- Pharo-dev - lists.pharo.org

Commenting the Session Manager

by Guillermo Polito

Hi all, I took some minutes to write down a class comment for the SessionManager class. There was an issue asking for it: https://pharo.fogbugz.com/f/cases/19463/Improve-SessionManager-class-comment Since it is an important topic, and sometimes too low level for some people, I'd like to have some feedback. Is it well explained? Is there something that is key for you and is missing? I know that the comment is not exhaustive, it can be iterated and enhanced, but we can have a nice first version of it for the release. Thanks, Guille =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= I am the object responsible of managing how sessions work in Pharo. A session defines the boundaries of work in the image. A new session starts when the image starts or when the image is saved. A session ends when the image quits or it is saved. There is only one active session at a single point of time. Saving the image causes the active to stop, and starts a new session. The current active session is hold by myself, the singleton session manager. It can be accessed by doing: SessionManager default currentSession. The most important responsibility of the session manager is to manage how resources and services in the image are started up and shut down at the beginning and end of a session respectively. For example, when the image starts, several initialization routines should be executed to make sure that the image has access to the graphic drivers, the standard input/output file descriptors and so on. Such initialization happens in the #snapshot:andQuit: method. #snapshot:andQuit: will: - stop current session - save current image if requested - quit if requested - start a new session When a session is started, all elements registered in the startup list are started up. When a session is stopped, all elements registered in the shutdown list are shut down. # Managing Startup and Shutdown lists The startup and shutdown lists can be accessed through the messages: SessionManager default startupList. SessionManager default shutdownList. In general terms, the shutdown list is the startup list reversed. Upon a startup [shutdown], all elements in the startup list are sent the message #startup: [#shutdown:] with a boolean as argument that indicates wether the image is being saved [closed]. Internally, startup and shutdown lists are prioritised. Priorities are managed by startup categories. By default the session manager includes the following categories in decreasing priority order: - System - Network - Graphical User Interface - Tools - User Categories can be accessed as follows: SessionManager default categoryNamed: aName. New categories can be registered in the system using the messages: SessionManager default createCategory: aCategoryName. SessionManager default createCategory: aCategoryName after: anotherCategoryName. Finally, to subscribe some resource handler to the startup shutdown lists, we need to subscribe a handler, subclass of AbstractSessionHandler. The most common handler implementation so far is the ClassSessionHandler, that allows to subscribe a class for startup and shutdown, keeping backwards compatibility to the old startup mechanism. ClassSessionHandler forClassNamed: aClassName We can register a session handler as follows SessionManager default register: (ClassSessionHandler forClassNamed: self name) inCategory: SessionManager default systemCategory. Or alternatively, by talking to the corresponding category: SessionManager default systemCategory register: (ClassSessionHandler forClassNamed: self name) # System Category Priorities A system category internally prioritizes its elements to provide a fine grained control on the startup and shutdown order. All methods above have variants that allow developers to specify the priority inside the category: SessionManager default register: (ClassSessionHandler forClassNamed: self name) inCategory: SessionManager default systemCategory atPriority: 100. SessionManager default systemCategory register: (ClassSessionHandler forClassNamed: self name) atPriority: 100 By default, if no priority is specified, a default priority is used. Every category answers to the message #defaultPriority. # How does an image restart from the point it was before An important point in the image startup is how does it manage to restart from the point where it was executing when it was saved. When the image is saved, using the snapshot primitive, the entire image is freezed at the point of the snapshot. More particularly, the process that invoked the snapshot primitive is freezed at the point of the primitive call. This works as a process fork: the running image will return from the snapshot primitive and the saved file will also start from the freezed point. To differentiate whether we are executing in the running image or in the freshly-saved image, the snapshot primitive returns a boolean indicating it. Read the comment of #snapshotPrimitive for more details.

April 13, 2017

CFP - IWST 2017 - International Workshop on Smalltalk Technologies

by jannik.laval＠gmail.com

CFP - IWST 2017 - International Workshop on Smalltalk Technologies [Please accept our apologies if you receive multiple copies of this call] [Please send to interested colleagues / mailing-lists] **************************************************************************** ******************************** CALL FOR PAPERS IWST 2017 Â International Workshop on Smalltalk Technologies http://www.esug.org/wiki/pier/Conferences/2017/International-Workshop-IWST_1 7 Maribor, Slovenia; Between September 4th to 8th, 2017 **************************************************************************** ******************************** ------------------- Goals and scopes ------------------- The goals of the workshop is to create a forum around advances or experience in Smalltalk and to trigger discussions and exchanges of ideas. The topics of your paper can be on all aspect of Smalltalk, theoretical as well as practical. Participants are invited to submit research articles or industrial papers. This year we want to open two different tracks: one research track and one industrial track with less scientific constraints. We expect papers of three kinds: Short position papers describing emerging ideas Long research papers with deeper description of experiments and of research results. Industrial papers with presentation of real and innovative Smalltalk applications; this kind of paper should enlighten why Smalltalk is really appropriate for your application. We will not enforce any length restriction. -------------------- Important Dates -------------------- Submission deadline: June 16th, 2017 Notification deadline: July 21th, 2017 Workshop : between September 4th and 8th, 2017 All accepted papers will be published in ACM DL (To be confirmed) ------------------- Topics ------------------- We welcome contributions on all aspects, theoretical as well as practical, of Smalltalk related topics such as: -Aspect-oriented programming, -Design patterns, -Experience reports, -Frameworks, -Implementation, new dialects or languages implemented in Smalltalk, -Interaction with other languages, -Meta-programming and Meta-modeling, -Tools ------------------- Best Paper Award ------------------- To encourage the submission of high-quality papers, the IWST organizing committee is very proud to announce a Best Paper Award for this edition of IWST. We thank the Lam Research Corporation for its financial contribution which makes it possible for prizes for the three best papers: 1000 USD for first place, 600 USD for second place and 400 USD for third place. The ranking will be decided by the program committee during the review process. The awards will be given during the ESUG conference social event. The Best Paper Award will take place only with a minimum of six submissions. Notice also that to be illegible, a paper must be presented at the workshop by one of the author and that the presenting author must be registered at the ESUG conference. ------------------- Publication ------------------- Both submissions and final papers must be prepared using the ACM SIGPLAN 10 point format. Templates for Word and LaTeX are available at http://www.acm.org/sigs/sigplan/authorInformation.htm. This site also contains links to useful informations on how to write effective submissions. ------------------- Submission ------------------- All submissions must be sent via easychair: https://easychair.org/conferences/?conf=iwst2017 ------------------- Program chairs ------------------- Anne Etien (UniversitÃ© de Lille 1, France) Jannik Laval (UniversitÃ© Lyon 2 LumiÃ¨re, France) -- ~~Jannik Laval~~ Enseignant-chercheur Responsable PÃ©dagogique Licence Coordonnateur de Projet en SystÃ¨me d'Information <http://iut.univ-lyon2.fr/> IUT LumiÃ¨re, <http://www.univ-lyon2.fr/> UniversitÃ© LumiÃ¨re Lyon 2 <http://disp-lab.fr/> laboratoire DISP +33 4 78 77 43 06 <http://www.jannik-laval.eu/> http://www.jannik-laval.eu <http://www.phratch.com/> http://www.phratch.com <http://www.approchealpes.info/> http://www.approchealpes.info --- L'absence de virus dans ce courrier Ã©lectronique a Ã©tÃ© vÃ©rifiÃ©e par le logiciel antivirus Avast. https://www.avast.com/antivirus

April 13, 2017

Re: [Pharo-dev] Security in the image

by Denis Kudriashov

2017-04-12 18:32 GMT+02:00 Ben Coman <btc(a)openinworld.com>: > If you want hostile actors working directly within the Image with a full > environment, then Pharo is probably not suitable. Its easy to get hold of > global class from the Playground references and overwrite/compile any > method in the system like this... > > SomeClass compile: 'initialize > MyEvilHack dostuff. > ^ super initialize ' > > You might want to consider Newspeak, which runs on the same VM as Pharo > and has a focus on security. > Interesting how they address your example?

April 13, 2017

Re: [Pharo-dev] Please, test our 64bits builds

by Alistair Grant

Hi Thierry, On Wed, Apr 12, 2017 at 07:33:21PM +0200, Thierry Goubier wrote: > Hi Alistair, > > for that I switch to OSProcess when on Pharo 64 bits. > > Regards, > > Thierry Thanks, that's working. Cheers, Alistair

April 13, 2017

Re: [Pharo-dev] Security in the image

by LaeMing

Ben Coman wrote > On Wed, Apr 12, 2017 at 3:34 PM, LaeMing < > laeming@.com > > wrote: > >> Hi, I am fishing about for an environment to try to develop a simple >> virtual >> world system in (yes, I'm aware of croquet/cobolt!). >> >> In a smalltalk-like environment, that might conceivably involve inviting >> multiple people into the 'image' where the world is hosted to interact >> with >> objects within it (if I was to use a smalltalk, it would want it to >> interleave the entire system, not just host it on top.) >> >> What sort of security implications might that have and are there any >> current >> solutions to a multi-user single-image situation? >> > > If you want hostile actors working directly within the Image with a full > environment, then Pharo is probably not suitable. Its easy to get hold of > global class from the Playground references and overwrite/compile any > method in the system like this... > > SomeClass compile: 'initialize > MyEvilHack dostuff. > ^ super initialize ' > > You might want to consider Newspeak, which runs on the same VM as Pharo > and > has a focus on security. > * https://www.slideshare.net/esug/8-gilad-brachaesug08 > * https://www.youtube.com/watch?v=UwkROn7OmNQ > > cheers -ben Thanks, Ben. I am looking into Newspeak now. I am rather impressed so far! -- View this message in context: http://forum.world.st/Security-in-the-image-tp4941857p4941932.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

April 13, 2017

Re: [Pharo-dev] Security in the image

by LaeMing

Stephan Eggermont wrote > On 12/04/17 09:34, LaeMing wrote: >> What sort of security implications might that have and are there any >> current >> solutions to a multi-user single-image situation? > > Take a look at gemstone. > https://www.youtube.com/user/JamesGFoster/videos > Starting at https://www.youtube.com/watch?v=U0z5TddqyQI > James made a really good series of videos > > > Stephan Thanks, Stephan. Gem looks rather good, but I need an OSS licensed environment. -- View this message in context: http://forum.world.st/Security-in-the-image-tp4941857p4941931.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

April 13, 2017

Re: [Pharo-dev] Alien osVersion is not defined. Pharo 5

by nacho

Thanks Esteban I will look at that. ----- Nacho Smalltalker apprentice. Buenos Aires, Argentina. -- View this message in context: http://forum.world.st/Alien-osVersion-is-not-defined-Pharo-5-tp4941906p4941… Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

April 12, 2017

Re: [Pharo-dev] Please, test our 64bits builds

by Thierry Goubier

Hi Alistair, for that I switch to OSProcess when on Pharo 64 bits. Regards, Thierry Le 12/04/2017 Ã 19:08, Alistair Grant a Ã©crit : > Hi Esteban, > > On Wed, Apr 12, 2017 at 06:54:16PM +0200, Esteban Lorenzano wrote: >> >>> On 12 Apr 2017, at 18:43, Alistair Grant <akgrant0710(a)gmail.com> >>> wrote: >>> >>> Hi Esteban, >>> >>> On Wed, Apr 12, 2017 at 06:08:56PM +0200, Esteban Lorenzano wrote: >>>> forget to say that iceberg will not work 100% yet :) but here, >>>> problem is because you need to install the threaded version, which >>>> of course I forget to point :P >>>> >>>> curl get.pharo.org/64/60+vmTLatest | bash >>>> >>>> this is the correct one for linux users. >>> >>> Thanks, I'm successfully past the Metacello load. >>> >>> Next problem: >>> >>> PrimitiveFailed: primitive #integerAt:put:size:signed: in >>> ExternalAddress failed >> >> the OSSubprocess maintainer needs to adapt his sources, I guess??? we >> cannot provide automatic compatibility??? :) > > OK, thanks. Hopefully Mariano is listening. ;-) > > Cheers, > Alistair > > >

April 12, 2017

Re: [Pharo-dev] Please, test our 64bits builds

by Alistair Grant

Hi Esteban, On Wed, Apr 12, 2017 at 06:54:16PM +0200, Esteban Lorenzano wrote: > > > On 12 Apr 2017, at 18:43, Alistair Grant <akgrant0710(a)gmail.com> > > wrote: > > > > Hi Esteban, > > > > On Wed, Apr 12, 2017 at 06:08:56PM +0200, Esteban Lorenzano wrote: > >> forget to say that iceberg will not work 100% yet :) but here, > >> problem is because you need to install the threaded version, which > >> of course I forget to point :P > >> > >> curl get.pharo.org/64/60+vmTLatest | bash > >> > >> this is the correct one for linux users. > > > > Thanks, I'm successfully past the Metacello load. > > > > Next problem: > > > > PrimitiveFailed: primitive #integerAt:put:size:signed: in > > ExternalAddress failed > > the OSSubprocess maintainer needs to adapt his sources, I guess??? we > cannot provide automatic compatibility??? :) OK, thanks. Hopefully Mariano is listening. ;-) Cheers, Alistair

April 12, 2017

Re: [Pharo-dev] Please, test our 64bits builds

by Esteban Lorenzano

> On 12 Apr 2017, at 18:43, Alistair Grant <akgrant0710(a)gmail.com> wrote: > > Hi Esteban, > > On Wed, Apr 12, 2017 at 06:08:56PM +0200, Esteban Lorenzano wrote: >> forget to say that iceberg will not work 100% yet :) >> but here, problem is because you need to install the threaded version, which of course I forget to point :P >> >> curl get.pharo.org/64/60+vmTLatest | bash >> >> this is the correct one for linux users. > > Thanks, I'm successfully past the Metacello load. > > Next problem: > > PrimitiveFailed: primitive #integerAt:put:size:signed: in ExternalAddress failed the OSSubprocess maintainer needs to adapt his sources, I guessâ¦ we cannot provide automatic compatibilityâ¦ :) Esteban > > while executing: > > proc := OSSUnixSubprocess new > command: 'git'; > arguments: #('config' '--global' 'user.name' 'AlistairGrant'); > runAndWaitOnExitDo: [ :cmd :out :err | self assert: cmd isSuccess ]. > > I'm going to have to learn about FFI some time... > > Full stack below. > > I should have also added I'm on Ubuntu 16.04, and this obviously all > works fine in the 32bit VM / image. > > Cheers, > Alistair > > > > ExternalAddress(Object)>>primitiveFailed: > ExternalAddress(Object)>>primitiveFailed > ExternalAddress(ByteArray)>>integerAt:put:size:signed: > ExternalAddress(ByteArray)>>unsignedLongAt:put: > ExternalAddress(ByteArray)>>nbUInt32AtOffset:put: > [ :string :index | > | pointer | > "0 terminated" > pointer := ExternalAddress allocate: string size + 1. > self registerPointer: pointer. > LibC memCopy: string to: pointer size: string size. > pointer nbUInt8AtOffset: string size put: 0. "string terminating null char" > aPointer > nbUInt32AtOffset: (index - 1) * self systemAccessor sizeOfPointer > put: pointer value ] in OSSUnixSubprocess>>collectArgumentPointersInto: in Block: [ :string :index | ... > OrderedCollection(SequenceableCollection)>>withIndexDo: > OSSUnixSubprocess>>collectArgumentPointersInto: > OSSUnixSubprocess>>buildPrimitiveArgVArgument > OSSUnixSubprocess>>internalRun > [ self internalRun ] in OSSUnixSubprocess>>run in Block: [ self internalRun ] > BlockClosure>>ensure: > OSSUnixSubprocess>>run > OSSUnixSubprocess>>runAndWaitOnExitDo: > UndefinedObject>>DoIt > OpalCompiler>>evaluate > DoItChunk>>importFor: > [ :declaration | > requestor ifNotNil: [ requestor contents: declaration contents ]. > value := declaration importFor: requestor ] in CodeImporter>>evaluateDeclarations in Block: [ :declaration | ... > OrderedCollection>>do: > CodeImporter>>evaluateDeclarations > CodeImporter class>>evaluateFileStream: > MultiByteFileStream(FileStream)>>fileIn > FileStream class>>fileIn: > SimpleServiceEntry>>performServiceFor: > PluggableButtonMorph>>performAction: > [ :m | > (m containsPoint: evt cursorPoint) > ifTrue: [ m enabled > ifTrue: [ m performAction: evt ] ] ] in PluggableButtonMorph>>mouseUp: in Block: [ :m | ... > Array(SequenceableCollection)>>do: > PluggableButtonMorph>>mouseUp: > PluggableButtonMorph(Morph)>>handleMouseUp: > MouseButtonEvent>>sentTo: >

April 12, 2017