Squeak-dev May 2002

squeak-dev@lists.squeakfoundation.org

160 participants
453 discussions

Re: Sorting oddity - feature or ??
by Hannes Hirzel 01 May '02

01 May '02

Bob Arning <arning(a)charm.net> wrote: > Hi all, > > Is the following a known and accepted feature of our sorting algorithm? > > [((1 to: 10000) collect: [ :x | x]) asSortedCollection] timeToRun ==> 98 > > [((1 to: 10000) collect: [ :x | 3]) asSortedCollection] timeToRun ==> 20826 > > Cheers, > Bob No, I don't think it is known and accepted. Another test: [((1 to: 10000) collect: [ :x | 10001 - x]) asSortedCollection] timeToRun ==> 87 asSortedCollection calls newFrom: aCollection "implemented in OrderedCollection" "Answer an instance of me containing the same elements as aCollection." | newCollection | newCollection _ self new: aCollection size. newCollection addAll: aCollection. ^newCollection And #addAll: is implemented in SortedCollection addAll: aCollection aCollection size > (self size // 3) ifTrue: [aCollection do: [:each | self addLast: each]. self reSort] ifFalse: [aCollection do: [:each | self add: each]]. ^ aCollection which calls #add: add: newObject ^ super insert: newObject before: (self indexForInserting: newObject) which finally calls #insert:before: and #indexForInserting: Two methods which are fairly complex and must contain an error somewhere. Are there sorting specialists out there? Regards Hannes

4 3

[FIX] SmallIRCTweaks
by bparsia＠email.unc.edu 01 May '02

01 May '02

If you want to incorporate these and bounce a .cs back to me, I will send this to Scott and Dan. I figured out the mail to list issue, and it was User Error :) Cheers, Bijan. from preamble: "Change Set: SmallIRCTweaks Date: 29 April 2002 Author: Bijan Parsia Two bits: 1) Fixed the '/me' handling (it should convert to an /action, not sure if I hacked it into the best spot. 2) Pulled Stephan's world menu changes and just made the 'irc client' item open the enhanced client. A preference could be added if anyone really wants it."!

6 8

Re: Sorting oddity - feature or ??
by Jon Hylands 01 May '02

01 May '02

At Wednesday, 1 May 2002, you wrote: >No, I don't think it is known and accepted. >And #addAll: is implemented in SortedCollection > >addAll: aCollection > aCollection size > (self size // 3) > ifTrue: > [aCollection do: [:each | self addLast: each]. > self reSort] > ifFalse: [aCollection do: [:each | self add: each]]. > ^ aCollection When you send asSortedCollection to an unsorted collection, the ifTrue part is what is used -- it adds all the elements, and then does an in-place sort. I haven't checked Squeak, but I assume it uses some variation of Quicksort. This (a large collection equal values) is probably just a worst-case scenario for whatever sorting algorithm is used. Later, Jon -------------------------------------------------------------- Jon Hylands Jon(a)huv.com http://www.huv.com/jon Project: Micro Seeker (Micro Autonomous Underwater Vehicle) http://www.huv.com

1 0

[ENH] Dates-raok
by ok＠cs.otago.ac.nz 01 May '02

01 May '02

from preamble: "Change Set: Dates-raok Date: 19 April 2002 Author: Richard A. O'Keefe My original intent was just to add a few class comments, but I soon noticed a few missing methods (enumerating over the days of a month, for example). That's not all that's missing, a DateInterval class (with Week and Month as subclasses of that instead of Date) is clearly called for. But that's too much refactoring for one morning. I also noticed that the instance creation methods of Date were inherited by Month, and could too easily be used to create invalid Months, so they had to be overridden. Very little has been done to Week; there should be two classes MondayWeek and SundayWeek instead of the 'startMonday' hack. Date>>previous: appeared to be broken, but Week relied on the breakage, so I have just noted the problem in a comment."!

2 1

Sorting oddity - feature or ??
by Bob Arning 01 May '02

01 May '02

Hi all, Is the following a known and accepted feature of our sorting algorithm? [((1 to: 10000) collect: [ :x | x]) asSortedCollection] timeToRun ==> 98 [((1 to: 10000) collect: [ :x | 3]) asSortedCollection] timeToRun ==> 20826 Cheers, Bob

1 0

[FIX] InterpreterPlugin-halt-dgd
by DiegoGomezDeck＠ConsultAr.com 01 May '02

01 May '02

from preamble: "Change Set: InterpreterPlugin-halt-dgd Date: 1 May 2002 Author: Diego Gomez Deck <DiegoGomezDeck(a)ConsultAr.com> fixed bug: InterpreterPlugin>>halt send #halt to self instead of super"!

1 0

Re: Remote access to Celeste db?
by Bruce ONeel 01 May '02

01 May '02

"Lex Spoon" <lex(a)cc.gatech.edu> wrote: > So many possibilities, none of which are implemented. :) Cute, that's worth a .sig file entry. Do you mind? cheers bruce

1 0

Re: rounded corners (was: Re: straw-man 3.2 default preferences)
by Martin Drautzburg 01 May '02

01 May '02

Scott Wallace <scott.wallace(a)squeakland.org> writes: > I suppose I agree, reluctantly, with Henrik's position about rounded > corners, and would be inclined to change the out-of-box look of 3.2 > accordingly. > > Opinions? I turned rounding off on my box.

2 1

Re: [e-lang] Java Serialization Re: Alan Kay in the News
by Rob Withers 01 May '02

01 May '02

"Mark S. Miller" <markm(a)caplet.com> wrote: > At 04:19 PM 4/30/2002 Tuesday, Blaine Buxton wrote: > >I am currently working on reading in Java Serialized Files as we speak into > >Squeak. Right now, I have simple serialized java objects working (in other > >words, they don't override the serialization interface (custom serialization) nor implements > >externalizable), but I should have that working soon. > > That's great! CapTP currently uses Java serialization. I don't like that That is great! Blaine, I'd be very interested to see your work. I tried extending Goran's Brimstone framework with a Java Parser, but quicky realized that it was a tangent I didn't want to take. The result is the eventual sending framework. Do you happen to have a Java Parser in squeak? > for all the reasons you mention, and wish to eventually switch to an > alternative, as we've been discussing. However, it sounds like your work > would enable Rob's Squeak-E efforts to interoperate with the current E > before we figure out the right serialization answer. That'd be great for a Yes! This would allow a quick jump from a baseline squeak distributed model (what I am doing now) to a CapTP implementation. :) > number of reasons, but especially because we'd then be trying to find the > right answer after some experience running cross platform. Do you also This is the exact reason I am doing a squeak only distribution first. To get some experience running eventual sending cross image. It seems that I am a step behind. > expect to be able to write Java serialization streams? That would be > necessary of course. > > Rob, does this sound like it could work for you? Certainment! cheers, Rob

1 0

[e-lang] Squeak-E
by Rob Withers 01 May '02

01 May '02

Cees, Mark, I am completely blown away by the enthusiasm. I am so happy that you all find this an interesting direction for squeak. Let me say welcome as well! At the moment I have very little of substance to bring to this discussion, aside from the implementation I am working on. I hope to have a distributed squeak version of eventual sending, RSN. It's not yet a true Squeak-E since it doesn't support the E language or VatTP...yet. I will post working code as soon as I get it going. The project roadmap I have had in mind is: - eventual-sends (done) - distributed eventual-sending in squeak (close) - distributed eventual Linda (should be easy to extend LindaTalk) - solidify event-loop concurrency (close look at Andreas' User Scripting) - security - authentication/encryption (nada planned....any security experts want to tackle CapTP's security in squeak?) - security - sealer/unsealer/grant/... (ditto) - VatTP implementation in Squeak (The message protocol described on the erights site) - common serialization (experts needed...experts found :) - migrate Lex' Islands work (compiler changes, globals, primitives) - multiuser squeak (pipe dream) - persistence/transactions (Kats/LOOM) clearly there is much room for collaboration :) ... cg(a)cdegroot.com wrote: > Mark S. Miller <markm(a)caplet.com> said: > >So don't be shy about proposing Squeak-based or Parcels-based serialization > >formats. > > > Braindump: > > - SRP (State Replication Protocol) is a Smalltalk-platform-neutral > protocol with some nice properties, and is expressly meant to be usable > on platforms besides Smalltalk. Especially the data encoding is cool, but > apart from that it also has good support for class evolution, etcetera. > http://wiki.cs.uiuc.edu/CampSmalltalk/About+State+Replication+Protocol+(SRP) The generic serialization framework which I mentioned in a previous email, was SRP. I was exposed to SRP several years ago, at Camp Smalltalk 1. I have CCed Paul Baumann, in case he wished to say some things about using SRP in a cross-language environment. cheers! Rob :-)

1 0

Jump to page: