... I keep an eye out for posts with Metacello in the subject line so I missed this one ... So if i understand your question, you have a GitHub project that you normally download like this: Metacello new baseline: 'Sample'; repository: 'github://GsDevKit/sampleProject:master/repository'; load. now you have cloned the repository to a local git repository: cd /opt/git git clone git@github.com:GsDevKit/sampleProject.git and you can directly load the project like this: Metacello new baseline: 'Sample'; repository: 'filetree:'///opt/git/sampleProject/repository'; load. but in another configuration (that you may not control) there is dependency upon Sample that looks like this: spec baseline: 'Sample' with: [ spec repository: 'github://GsDevKit/sampleProject:master/repository' ]. and you are wondering how you can force you local clone to be used ... is that correct? If so then the trick is to use the Metacello `lock` ... Do the following BEFORE loading the project that depends upon you sample project: Metacello new baseline: 'Sample'; repository: 'filetree:'///opt/git/sampleProject/repository'; lock. Then whenever Metacello finds a dependency upon 'Sample' the local clone (the `lock`ed) project will be used instead. This is use case is the primary reason for the existence of the `lock` command. Also note that you need to use the scripting api for all loads to make sure that the lock is honored. Finally you can use the onLock: message to break the lock in certain cases ... the default Metacello load honors all locks (a Warning is dumped to the transcript) ... As Thierry mentions, you should be using the latest version of Metacello on github ... If I missed the mark here let know ... Dale On 3/21/15 6:41 AM, Peter UhnÃ¡k wrote: > Hi, > > I'm trying to load local package over git one... > basically: > > Projects A and B. > A depends on B. > A loads B from its BaselineOfA, the reference points to github. > > Now I have also local copy of B git repository, and I would like to > either > a) Override the load of B with local B (so the packages will point to > local instead of remote). > Currently it fails on MetacelloRepositorySpec>>hasNoLoadConflicts: > because each (local and git) repo points to different place. > This would be useful if I want to make change to B AFTER the project B > is already loaded. > > b) Add all packages from B to local B repo (same as manually using > "add to package..." in Monticello Browser") > This will work only if both the repos have the same packages, so > probably not that useful. (i.e. no code differences) > > c) Override the GitHub B repository in the BaselineOfA BEFORE the > project is loaded. > This would require some mapping in Monticello browser, I've seen > something like "repositoryOverrides:" but I don't know how it works. > This would be useful if I knew beforehand that I want to make changes > also to B. > > So how would one go about this? Having option "c)" would be probably > the best; I can live without "a)". :) > > Any pointers appreciated, > Peter

March 22, 2015

Re: [Pharo-users] Collection with memory?

by Alain Rastoul

Le 22/03/2015 22:39, stepharo a Ã©crit : > This is like the Iterator Design pattern and this is interesting. Yes, usually in dotnet IQueryable is implemented by (or on top of) a IEnumerable. It is one of the key elements of Linq that makes Linq very user friendly for developers > Why don't you give a try to see how it grows? I did a small try . it works, and there was not too much errors in my suggestion . sometimes I surprise myself ... ;-) I think it could be cool not only for collections, but with Glorp for example (translate predicates to sql, some kind of "Linq for smalltalk"). I think I'll dig a bit in that direction May be time for some explorations with Glorp. -- Regards, Alain

March 22, 2015

Re: [Pharo-users] Collection with memory?

by Alain Rastoul

Le 22/03/2015 22:42, stepharo a Ã©crit : > https://code.google.com/p/xtreams/ Thank you, I just found it too by watching the Martin Kobetic video on youtube. Very interesting -- Regards, Alain

March 22, 2015

Re: [Pharo-users] Collection with memory?

by Sven Van Caekenberghe

> On 22 Mar 2015, at 22:40, stepharo <stepharo(a)free.fr> wrote: > > > > Le 22/3/15 22:05, Sven Van Caekenberghe a Ã©crit : >>> On 22 Mar 2015, at 21:44, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: >>> >>> Hi >>> >>> It is another perfect task for XStreams: >>> >>> r := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') reading selecting: [:each | each first = $a]. >>> r get > a12 >>> r get > a13 >> Beautiful ! > > Yes but with > r next. > r next The choice for different selector names was intentional, by design. To avoid confusion, because #next and #get are not 100% identical (semantically). This is an important point. > I was reading the XTreams API recently but this is forbidden for me to open another project before finishing > what I started. Now if somebody else would like to help pushing Xtreams in Pharo 50. > :) Well, it does/did load OK in Pharo 3 and probably will in 4 too. https://ci.inria.fr/pharo-contribution/job/Xtreams/ >> >> Thx Denis. >> >>> 2015-03-22 20:47 GMT+03:00 stepharo <stepharo(a)free.fr>: >>> Hi >>> >>> I need a collection that does the following: >>> >>> col := #('a12' 'b12' 'a13' 'a14' 'c23' 'a16') asMyNewCollection. >>> col findFirst: [:each | each first = $a] >>> > a12 >>> col findFirst: [:each | each first = $a] >>> > a13 >>> >>> Do you know a collection that would work? >>> So I have the impression that such behavoir could be defined with a kind of methods wrapper. >>> >>> Stef

March 22, 2015