Squeak-dev March 2006

squeak-dev@lists.squeakfoundation.org

177 participants
255 discussions

Re: hash and equal problem
by ncellier＠ifrance.com 28 Mar '06

28 Mar '06

Hi, Stef and Boris, We know from 1997-99 discussions that VA implementation is not good. It computes hash code fast, but handle Set of Float inefficiently. Example: ((1 to: 100) collect: [:i | 10.0 raisedTo: i negated]) asSet. All elements have same hash code (0 hash) and access performance turn to linear search. In other dialects, self=self truncated ifFalse: [...] is also here to handle fraction/float equality (3/2) hash = 1.5 hash VW used to convert everything to Fraction (and had nasty bugs because their imperfect asRational algorithm sometimes have (aDouble = aDouble asRational asDouble) answering false, see SYSBUG-asRational at cincom public store). Now they convert everything to Float, then to Double if anything fail (Overflow) then to old Fraction hash (Large Fraction can also overflow max IEEE754 Double). I think new implementation is faster, and also covers other nasty bugs of aFloat=aDouble answering true, but having different fractional representations. This also cover scaled numbers. I do not have Dolphin, ST/X, ... image here, but i imagine they follow one of the above paths. Sorry, I have no original solution in mind, but borrow solutions of our Smalltalk cousins. ________________________________________________________________________ iFRANCE, exprimez-vous ! http://web.ifrance.com

3 2

hash and equal problem
by nicolas cellier 28 Mar '06

28 Mar '06

Hi all, i do not think that this behavior is correct (2 = 2.0) = true. (2 hash = 2.0 hash) = false. i saw a discussion about it in an old old thread (1999), but did not understand the rationale for such a behavior. i can exhibit a lot of examples of code completely broken by this behavior. Try this one: | aSet | aSet := Set new. aSet add: #( 1.0 2 3 4 5 6). aSet add: #( 1 2.0 3 4 5 6). aSet add: #( 1 2 3.0 4 5 6). aSet add: #( 1 2 3 4.0 5 6). aSet add: #( 1 2 3 4 5.0 6). aSet add: #( 1 2 3 4 5 6.0). aSet size Then, try this slight modification: | aSet | aSet := Set new: 12. aSet add: #( 1.0 2 3 4 5 6). aSet add: #( 1 2.0 3 4 5 6). aSet add: #( 1 2 3.0 4 5 6). aSet add: #( 1 2 3 4.0 5 6). aSet add: #( 1 2 3 4 5.0 6). aSet add: #( 1 2 3 4 5 6.0). aSet size You will get 6 or 5 depending on initial Set capacity... Secondly, Array hash, Matrix hash, WideSstring hash are hashing each and every element... Sawing this, I imagine one never store 100x100 Matrices in a Dictionary, unless prepared to spend a lot of time at coffee machine... In other words, finding a good balance between long to run hash algorithm and too many identical hash values is difficult. I am not sure that the exhaustive way is a good balance. The bet that string will be short is maybe not the worst one (unless you store whole sentence like i saw in translation), but betting that every array in Smalltalk will be short is nonsense. Try this one if you are not convinced: (SequenceableCollection allSubInstances collect: [:e | e size]) asBag valuesAndCounts associations asSortedCollection: [:a :b | a key > b key] In a 3.9a image, i have an average of 74 and a maximum over the millon. So i raise the question again. Should we really let above behavior exist in squeak ? Shouldn't we look for more efficient hashing ? I'm asking for a good rationale (compatibility issues included).

3 3

REMINDER: Squeak Chat Wednesday March 29th 04:00 GMT
by Ken Causey 28 Mar '06

28 Mar '06

Just a reminder that the next Squeak Chat will take place between 04:00 (4 AM) and 08:00 (8 AM) GMT Wednesday March 29th, 2006. As usual we will meet in #squeak on irc.freenode.net. More information about the #squeak IRC channel can be found at http://people.squeakfoundation.org/article/7.html This is a good time to contribute to the development of Squeak by participating in the Harvesting Process. One of the main ways to contribute is by Reviewing submitted fixes and enhancements. For information about the Harvesting Process see http://minnow.cc.gatech.edu/squeak/3152 If you have the time please attend and work toward the future of Squeak. The next seven chats are scheduled for 08:00 (08:00 AM) GMT Sunday April 02, 2006 12:00 (12:00 PM) GMT Thursday April 06, 2006 16:00 (04:00 PM) GMT Monday April 10, 2006 20:00 (08:00 PM) GMT Friday April 14, 2006 00:00 (12:00 AM) GMT Wednesday April 19, 2006 04:00 (04:00 AM) GMT Sunday April 23, 2006 08:00 (08:00 AM) GMT Thursday April 27, 2006

1 0

Re: [Seaside] WeakArray (again)
by David Shaffer 28 Mar '06

28 Mar '06

William Harford wrote: > > The application uses call/render a lot. This was one of the main > draws to using Seaside. It allows us to easily customize the > application for our clients. We build our entire application out of > smaller reusable components that can be easily customize/replaced. > There are ~= 12 (and could go up to 30) components created for each > request. In most cases the user moves on. Would limiting the number > of pages a user can backtrack help the situation ? I'm cross-posting to Squeak-dev and I think this should migrate to that list (since it has come up there several times). Please follow up to Squeak-dev only. I don't believe this is a problem for VisualWorks which seems to have a significantly more robust weak reference mechanism. ...as for your question... I don't have a good feeling for the impact of that. It would limit the size of the WALRUCache (?) used to track continuations but I'm not sure that it would actually impact the lifetime of objects in the registry. Actually I don't have much intuition at all regarding lifetime of the objects in the registry. I think the idea was that when a continuation "expires" (gets pushed out of the LRU cache) then data in its "snapshot" should be available for GC. Keeping the cache small might limit the number of items in the weak dictionary if you GC often enough. Sorry I can't say more. Maybe Julian or Avi would care to chime in ;-) \begin{amateurHour} It seems to me that the notification needs to be changed to actually queueing information about the objects which the GC deams un(strongly)reachable. I spent some time staring at ObjectMemory>>sweepPhase, #finalizeReference: and #signalFinalization: which seem to be the cornerstones of this process. All that #signalFinalization: is currently doing is signaling a semaphore (well, indicating that one should be "signaled" later). Why not keep a list of (oop,i) [i is the offset of the weak reference in the oop] pairs and somehow communicate those back to a Smalltalk object? As a total VM novice it just seems too simple ;-) What I think I would do is associate a queue like thing with every weak reference container. Then when an object becomes GC-able I'd place the (oop,i) pair in that shared queue. What I need is someone to hold my hand through... ...designing this "queue like thing". How about a circular array which can only be "read" (move the read index) by ST code and only be written by the VM code? This avoids a lot of concurrency issues. Are there any examples like this in the VM? \end{amateurHour} Andreas said it wasn't trivial and I believe him but I think we've got to give it a try or risk having Squeak/Seaside being ignored for larger projects. It is also a big problem for anyone using GOODS or GLORP, for example, since those libraries make extensive use of weak references. So...is there anyone with a knowledge of the VM willing to step up, design the solution and divvy up parts of it for those of us with a good knowledge of C but little of the VM. David

8 22

mac VM universal 3.8.11b5U pending
by John M McIntosh 27 Mar '06

27 Mar '06

I've posted a Squeak 3.8.11beta5U.app Mac Carbon VM to my idisk and FTP server. http://www.smalltalkconsulting.com/squeak.html This version has a new FFI bundle for PPC and intel which fixes the crash issue when we attempt to start Croquet, it also has a start on building the browser plugin which isn't done yet. However you need this VM if you want to run Croquet. 3.8.11b5 sqMacOpenGL.c print diagnostics sqMacFFIPPC.c Migrate back to carbon FFI prims and Andreas' FFI interface because unix one is broken for Croquet. print diagnostics, use UI Locking code to call FFI functions to avoid crash with quicktime. sqMacNSPlugin.c sqMacNSPLuginUILogic.c sqMacWindowUniversal.c Lots of changes to reconsider when SetWindow() is called, and migrate to Context draw logic Add logic so that we can do OpenGL. Still broken -- ======================================================================== === John M. McIntosh <johnmci(a)smalltalkconsulting.com> 1-800-477-2659 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===

1 0

Trying to find the Kay discussion of invention of term "OO", cell.
by blake＠kingdomrpg.com 27 Mar '06

27 Mar '06

Hey, gang, I'm trying to find that e-mail that's been posted a couple of times where Alan discusses the origin of OO and cells. I think I've been using the version here: userpage.fu-berlin.de/~ram/pub/pub_jf47htdHt/doc_kay_oop_en But this seems to be down. I need to contrast the original ideas with the Inheritance/Encapsulation/Polymorphism that became OO as it is currently practiced. ===Blake=== "Spoon!"

3 4

64bit unix vm, built with WIDTH=32, seg faults when it tries to do something net-ish
by Stephen Davies 27 Mar '06

27 Mar '06

Hi, Can anyone help a newbie with the UNIX VM. I have a Gentoo 64bit system. I pulled down Squeak-3.8a-2.src.tar.gz, the 64bit sources. I compiled it with "make WIDTH=32" (I want to run an ordinary 32bit image). After installing, squeak does start OK. However, if I try to do anything that does netish stuff - for example "update code from server", or try to open SqueakMap Package Loader, then my vm crashes with a Segmentation Fault, something like so: Segmentation fault 19565880 [] in NetNameResolver class>addressForName:timeout: 19565972 [] in Semaphore>critical: 19565604 BlockContext>ensure: 19565512 Semaphore>critical: 19564480 NetNameResolver class>addressForName:timeout: 19564228 HTTPSocket class>httpGetDocument:args:accept:request: 19562540 HTTPSocket class>httpGetDocument: 19562448 HTTPClient class>httpGetDocument: 19562356 HTTPClient class>httpGet: 19563796 [] in Utilities class>newUpdatesOn:special:throughNumber: 19561652 SequenceableCollection>do: 19563480 Utilities class>newUpdatesOn:special:throughNumber: 19562104 [] in Utilities class>readServer:special:updatesThrough:saveLocally:updateImage: 19561468 BlockContext>ensure: 19561376 Cursor>showWhile: 19561852 Utilities class>readServer:special:updatesThrough:saveLocally:updateImage: 19514288 Utilities class>readServerUpdatesThrough:saveLocally:updateImage: 19515092 Utilities class>readServerUpdatesSaveLocally:updateImage: 19515000 Utilities class>updateFromServer 19514816 TheWorldMenu>doMenuItem:with: 19514724 [] in MenuItemMorph>invokeWithEvent: 19514632 BlockContext>ensure: 19514540 Cursor>showWhile: 19514196 MenuItemMorph>invokeWithEvent: 19514100 MenuItemMorph>mouseUp: 19514008 MenuItemMorph>handleMouseUp: 19513776 MouseButtonEvent>sentTo: 19513684 Morph>handleEvent: 19513592 MorphicEventDispatcher>dispatchDefault:with: 19513500 MorphicEventDispatcher>dispatchEvent:with: 19513408 Morph>processEvent:using: 19513316 MorphicEventDispatcher>dispatchDefault:with: 19513208 MorphicEventDispatcher>dispatchEvent:with: 19513076 Morph>processEvent:using: 19512984 Morph>processEvent: 19512892 MenuMorph>handleFocusEvent: 19512616 [] in HandMorph>sendFocusEvent:to:clear: 19512708 [] in PasteUpMorph>becomeActiveDuring: 19512524 BlockContext>on:do: 19512432 PasteUpMorph>becomeActiveDuring: 19512248 HandMorph>sendFocusEvent:to:clear: 19512156 HandMorph>sendEvent:focus:clear: 19512012 HandMorph>sendMouseEvent: 19511920 HandMorph>handleEvent: 19511664 HandMorph>processEvents 19511756 [] in WorldState>doOneCycleNowFor: 19511572 SequenceableCollection>do: 19511480 WorldState>handsDo: 19511388 WorldState>doOneCycleNowFor: 19511296 WorldState>doOneCycleFor: 19511204 PasteUpMorph>doOneCycle 17625536 [] in Project class>spawnNewProcess 17625720 [] in BlockContext>newProcess This was the same with a 3.9 alpha image, and a seaside-specific image. I really feel at a loss for how to find the problem. gdb won't work on this box due to the use of PaX. Any pointers would be appreciated. Thanks, Steve

1 0

Mantis Issue: 0003355 - TTCFont ascent reports wrong result
by John Pierce 27 Mar '06

27 Mar '06

Hi all, I got in trouble last time I posted a fix to this list. I'm not sure how to post fixes, but I did make a bug report in Mantis (issue #0003355). The problem is TTCFont used to have an "ascent" instance variable that the "ascent" method expects to use. The variable used to exist about October of 2005 and no longer does, yet the method still references it. Thus, the "ascent" method seems to report the wrong answer and most TTCFonts are not appropriately sized if "autofit" is enabled on the TextMorph. So, this fix is easy. Not sure where to post it, but just add "ascent" to the instance variable list of TTCFont. Regards, John -- It's easy to have a complicated idea. It's very very hard to have a simple idea. -- Carver Mead

3 2

Fwd: SUnit: Skipping tests?
by stéphane ducasse 27 Mar '06

27 Mar '06

Begin forwarded message: > From: Joseph Pelrine <jpelrine(a)metaprog.com> > Date: 27 mars 2006 09:15:11 HAEC > To: stéphane ducasse <ducasse(a)iam.unibe.ch> > Subject: Re: SUnit: Skipping tests? > > The simple way is to set up "environment variables" for tests, and > have a TestCase subclass check that these vars are valid as a > preconditoin to running the tests. The perfect way is to re-do > SUnit, adding yet another state to {pass|fail|error}, and another > exception to mark this. Note the additional level of complexity > this brings to TestResult>>#runCase: > > I did the first, and wrote to the Transcript a log of which tests > were skipped. I think that's the was Alan did it too. > > Joseph > At 08:20 27.03.2006, stéphane ducasse wrote: >> And whaat was your solution? >> >> Stef >> On 27 mars 06, at 07:39, Joseph Pelrine wrote: >> >>> Sure, I've done that already. IIRC, Glorp has these kind of tests. >>> >>> Joseph >>> At 07:30 27.03.2006, stéphane ducasse wrote: >>>> what do you think on that? >>>> I thought I would add a test kind. >>>> >>>> Stef >>>> >>>> Begin forwarded message: >>>> >>>>> From: Andreas Raab <andreas.raab(a)gmx.de> >>>>> Date: 27 mars 2006 01:08:17 HAEC >>>>> To: The general-purpose Squeak developers list <squeak- >>>>> dev(a)lists.squeakfoundation.org> >>>>> Subject: SUnit: Skipping tests? >>>>> Reply-To: The general-purpose Squeak developers list <squeak- >>>>> dev(a)lists.squeakfoundation.org> >>>>> >>>>> Hi Folks - >>>>> >>>>> I am in the interesting situation that I'm writing a few tests >>>>> that >>>>> require large data sets for input and where I don't want people to >>>>> require to download the data sets. My problem is while it's >>>>> easy to >>>>> determine that the data is missing and skip the test there isn't a >>>>> good way of relaying this to the user. From the user's point of >>>>> view "all tests are green" even though that statement is >>>>> completely >>>>> meaningless and I'd rather communicate that in a way that says "X >>>>> tests skipped" so that one can look at and decide whether it's >>>>> useful to re-run the tests with the data sets or not. >>>>> >>>>> Another place where I've seen this to happen is when platform >>>>> specific tests are involved. A test which cannot be run on some >>>>> platform should be skipped meaningfully (e.g., by telling the user >>>>> it was skipped) rather than to appear green and working. >>>>> >>>>> Any ideas? >>>>> >>>>> Cheers, >>>>> - Andreas >>> >>> -- >>> Joseph Pelrine [ | ] >>> MetaProg GmbH >>> Email: jpelrine(a)metaprog.com >>> Web: http://www.metaprog.com >>> >>> You don't become enormously successful without encountering some >>> really interesting problems. >>> - Mark Victor Hansen >>> >> > > -- > Joseph Pelrine [ | ] > MetaProg GmbH > Email: jpelrine(a)metaprog.com > Web: http://www.metaprog.com > > You don't become enormously successful without encountering some > really interesting problems. > - Mark Victor Hansen > >

1 0

Monticello infinite loop
by nicolas cellier 26 Mar '06

26 Mar '06

In a fresh 3.9a 7020 image, i happened to have infinite loops on most actions in the MC Browser (showing the Diffing... progress bar repeatedly). Interrupting the process, the looping message was the #addToCache method sent on most MC actions. Stepping into, this was due to the storeDiffs instVar set to true for my MCCacheRepository default. Inspecting instVar refs and senders, the only guilty is the 'store diffs' menu item in MC Browser. So, i am guilty for this (unwanted) wrong manipulation, but i feel like it's a little too easy to break things down... Shouldn't this store diffs item be protected in case of cache repository ? (it's easy to redefine MCCacheRepository>>#alwaysStoreDiffs as always answering false, and thus have it ignoring the menu action, but i would not change a code i do not understand enough). Couldn't this item be reserved to advanced users and hidden to Monticello novices like me ? (having an advanced submenu implicitly stating don't put your nose therein). I put this one in mantis http://bugs.impara.de/view.php?id=3359 Nicolas

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
26
Older →

Jump to page: