Squeak-dev September 2002

squeak-dev@lists.squeakfoundation.org

165 participants
288 discussions

[Enh]Auto Scroll
by Torge Husfeldt 01 Oct '02

01 Oct '02

subtitle: "[Q]how many wheels am i reinventing right now." Hi List, I am planning to program a proof of principle for a 2D Point&click advinter a la "Baphomets Fluch"/"Monkey Island"/"Indiana Jones 3/4". As a first step i wrote a sibling to TwoWayScrollPane called PanningView that has no scrollbars but scrolls automatically when the mouse comes near the edges. I made this behavior also accesable to all ScrollPanes by means of the CustomMenu. Please see the attached changset and comment on one or more of the following points: -do i reinvent wheels -does this seem remotely useful -does this file in correctly -does this work in your image -does my solution appeal in any way (i.e. what would you have done better) Thankful for any feedback. Torge

2 2

Re: [ANN] Logic Circus are ready
by Bob Arning 01 Oct '02

01 Oct '02

On Mon, 30 Sep 2002 16:27:20 -0700 Tommy Thorn <thorn(a)meko.dk> wrote: >Is it supposed to be editable? I tried middle-clicking on the yellow pad >of a new and gate (to connect it to the circuit), but I got >MessageNotUnderstood: cambiar, and it seems nobody implements cambiar. Hmm... I wonder if you changed that to cambiado would it work? Cheers, Bob

1 0

[BUG][FIX] Thumbnails in Holders (EToy)
by Bert Freudenberg 01 Oct '02

01 Oct '02

It seems like noone used thumbnails in Holders before? My son needed this to hold levels for a game. Since each level is almost full-screen, thumbnailing them was the only viable solution I could think of. So here's a one-line fix to make it work: "Change Set: ThumbHolderFix-bf Date: 30 September 2002 Author: Bert Freudenberg Makes getValueAtCursor and getGraphicAtCursor for a Holder showing thumbnails return the represented morph. Before, the thumbnail itself was returned which caused the animation to be played at thumbnail size." -- Bert

2 1

Migrating old AlignementMorph code to TableLayout
by Stephane Ducasse 30 Sep '02

30 Sep '02

Hi all I stupidly played around with SameGame and naively copied the AlignmentMorph code ;( Now I would like to use TableLayout and other new stuff (so that my book is not deprecated before been finished). Is there any path to follow. I'm rading the project on MorphLayoutPolicy Here is the code I have: Dr. Stéphane DUCASSE (ducasse(a)iam.unibe.ch) http://www.iam.unibe.ch/~ducasse/ "if you knew today was your last day on earth, what would you do different? ... especially if, by doing something different, today might not be your last day on earth" Calvin&Hobbes

1 0

Re: fading Color
by Bob Arning 30 Sep '02

30 Sep '02

On Mon, 30 Sep 2002 17:42:02 +0200 Chris Burkert <christian.burkert(a)s2000.tu-chemnitz.de> wrote: >I'm searching for a way to fill a Morph with more than one Color. >I hope 'fading Color' describes it well. The best example is the >MoviePlayer from the Widgetsflap but I didn't find how to get this >working. Can you help me ? Chris, "Gradient" is the general word you are looking for. Check methods that have 'gradient' in them using the method finder. One implementor of several such methods is the ObjectPropertiesMorph - you can get one for any morph via the eyedropper (magenta) halo handle. Cheers, Bob

1 0

ECOOP Tutorial, Worksheet1B - problem in Windows 2000
by Alex Peake 30 Sep '02

30 Sep '02

I have been working through the tutorial (ECOOP Tutorial, Worksheet1B). When building the Roladex, I get to the "eye" handle and get the "miscellaneous" script menus. I drag it into the desktop. However, clicking on the up and down arrows mostly does nothing. It starts with "send to back", and very occasionally I see "send to backt" and if I select the text it becomes "Bring to front". Sometimes I see "GIF file" and once I saw "BMP file". Every once in a while a large (two page) menu pops up. I select "serach for text" (the tutorial selection) and I get the "GIF file" selection. I cannot consistently get the menu. I think something is broken? Alex

1 0

Call for proposals for hosting the ESUG Smalltalk Conference 2003
by Noury Bouraqadi 30 Sep '02

30 Sep '02

Apologies for multiple receptions. -------------------- *** Call for Proposals for Hosting the ESUG 2003 Smalltalk Conference ESUG (http://www.esug.org) is a non-profit organization which was founded in 1991. Its goal is promoting object technology within the context of the Smalltalk programming language. To acheive this goal, ESUG undertakes many activities, such as: -organizing conferences and workshops that allow users and vendors and also industry people and academics to meet, -providing free teaching materials and books to students and universities, -supporting local Smalltalk users groups and helping establish new ones, -... ESUG is looking for local organizers to host its 2003 conference. The conference will take place the last week of August 2003 (i.e. from Saturday August 23rd to Friday August 30th). Please send proposals to mailto:esug-info@esug.org by November 15th, 2002. Each proposal should include three names and email addresses for the local organization contact people, and give information about location, hosting, ... For a sample, please look at http://www.iutc3.unicaen.fr:8000/esug2002/ Some practical information: - At the 2002 edition of the conference at Douai, France, we had around 100 participants. A room with a broadband internet connection and an excellent infrastructure was open 24h/day. - We had a student volunteer program to help the local organizers to take care of the coffee break, material, badges, projectors... - If you want to offer to organize ESUG, you should consider that we would like to also host CampSmalltalk, and to have a two track sessions. For this reason we will need two amphitheaters, and a room with network access. - The ESUG Board takes care of the conference program. - Before chosing your location we will contact you to check certain details about the organization of the social event, and about hosting possibilities. Ideally, we would like to be able to have everybody in a single location. - We will also provide you a list of tips and tricks to organize ESUG, so that you can profit from the experience we have accumulated over the years. We hope to hear about you soon, and hope that you will suggest a great place for the 11th edition of the ESUG conference. The ESUG Board. -- ------------------------------------------ Dr. Noury Bouraqadi Ecole des Mines de Douai - Dept. G.I.P http://csl.ensm-douai.fr/noury ------------------------------------------

1 0

Yahoo archive truncated
by Scott Wallace 30 Sep '02

30 Sep '02

I noticed that the Yahoo archive of the Squeak mailing list (http://groups.yahoo.com/group/squeak/) has been trimmed so that everything before December 2001 seemingly is no longer accessible. Does anyone know why this happened, and how we can get the history restored, and how we can avoid this happening again in the future? Cheers, -- Scott

12 20

RE: [FIX] SortedCollectionFix-sr
by Richard A. O'Keefe 30 Sep '02

30 Sep '02

Ian Piumarta <ian.piumarta(a)inria.fr> wrote that: [performant i]s a perfectly good french adjective, and means precisely... > "[something] performs well, has good performance" and is therefore > "performant". ...that. If it isn't part of English, it should be. (Who do we petition? L'Académie anglaise ?? ;-) I like to count syllables. I note that "more performant" is four syllables, "faster" is two. The problem is that "performs well, has good performance" is seriously ambiguous. For cars, for example, we can say unambiguously: X accelerates faster than Y. X uses less fuel than Y. X is quieter than Y. X uses less oil than Y. X has more leg-room than Y. X emits fewer particles than Y. X costs less than Y to buy. X costs less than Y to run. X smells nicer than Y. X pulls more birds than Y. (Sorry about the dated idiom.) ... But what does 'performs better' mean? With respect to computer software, X is faster than Y. X is faster than Y for large problems. X uses less memory than Y. X is more accurate than Y. X passes more tests than Y. X is cheaper than Y. ... are all aspects of performance. If someone tells me 'X is more performant than Y', then even with the French dictionary in front of me I know nothing more than this: the speaker approves of X. I do not know *why*. In computing discussions, let us prefer *specific* terms to words and phrases like "performant" and "performs well".

1 0

Re: [FIX] SortedCollectionFix-sr
by Richard A. O'Keefe 30 Sep '02

30 Sep '02

Stephan Rudlof <sr(a)evolgo.de> wrote: your proposal implementing SortedCollection with 'lazy data structures' is very interesting. ... One question is: How is the relation between changes of the SortedCollection and looking >>at: it, since some time is spend for the checks there. If accessing it outweights the changes by magnitudes there could be a *small* performance penalty. The first question is whether the overhead can be made small. Retaining the names a, m, z, instead of calling self pvtEnsureSorted one would do m == z ifFalse: [self pvtEnsureSorted] which costs push, push, compare, branch. I made a subclass SlowerCollection of SortedCollection. The only change was that I made a copy of #at: and stuffed firstIndex == firstIndex ifFalse: [self error: 'Impossible']. into it at the appropriate place. SortedCollection time for a million sends of #at: 1.901 seconds. SlowerCollection time for a million sends of #at: 2.071 seconds. That is, adding this check made #at: about 9% slower. BUT WE CAN DO BETTER! SortedCollection inherits #at: from OrderedCollection, which does some dynamic bounds checking. Use four variables: firstIndex (from OrderedCollection), lastIndex (from OrderedCollection), lastSortedIndex (for laziness), and fakeLastIndex (to make #at: fast). Ensure that fakeLastIndex = (lastSortedIndex = lastIndex ifTrue: [lastIndex] ifFalse: [-1]). Then have #at: check against fakeLastIndex instead of lastIndex. In the branch that currently handles out-of-range indices, #at: would restore the stronger invariant and then resend to super. With that hack the time penalty would be ZERO except for the first call to #at: after an addition. The basic reason we can make the relative time penalty for SortedCollection effectively zero is that the time penalty for #at: in OrderedCollection is already rather high. If anyone was really worried about the speed of #at:, they should be thinking about writing a primitive for OrderedCollection>>at:.

1 0

Jump to page: