Squeak-dev June 2007

squeak-dev@lists.squeakfoundation.org

120 participants
175 discussions

Re: Re: [Q] Maximum Size of Squeak Image in 32Bit Linux
by "S.J.Chun" 27 Jun '07

27 Jun '07

No, there's no such requirement. I can use file system instead of image. For multiple image case, I think that's possible up to 10 or 20 images. But number of users are 100 to 200. So I think single image per user is not possible solution. ----- Original Message ----- From: Craig Latta <craig(a)netjam.org> To: squeak-dev(a)lists.squeakfoundation.org Sent: 07-06-27 13:24:01 Subject: Re: Re: [Q] Maximum Size of Squeak Image in 32Bit Linux Hi-- > The problem is that how large can I make the size of image? Is there an actual requirement that there be a single image, and not a communicating set of them? -C -- Craig Latta improvisational musical informaticist www.netjam.org Smalltalkers do: [:it | All with: Class, (And love: it)]

1 0

[Q] Maximum Size of Squeak Image in 32Bit Linux
by "S.J.Chun" 27 Jun '07

27 Jun '07

Hi, I'm currently developing Seaside application and the application accepts file upload from user. Given that smalltalk treats everything as objects (not files), I do want to store user's data to the image. The problem is that how large can I make the size of image? User's normal, average size of data will be in the range of 100 MB to 64 GB( single file). Thanks in advance.

3 2

[Monticello2] MDBogusA a sharedPool isNil
by Damien Cassou 26 Jun '07

26 Jun '07

Hi Avi and Colin, after loading Monticello2, a class has strange shared pools: MDBogusA sharedPools an OrderedCollection(MDBogusPoolA aDictionary(...) nil) So, there is 3 different kinds of element in the shared pools. Is this normal? At least the nil causes problems. Bye -- Damien Cassou

1 0

Some thoughts about native multithreading in smalltalk
by sig 26 Jun '07

26 Jun '07

Once again i asked myself, is it possible to run single squeak image using multiple native threads without a need of excessive locking anything or deeply rewriting a smalltalk code base.. and came to following ideas (please, don't blame me if this was already discussed and proven to be unusable... i hope this is original idea, right from my mind :) All my ideas below is around simple isolation strategy: only single native thread can write to object's memory at some point of time. Basically, this means that all primitive operations on object, which somehow altering its memory, like #basicAt:put:, or assigning ivar must be synchronized using single thread. There may be another issues, preventing from using multiple threads, but this one is most fundamental i think. Lets review the situation: suppose we having 1000 native threads performing implementation cycle, and now one of these threads came to point where it needs to access object's memory. Common approach would be to setup a write barrier for given memory region, where object's state stored (using mutex/semaphore or critical section) and any thread which willing to access there must take ownership on that particular barrier. Things seem pretty straightforward, but thinking more in detail, you'll see that this leads to nightmare - it will require to rewrite a code base, putting locks here and there.. and in result you'll have a common situation when 1 thread running while others waiting. Another way is based on message passing and what we have here: - any number of threads can run simultaneously under condition that there's no two threads which currently interpreting a message with same receiver. This condition is enough for making object's memory consistent. Since all methods can directly access only receiver's memory. In detail this means, that each thread having an unique receiver, and each object having a slot, which pointing to thread, which is currently interpreting a message with given object as receiver. To check, if we can send message to object without risk of getting into race condition we need to check object's thread id, and if it's nil, then we can safely send a message to it, by assigning object's thread id slot to current thread id, and entering method. Now, what will happen, when other thread needs to send message to object, which is already used by other thread as receiver? Things get complex when we using a stack for message passing, where we need to push/pop arguments e.t.c, using stack frames and all other things dependent on that. But what if we use a message queues instead of stack? The state of message evaluation can be represented by following structure (method context): - receiver - arguments - method - thread id (responsible for evaluating a message) - completion flag (indicating that message evaluation is complete) - message result - 'next queue object' (indicating that this message context are waiting for complete evaluation of other message) - bytecode position - space for temps Now, interpretation cycle will look like: A particular thread entering a method by receiving structure above and starting implementing method bytecodes. Each time, it needs to send message to some object, it must create a new structure, fill it with proper values and append to message queue of thread, which will be responsible for evaluation of sending message. After appending a message to queue it must wait for it's completion isn't? No! Yes, it stops interpretation of current method, but instead of waiting for complete evaluation of message required to continue evaluating current method, thread checking its message queue to see, if there's more messages awaiting for evaluation (having 'next queue object' = nil), or if it is not nil, having complete flag set. When finished evaluating a message, thread stores the message result and sets completion flag to true. Then starting listening queue as always. When thread chooses to continue evaluation of method context which was waiting for result of other context, it reads a result, then removes completed method context from queue. Now, since each object having a slot, indicating where it currently used as receiver we can simply decide to which thread's queue we must add a method context for evaluation of message send to prevent race condition. If thread A evaluating some message and needs to send message to object B, which is currently involved in message evaluation as receiver in thread C, we simply appending new method context to thread C queue and waiting for it's completion. A message queuing described above can be combined with technique, known as 'future refs' - we can continue evaluating a current message by not waiting message send result, but using a temporary 'future object' which is actually our method context described above. We can continue up to the point, where result of an evaluation needs to be stored in other object's memory or used in branching, any additional message sends to 'future object' can be simply queued in same thread which currently evaluating the future object. Ok, seems I described a rough view of idea. Awaiting for your comments! Also, consider that English is not my native language :) Regards, sig

1 0

Areithfa Ffenestri where?
by Herbert König 26 Jun '07

26 Jun '07

Hello, I wanted to have a look at Areithfa Ffenestri and couldn't find a link for download, only a lot of references on the mailing list. Any help? Thanks, Herbert mailto:herbertkoenig@gmx.net

6 9

Re: Re: [Q] Problem in Soap Processing
by "S.J.Chun" 26 Jun '07

26 Jun '07

Thank you very much. ----- Original Message ----- From: Masashi UMEZAWA <masashi.umezawa(a)gmail.com> To: "\"S.J.Chun\"" <chunsj(a)embian.com>, The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org> Sent: 07-06-26 14:57:00 Subject: Re: Re: [Q] Problem in Soap Processing Hi, > It seems that main cause of error is in Array>>elementsForwardIdentityTo: otherArray, which > is primitive and this primitive failure is main cause of error. This message is sent to array of single > element(a SoapHref) with #(0) as a argument. > The error seems to be related to SoapHref>>safeBecomeForward:. I thought it was fixed in the 20070508 version of SOAP on SM. Anyway the latest version (not released on SM yet) is here: http://squeaksource.blueplane.jp/SoapCore.html I could not reproduce the error when I read the envelope like: envelopeBuilder := SoapEnvelopeBuilder new. testXML := (FileStream readOnlyFileNamed: 'yourEnvelope.txt') contentsOfEntireFile. returnEnvelope := envelopeBuilder buildSoapEnvelopeFromXmlString: testXML. message := SoapMessage envelope: returnEnvelope. resp := SoapResponse fromSoapMessage: message. value := resp returnValue. "inspect it" So, could you send me SqueakDebug.log? (or DebugReport html is best, if possible). > Thanks in advance. > > PS) > Href's in Soap Request/Response is normal? I haven't seen this before. Can I turn this off? > It is normal and .NET SOAP implementations use Hrefs quite often, so I had to add this feature. However, for primitive types, the latest SoapCore does not use Href. So your problem might be solved by updating SOAP. Cheers, -- [:masashi | ^umezawa]

1 0

[Q] Problem in Soap Processing
by "S.J.Chun" 26 Jun '07

26 Jun '07

Hi, I've create both server/client in SoapCore in Squeak. The problem is that when I send request to receive array of dictionary the client, the client shows error which seems related to SoapHref. The response from server is <env:Envelope xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:env="http://schemas.xmlsoap.org/soap/envelope/" xmlns:enc="http://schemas.xmlsoap.org/soap/encoding/" env:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"> <env:Body> <m:directoryContentsResponse xmlns:m="http://www.nxiss.com/service"> <result xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:enc="http://schemas.xmlsoap.org/soap/encoding/" enc:arrayType="xsd:anyType[14]"> <item> <Name xsi:type="xsd:string">bgnet_A4.pdf</Name> <CreationTime xsi:type="xsd:positiveInteger">3360225158</CreationTime> <IsDirectory id="4035" xsi:type="xsd:boolean">false</IsDirectory> <FileSize xsi:type="xsd:integer">414313</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3360225155</ModificationTime> </item> <item> <Name xsi:type="xsd:string">Main.js</Name> <CreationTime xsi:type="xsd:positiveInteger">3360158456</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">48579</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3360157683</ModificationTime> </item> <item> <Name xsi:type="xsd:string">Seaside2.8a1-pmm.382.mcz</Name> <CreationTime xsi:type="xsd:positiveInteger">3360309391</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">453934</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3360309391</ModificationTime> </item> <item> <Name xsi:type="xsd:string">BmdEmail-bmd.1.mcz</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144764</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">14562</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358431079</ModificationTime> </item> <item> <Name xsi:type="xsd:string">FileMan-mu.98.mcz</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144922</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">32655</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358410615</ModificationTime> </item> <item> <Name xsi:type="xsd:string">jUploadr</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144212</CreationTime> <IsDirectory id="217" xsi:type="xsd:boolean">true</IsDirectory> <FileSize id="2491" xsi:type="xsd:integer">0</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358777691</ModificationTime> </item> <item> <Name xsi:type="xsd:string">SoapCore20070508</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144212</CreationTime> <IsDirectory href="#217"></IsDirectory> <FileSize href="#2491"></FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358419762</ModificationTime> </item> <item> <Name xsi:type="xsd:string">jets3t-0.5.1-20070425.zip</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144922</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">9471054</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358539537</ModificationTime> </item> <item> <Name xsi:type="xsd:string">s3-raw-http-example.zip</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144922</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">1535</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358539573</ModificationTime> </item> <item> <Name xsi:type="xsd:string">squirrelmail-1.4.10a</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144212</CreationTime> <IsDirectory href="#217"></IsDirectory> <FileSize href="#2491"></FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3356271056</ModificationTime> </item> <item> <Name xsi:type="xsd:string">capri_Naver_Dictionary_Widget_v1.9</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144212</CreationTime> <IsDirectory href="#217"></IsDirectory> <FileSize href="#2491"></FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358444358</ModificationTime> </item> <item> <Name xsi:type="xsd:string">Seaside2.8a1-lr.381.mcz</Name> <CreationTime xsi:type="xsd:positiveInteger">3360217379</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">453339</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3360217379</ModificationTime> </item> <item> <Name xsi:type="xsd:string">AWSECommerceService.wsdl</Name> <CreationTime xsi:type="xsd:positiveInteger">3360144922</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">146741</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3358675341</ModificationTime> </item> <item> <Name xsi:type="xsd:string">Scriptaculous-pmm.202.mcz</Name> <CreationTime xsi:type="xsd:positiveInteger">3360217385</CreationTime> <IsDirectory href="#4035"></IsDirectory> <FileSize xsi:type="xsd:integer">310641</FileSize> <ModificationTime xsi:type="xsd:positiveInteger">3360217385</ModificationTime> </item> </result> </m:directoryContentsResponse> </env:Body> </env:Envelope> It seems that main cause of error is in Array>>elementsForwardIdentityTo: otherArray, which is primitive and this primitive failure is main cause of error. This message is sent to array of single element(a SoapHref) with #(0) as a argument. Thanks in advance. PS) Href's in Soap Request/Response is normal? I haven't seen this before. Can I turn this off?

2 2

ESUG Event - Hotel special prices dead-line
by Noury Bouraqadi 25 Jun '07

25 Jun '07

Dear Smalltalkers, Hurry up if you didn't book your room for the ESUG conference. Noury Début du message réexpédié : > De : Michele Lanza <michele.lanza(a)unisi.ch> > Date : 25 juin 2007 15:44:40 HAEC > À : Noury Bouraqadi <bouraqadi(a)ensm-douai.fr>, Stéphane Ducasse > <stephane.ducasse(a)univ-savoie.fr> > Cc : ESUG Board <esug-info(a)esug.org> > Objet : hotel deadline > > Folks, > the deadline for the hotels is arriving, i.e., in a couple of days > we won't be able to guarantee the room prices anymore, and then > everybody is "on his own" on this..remember that in Lugano August > is high season. Stef, Noury, could we send out an e-mail to the > previous esug attendees saying that they need to hurry up. The > other option is that I basically block rooms, for example in Hotel > Zurigo, and then we allocate people in there, with the risk of > remaining with some empty rooms.. > > let me know, the hotels are stressing on this..! > > Michele > Noury ------------------------------------------------------------------ Dr. Noury Bouraqadi - Enseignant/Chercheur Responsable de l'option I.S.I.C. ARMINES - Ecole des Mines de Douai - Dept. I.A. http://csl.ensm-douai.fr/noury European Smalltalk Users Group Board http://www.esug.org ------------------------------------------------------------------

1 0

Traits for a Thesis?
by claudio acciaresi 25 Jun '07

25 Jun '07

Hello, we (Nicolas Butarelli and Claudio Acciaresi) are from Buenos Aires, Argentina, we’re trying to make a thesis at the University of Buenos Aires. Our initial idea is to reimplement a model using traits, we would like to know what your opinions are (if you think that this is useful, etc) about this idea and which models do you think that would be nice to rethink using traits. Any material related to this idea will be really useful. Any opinion on this idea will also be useful, if you know anther idea related to traits, please tell us. Thanks a lot. Nicolas and Claudio. __________________________________________________ Preguntá. Respondé. Descubrí. Todo lo que querías saber, y lo que ni imaginabas, está en Yahoo! Respuestas (Beta). ¡Probalo ya! http://www.yahoo.com.ar/respuestas

6 9

os-x 10.4.10 and SmalltalkImage current osVersion
by John M McIntosh 25 Jun '07

25 Jun '07

Oddly enough I was reading a few months back that the call we make in the os-x carbon VM to get the version number would lie when the minor version number went over the number 9. Sure enough running os-x 10.4.10 and asking for SmalltalkImage current osVersion -> 1049 The same value as when running 10.4.9 As far as I know any use of this version number in squeak on the mac usually cares about osVersion < 1000 or >= 1000 However I thought I'd note this event in case someone becomes confused about what the VM is reporting, versus what they know about the version of the OS they are running. -- ======================================================================== === John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
18
Older →

Jump to page: