Squeak-dev May 2009

squeak-dev@lists.squeakfoundation.org

102 participants
131 discussions

About best ways preserving backward compatibility
by Igor Stasenko 06 May '09

06 May '09

Hello again, I implemented a new delays, called AdvancedDelay (such a bold name :)), and now during installation of new Processor need to switch Delay -> AdvancedDelay. Now, the problem is, that we having a Delay's sitting in image, and to make transition w/o problems i need to make all Delays migrate to a new instances. Also, i'd like to being able to switch back to old class, in case if image running under older VM, which does not supports new primitives. There are a couple of alternatives, rename a Delay class, to something like OldDelay, and use a Delay global var as a placeholder for currently installed class, so it can switch between OldDelay and AdvancedDelay at image startup. But newest Delays having a subclasses (DelayWaitTimeout, MonitorDelay) which makes this a bit more problematic. Another option which i thought, is to make a Delay class abstract, and have a subclasses with concrete implementations, while Delay class>>someMessage implement as: ActiveDelayClass someMessage "ActiveDelayClass is class var" Third option would be forget about backward compatibility and overwrite Delay with new Delay class. So, i want to ask people: what you think , what way is preferable? Do i need to care about backward compatibility (in order to be able to run newer images on older VMs), or not? And if do, then what the best way? VM changes is backward compatible (you can run old images with new VM w/o problems). -- Best regards, Igor Stasenko AKA sig.

1 0

Interaction between objects which can be waited on, and their clients(waiters)
by Igor Stasenko 05 May '09

05 May '09

Hello list, I stared @ DelayWaitTimeout class, and thinking, how to implement it in new scheduler.. First i thought , just make it quickly and forget about it, but one thing which don't makes me feel good about it, that DelayWaitTimeout is a perfect example, when you need to stop a Process until one of the multiple events occurs - in case of DelayWaitTimeout its either semaphore signaled or delay expired. So, i thinking, what if instead of implementing a DelayWaitTimeout , which works only for 2 objects - delay & semaphore, implement a more generic class , which would allow us to wait for any number of objects and blocks process until one of them is signaled. Or, blocks process until all of them signaled. There's nothing new, in fact, in Windoze there an API function, called WaitForMultipleObjects, which does exactly what i describing - it could wait on multiple objects, releasing a waiter once a single object signaled, or release waiter only after all objects signaled. Here is my thoughts, how to implement that. First, we need to allow Delay & Semaphore to work with abstract waiter object (not just Process). A proposed protocol between object which acts as one who can be waited on, and object who acts as waiter is following: To start waiting, we sending a #waitForSignalBy: waiter semaphore waitForSignalBy: waiter or delay waitForSignalBy: waiter then, default #wait could be implemented, in both cases, simply as: ^ self waitForSignalBy: Processor activeProcess. now, after issuing #waitForSignalBy: we have the following possible cases: - given object is already signaled (as a semaphore with excess signals), in this case we immediately sending #waitComplete: to waiter - given object is scheduled for being signaled later , in this case we sending #startWaitingOn: to waiter. - now after #startWaitingOn: issued and once a semaphore/delay signaled we sending #waitComplete: to waiter. - in case if semaphore/delay abandoned, we sending #waitAbandoned: to waiter. This happens, for instance, when we terminating a process who currently waits on semaphore. (If you have a better idea how to name these messages, please make a guess) Now, by having such protocol implemented for Delay, Semaphore & Process, we now can easily introduce another class which could allow us to wait for multiple events. Again, i'm not sure how properly call it.. Just want to illustrate the usage: semaphore := Semaphore new. delay := Delay forMilliseconds: 100. signaler := (MultipleWait on: { semaphore. delay }) waitForOne. "wait until one of the listed objects get signaled" or: result := (MultipleWait on: { semaphore. delay }) waitForAll. "wait until all of the listed objects get signaled" by default, MultipleWait>>waitForSignalBy: is releasing waiter once one of the listed objects signaled (as in waitForOne case). Not sure about return result of #waitForAll, IMO , it could be just a boolean value - return true, if all objects signaled, or false in case if waiting abandoned. Now, of course, since MultipleWait acts as a waiter and as an object who can be waited on, we can allow nesting: mwait1 := MultipleWait on: { sema1. sema2 }. (MultipleWait on: { mwait1. delay }) waitForOne. and finally, as you may guess , a Semaphore>>waitTimeoutMSecs: now can be implemented as: waitTimeoutMSecs: anInteger "Wait on this semaphore for up to the given number of milliseconds, then timeout. Return true if the deadline expired, false otherwise." | delay | delay := Delay forMilliseconds: (anInteger max: 0). ^ delay == (MultipleWait on: { self. delay }) waitForOne. -- Best regards, Igor Stasenko AKA sig.

2 7

Re: Pier-EmailSender and "Contact Us"
by John M McIntosh 05 May '09

05 May '09

I note if you go to the squeak.org website and tap but report it dumps you into Mantis. However for novices that likely isn't helpful and the link to a mail message from Feb isn't something anyone is going to find, or mmm or remember. On 5-May-09, at 9:05 AM, Andreas Raab wrote: > I'll check it out. Also note that cc-ing release(a)lists.squeakfoundation.org > will generally be useful. > > Cheers, > - Andreas > > John M McIntosh wrote: >> Well then for http://bugs.squeak.org/view.php?id=6581 >> who would stick that into 3.11? Let alone that I'm looking at >> 3.10.2 image >> Added (partly) >> http://bugs.squeak.org/view.php?id=7348 >> On 5-May-09, at 2:07 AM, Andreas Raab wrote: >>> Hi John - >>> >>> John M McIntosh wrote: >>>> (I had to fix the sq 3.10.2 image because of ignored bug using >>>> allInstances I reported years back, never fixed, but fixed in >>>> pharo last fall). >>> >>> [... snip ...] >>> >>>> (to be fair here I could bring some fixes into the 3.10.2 image >>>> to reduce the cpu time, but really those fixes should have gone >>>> in years ago...) >>> >>> Just as an FYI, there is now a well-defined process for >>> contributing stuff to 3.11: >>> >>> http://lists.squeakfoundation.org/pipermail/squeak-dev/2009-February/134095… >>> >>> If these fixes are on Mantis all you need is to provide an >>> installer script and update its status accordingly. Just try it! >>> >>> Cheers, >>> - Andreas -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

2 1

Rosetta Code site
by Daniel Ingalls 04 May '09

04 May '09

Hi all - Today while looking for an old reference to Rosetta Smalltalk, I came across this site: http://rosettacode.org In their own words, "The idea is to present solutions to the same task in as many different languages as possible, to demonstrate how languages are similar and different, and to aid a person with a grounding in one language in learning another. Rosetta Code currently has 278 tasks, and covers 135 languages, though we do not have solutions to every task in every language." It seemed to me that some of the Smalltalk examples are not as nice as they might be. For instance the very first thing I looked at was the Ackermann function which is written as a block... |ackermann| ackermann := [ :n :m | (n = 0) ifTrue: [ (m + 1) ] ifFalse: [ (m = 0) ifTrue: [ ackermann value: (n-1) value: 1 ] ifFalse: [ ackermann value: (n-1) value: ( ackermann value: n value: (m-1) ) ] ] ]. To begin with, this is cluttered with value:value:; it includes redundant parentheses which don't belong in an illustration of the language; finally, it will not run in Smalltalks that do not allow recursive blocks. Seems to me it might be more illustrative as an Integer method that will run in any Smalltalk, that runs faster too, and looks cleaner to me... <Integer> ack: n "The first parameter m is the receiver, self, in this Ackermann method" self = 0 ifTrue: [^ n+1]. self > 0 ifTrue: [n = 0 ifTrue: [^ self-1 ack: 1]. n > 0 ifTrue: [^ self-1 ack: (self ack: n-1)] ] " 0 ack: 0 ==> 1 3 ack: 4 ==> 125 " Anyway, I thought Smalltalk's representation on that site could use a little QA, and that folks on this list might have a good time playing TerseMan. I could imagine the Squeak solutions being an interesting area in our Wiki, both as a staging area for such a project, and as a possibly interesting collection of TerseMan discussions (terseness is great fun, but not always the ideal). A comment on the last line might point to our Wiki which might gain some new Squeak users, and would encourage checking with us before anyone overwrites our contributions.

3 3

Teams Review
by Ken Causey 04 May '09

04 May '09

http://squeakboard.wordpress.com/2009/05/04/teams-review/ "As part of our responsibility to the community to guide and nurture the community teams I sent out a simple survey to all current team leaders on April 8th. A response to this email is required and I set a deadline of May 9th. Today I sent out a reminder. If you are the leader of a Squeak community team and did not receive this email (subject of most recent is: “REMINDER: What is the Status of Your Team?”) please contact me as soon as possible. Expect a report based on this data in the coming weeks." Ken

1 0

[ANN] BobsBrowser again
by Hernán Morales Durand 04 May '09

04 May '09

Dear all, As result of a strange exercise in software resurrection, you have the chance of using the BobsBrowser [1] again , which is an exploration of different ways of browsing code. To install please evaluate: Installer ss project: 'BobsBrowser'; install: 'BobsBrowser'. You can start a new browser using the World menu or by evaluating: BobsBrowser browseAllClasses The creator of this goodie was Bob Arning, enjoy it. Best regards, Hernán [1] http://www.squeaksource.com/BobsBrowser.html

3 7

[ANN] Aida/Web 6.0 second beta is out!
by Janko Mivšek 04 May '09

04 May '09

Dear Smalltalkers, That time with a considerable more fresh meat and on both Squeak+friends and VisualWorks, a second beta of Aida/Web 6.0 web framework and application server is released. To download it visit our renewed website http://www.aidaweb.si With that beta we also changed the direction: now the 6.0 won't be just a companion of Scribo CMS but a standalone release with support of tree-like control flow, so that this release will really deserve a major number jump from 5.6 to 6.0. Here are just 3 most important new features in that beta: 1.Introduction of web execution contexts as a step towards tree-like control flow. These contexts give a scope to the "execution" of web request and are the precondition for controlling this flow of execution. 2.Translation support for text on the web apps, to allow the multilingual text on the pages, besides multilingual content, 3.Caching (memcached like) to improve performance of serving the content, which otherwise doesn't change frequently. But much more in release notes: http://www.aidaweb.si/release-notes-6.0.html Best regards Janko Mivšek -- Janko Mivšek AIDA/Web Smalltalk Web Application Server http://www.aidaweb.si

2 2

[ANN] Z39.50 FFI Client (ZOOM binding)
by Hernán Morales Durand 04 May '09

04 May '09

This is a binding of the ZOOM Yaz client for Squeak and Pharo [1]. To install: 1) Download the compiled library (DLL or SO) from http://www.indexdata.dk/yaz/ and copy the yaz3.dll (or equivalent in Unix) to your Squeak directory. 2) Make sure you have installed FFI package in your image: 3) If you have Installer in your image doIt this: Installer ss project: 'Z3950'; install: 'Z3950'; install: 'Z3950Tests'. To use: Right now you must know the infamous PQF/PQN query syntax (I'm working on it). Connect to a Z39.50 server and make a query for records with author Foucault : Z3950Client new server: #Alameda; query: '@attr 1=1003 foucault'. or using a random server for Kant Z3950Client new query: '@attr 1=1003 kant'. Best regards Hernán [1] SqueakSource page link : http://www.squeaksource.com/Z3950.html

1 0

[ANN] Dependency Browser update
by Hernán Morales Durand 04 May '09

04 May '09

Dear all, I've updated the Dependency Browser package and documentation, take a look here: http://cs.hernanmorales.com.ar/projects/dependencyBrowser/DBrowser-en.php Quick install if you have OmniBrowser and Installer packages loaded: | dwInstaller | dwInstaller := Installer ss project: 'DependencyWalker'. dwInstaller install: 'DependencyWalker'; install: 'OBDependencyWalker'; install: 'DWTraitsFix'. ( Smalltalk version beginsWith: 'Squeak' ) ifTrue: [ dwInstaller install: 'OBDWMethodNodeFix' ] Cheers. Hernán

1 0

FFI question [WAS] squeakopendbx, which VM?
by Mariano Martinez Peck 04 May '09

04 May '09

Hi folks! I have a problem with SqueakDBX and FFI. I install openDBX from a .deb and I get this: ldconfig -p | grep opendbx show the same as you: libopendbxplus.so.1 (libc6) => /usr/lib/libopendbxplus.so.1 libopendbx.so.1 (libc6) => /usr/lib/libopendbx.so.1 The problem is that we (in squeakDBX) have as module: opendbx. Linux looks for libMODULE_NAME.so. In this case, was searching libopendbx.so. But, my file was named libopendbx.so.1. And of course, I got a 'unable to find function address'. Solution: copy and rename (or just rename) libopendbx.so.1 to libopendbx.so and worked like a charm. But, the question is, shouldn't FFI resolve this ? is this a matter of the linker ? Thanks for the help! Mariano ---------- Forwarded message ---------- From: Mariano Martinez Peck <marianopeck(a)gmail.com> Date: Tue, Apr 28, 2009 at 10:40 PM Subject: Re: [Pharo-project] squeakopendbx, which VM? To: Pharo-project(a)lists.gforge.inria.fr I think I fix it :) I took a clean Ubuntu. I did: sudo dpkg -i libopendbx1_1.4.0-1_i386.deb libopendbx1-mysql_1.4.0-1_i386.deb Then ldconfig -p | grep opendbx show the same as you: libopendbxplus.so.1 (libc6) => /usr/lib/libopendbxplus.so.1 libopendbx.so.1 (libc6) => /usr/lib/libopendbx.so.1 The problem is that we have as module: opendbx. Ubuntu looks for libMODULE_NAME.so. In this case, was searching libopendbx.so. But, your (and my) file was named libopendbx.so.1. Solution: copy and rename (or just rename) libopendbx.so.1 to libopendbx.so and should work like a charm. Thanks for your testing Hilaire. We want SqueakDBX to be as much easier as possible. So, documenting the wiki with several cases will definitively help us. I add this information to the wiki. Let me know if that work. Cheers, Mariano On Tue, Apr 28, 2009 at 2:56 PM, Mariano Martinez Peck < marianopeck(a)gmail.com> wrote: > > > 2009/4/28 Hilaire Fernandes <hilaire(a)ofset.org> > >> >> >> 2009/4/28 Mariano Martinez Peck <marianopeck(a)gmail.com> >> >>> >>> >>>> I don't have it in my version of Pharo. I don't know if I can update it >>>> without going through a full pharo image which I don't want to do now. >>>> >>> >>> >>> Ok. Doesn't matter. Your current problem is not about FFI but you will >>> have that problem then (when you get squeakDBX working). You only need to >>> get the latests FFI version from MC to work in pharo with closure VM. >>> >> >> I am pretty sure the VM I used is not with the fixed closure. I am note >> sure about the image, I can tell you tomorrow the version number of the >> image when at the office. >> Is the fixed closure mandatory to get SqueakDBX working? >> Which Pharo # image and VM should be used? >> > > SqueakDBX works with both. If you use a closure image you must use a VM > with closure an latest FFI (from MC). But, you can also use Squeak in older > version (non closure) with FFI from Universes. So, no problem for you. > > > >> >> I haven't my linux machine here, I will look tonight in my home, but: >> >> > 1) what do you have /usr/lib in PATH ? >> No it points only to binary as it should. >> Does FFI looks at PATH for the libraries? If so it is a buggy behavior. >> I am really currious to know how FFI find the libraries... >> > > I am not sure about this. Perhaps someone (Andreas) can help us. > > >> >> >> 2) Just for testing, can you put the openDBX libraries in the same >> directory where the pharo >> image is. >> >> 3) The same as 2) but where the VM is. >> >> I will try tomorrow morning. >> > > Which Linux are you using? Perhaps there is a problem trying to resolve > the library. The last test you can do (just to see if it works) is this: > > Go t class OpenDBXUnix and change the methods > > apiBind: handle database: databaseName name: userName password: password > method: method > > and > > apiInitialize: handle backend: backend host: host port: port > > In both, put module: 'libopendbx.so.1' or module: 'libopendbx.so' instead > of module: 'opendbx' > > I hope you get it working. Ask all what you want. > > Greetings, > > Mariano > > >> >> >> Thanks! >> >> >> Hilaire >> >> >> -- >> http://blog.ofset.org/hilaire >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > >

3 10

Jump to page: