Squeak-dev July 2002

squeak-dev@lists.squeakfoundation.org

205 participants
393 discussions

[BUG] Scrolling in PluggableListMorph on Linux
by dsiegel＠acm.org 29 Jul '02

29 Jul '02

I'm running a 3.2 image on both Windows 98 and Mandrake Linux 8.2. The image is configured as delivered, no changed preferences. On Linux, clicking in the scrollbar of a PluggableListMorph sometimes results in several page-scrolls rather than just one. To see this open a Browser and click in the scrollbar of category pane. On my system, within half a dozen clicks, the scrollbar will start "autoscrolling". This does not happen on Windows. Can anyone duplicate this behavior? It's especially annoying when using Celeste -- it's infuriating when I try to click to the previous page of message headers, and they just stream by. I don't know how to begin to debug this one. -dms

4 3

Re: grabbing value of a variabe by its name?
by Richard A. O'Keefe 29 Jul '02

29 Jul '02

Ragnar Hojland Espinosa <ragnar(a)linalco.com> is still trying to defend a maintenance-disaster-type pun. If I got to choose which code to maintain, this: execute 'insert into table1 (firstName,lastName,position,address,zip,country) values (', firstName asQuotedPostgres,',', lastName asQuotedPostgres,',', position asQuotedPostgres,',', address asQuotedPostgres,',', zip asQuotedPostgres,',' country asQuotedPostgres,')'. "ad nauseum" This is a straw man; it's easy to knock over but there is no real substance to it. Nobody in his or her right mind would write code like this, and I have given several examples in this thead of how it could be done better. The starting point is the observation that the data are not magically born from Zeus's brow already in variables that have exactly the right names. The data come from somewhere. When they are stored, there is not the slightest reason to put them in method temporaries, and every reason to put them somewhere else. To give my example before, r := Table1 new. r firstName: "wherever the data REALLY come from". r lastName: "wherever the data REALLY come from". r position: "wherever the data REALLY come from". r address: "wherever the data REALLY come from". r zip: "wherever the data REALLY come from and what have hot water heaters got to do with it anyway?". r country: "wherever the data REALLY come from". "etc ad nauseam spon". r insert. This is doing things an OBJECT-ORIENTED way. You want to insert something into a table. What do you want to insert? Variables? No: you want to insert an OBJECT. So let there *BE* an object. In particular, this carries over to update: Table1 do: [:r | r country = #NZ ifTrue: [r price: r price * 0.80. r replace]]. or this: fields _ SQLFields new: #(firstName lastName position address zip country). execute 'insert into table1 (', fields asNames, ') values (', fields asValues, ')'. I know which I would choose. Please reconsider. This is the dark side of Perl. Not only its quicker and cleaner, It is quicker and cleaner than your straw man, but it is NOT quicker or cleaner than doing things in an object-oriented way. For an example of the way it's not object-oriented, you MAKE an object, but then you don't USE it as an object. Why not (SQLFields new: #(firstName lastName position address zip country)) insertInto: #table1. its also safer safer It is safer than your straw man. What wouldn't be? But it is NOT safer than doing things in an object-oriented way. For example, suppose we have a Table1 class, the way I keep suggesting, whose instances represent past, current, future, actual, or possible rows of 'table1'. The methods in that class can - provide default values - enforce integerity constraints before touching the data base - handle version changes For example, suppose the column 'zip' in your data base is renamed from the US-centric 'zip' to a more globally intelligible 'postcode'. ALL of your versions break, but the Table1 class can just add two methods zip ^self postcode zip: x self postcode: x and your code keeps running. THIS is what I call safety; depending on accidental puns is what I call HIGHLY RISKY. and encourages a coherent homogeneus naming convention across classes for data extracted from the database with a localized "trick" I completely fail to see how this trick can possibly encourage a 'coherent homogeneus (sic!)' naming convention. It fails at the first step: * It simply isn't *POSSIBLE* to use the same names in your Smalltalk code as in your data base. Here is a sample table I happen to have kicking around. create table Results ( Round decimal(2) unique not null, -> Home_Team character(14), -> Home_Score decimal(2) not null, -> Away_Team character(14), -> Away_Score decimal(2) not null, foreign key (Home_Team) references Teams(Team), foreign key (Away_Team) references Teams(Team), primary key (Round, Home_Team, Away_Team) ); I suppose you could use #round as a method temporary, but you could NOT use #home_team, #home_score, #away_team, or #away_score in Squeak. (ANSI Smalltalk, yes. Squeak, no.) * The trick does not particularly encourage consistency. SQL names are not case sensitive. So homeTeam, homeTEAM would both represent the same SQL name. How exactly does the trick encourage people to consistently pick one of them, in a way that gentle exhortation would not? * Sometimes in SQL you have to use qualified names. The example data base that my example comes from includes also create table Teams (Team ...) create table League (Team ...) You have not yet explained how your pun can accomodate SQL names like Teams.Team or League.Team. * Some names which are perfectly usable as SQL names are reserved words in Smalltalk, such as "self", "super", "nil". * You have not yet explained how your pun can accomodate SQL delimited identifiers. For people who never looked at SQL 92, "foo!" (with the double quotes) is a perfectly good SQL 92 'delimited identifier'. Delimited identifiers may be used pretty much anywhere that simple identifiers may be. * According to the SQL 1992 standard, an identifier may contain non-Latin letters if such exist according to the <module character set specification> or the <character set specification>; how are those to be handled by this pun? But above all, the trick fails at the second step. Suppose in one method you need to refer to TWO tables. t := Teams new. t team: 'Hawkeyes'. t insert. lg := League new. lg team: t team. lg wins: 0; losses: 0; draws: 0; forScore: 0; againstScore: 0. lg insert. In this case, the two rows happen to need the same value for their 'team' columns. But what if they needed DIFFERENT values for their team columns? winners := 'Hawkeyes'. losers := 'Windies'. w := Wins new. w team: winners; against: losers. w insert. l := Losses new. l team: losers; against: winners. l insert. I cannot for the life of me regard any approach which would force the entry into Wins and the entry into Losses to use the *same* local variables for opposite purposes as 'safe'.

2 1

Re: New Random Number Generator
by Richard A. O'Keefe 29 Jul '02

29 Jul '02

The random number generator is a strong candidate for being a plugin. It really is a fast one. A good way to use it is to fill a buffer at a time and dole the numbers out using a stream.

2 1

squeak on freshmeat, and rpms and debs.
by Bruce ONeel 29 Jul '02

29 Jul '02

Hi, I've just updated Squeak on freshmeat to 3.2. This brought up a good question, what changed between 3.0 and 3.2 besides packages and themes? The freshmeat request is waiting for approval. Along the lines of releases, does any one want to make RPMs and Debs for the ftp area? That would be great, just point me to them. Thanks! cheers bruce

1 0

Re: New Random Number Generator
by Mark4Flies＠aol.com 29 Jul '02

29 Jul '02

The information about the different way that Squeak and C handle unsigned long integers in shift operations was interesting but amazingly it made no difference in the outcome. Here is the revised class definition. -Mark 'From Squeak3.0 of 4 February 2001 [latest update: #3545] on 29 July 2002 at 8:24:39 am'! Random subclass: #MTRandom instanceVariableNames: 'mt mti ' classVariableNames: 'LowerMask M MatrixA N TemperingMaskB TemperingMaskC UpperMask ' poolDictionaries: '' category: 'Kernel-Numbers'! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:07'! mask: aWord1 and: aWord2 "Perform bit logic and maintain unsigned four-word 32-bit integers" ^ ((aWord1 bitAnd: UpperMask) bitOr: (aWord2 bitAnd: LowerMask))! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:06'! temperingShiftL: anInteger "Second of four Mersenne Twister shifts" ^ (anInteger >> 18 bitAnd: 2r11111111111111)! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:06'! temperingShiftS: anInteger "Second of four Mersenne Twister shifts" ^ (anInteger << 7 bitAnd: 16rFFFFFFFF)! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:06'! temperingShiftT: anInteger "Second of four Mersenne Twister shifts" ^ (anInteger << 15 bitAnd: 16rFFFFFFFF)! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:06'! temperingShiftU: anInteger "First of four Mersenne Twister shifts" ^ (anInteger >> 11 bitAnd: 2r111111111111111111111)! ! !MTRandom methodsFor: 'supporting' stamp: 'mwb 7/29/2002 08:07'! xor: aWord1 and: aWord2 ^ ((aWord1 bitXor: (aWord2 >> 1 bitAnd: 2097151)) bitXor: ((WordArray with: 0 with: MatrixA) at: (aWord2 bitAnd: 1) + 1))! ! !MTRandom methodsFor: 'initialization' stamp: 'mwb 7/29/2002 08:08'! initialize "Set up space for seeds, then fill them in" | s | mt _ WordArray new: N. mti _ N + 1. [s _ Time millisecondClockValue. s = 0] whileTrue. self sgenrand: s! ! !MTRandom methodsFor: 'accessing' stamp: 'mwb 7/27/2002 09:47'! next "Answer a random Float in the interval [0 to 1)." ^ (self nextValue / 16rFFFFFFFF) asFloat! ! !MTRandom methodsFor: 'accessing' stamp: 'mwb 7/26/2002 17:37'! nextInt: anInteger "Answer a random integer in the interval [1, anInteger]." ^ (self next * anInteger) truncated + 1! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/27/2002 11:16'! genrand | y | self mti >= N ifTrue: [self mti = (N + 1) ifTrue: [self sgenrand: 4357]. 1 to: N - M do: [:kk | y _ self mask: (self mt at: kk) and: (self mt at: kk + 1). self mt at: kk put: (self xor: (self mt at: kk + M) and: y)]. N - M + 1 to: N - 1 do: [:kk | y _ self mask: (self mt at: kk) and: (self mt at: kk + 1). self mt at: kk put: (self xor: (self mt at: kk + (M - N)) and: y)]. y _ self mask: (self mt at: N) and: (self mt at: 1). self mt at: N put: (self xor: (self mt at: M) and: y). mti _ 1]. y _ self mt at: mti. mti _ mti + 1. y _ (y bitXor: (self temperingShiftU: y)) bitAnd: 16rFFFFFFFF. y _ ((y bitXor: (self temperingShiftS: y)) bitAnd: TemperingMaskB) bitAnd: 16rFFFFFFFF. y _ ((y bitXor: (self temperingShiftT: y)) bitAnd: TemperingMaskC) bitAnd: 16rFFFFFFFF. y _ (y bitXor: (self temperingShiftL: y)) bitAnd: 16rFFFFFFFF. ^ y! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 11:38'! mt ^ mt! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 16:23'! mti ^ mti! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 16:23'! mti: anInteger mti _ anInteger! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 20:56'! nextValue "This method generates random instances of Integer. The algorithm is described in detail by M. Matsumoto and T. Nishimura in 'Mersenne Twister: A 623-dimensionally equidistributed uniform pseudorandom number generator', ACM Trans. on Modeling and Computer Simulation Vol. 8, No. 1, January pp.3-30 1998." ^ self genrand! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/29/2002 07:45'! seed: anInteger seed _ anInteger. self sgenrand: anInteger! ! !MTRandom methodsFor: 'private' stamp: 'mwb 7/26/2002 17:36'! sgenrand: anInteger mt at: 1 put: (anInteger bitAnd: 16rFFFFFFFF). 2 to: N do: [:i | self mt at: i put: (69069 * (self mt at: i - 1) bitAnd: 16rFFFFFFFF)]. self mti: N ! ! "-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- "! MTRandom class instanceVariableNames: ''! !MTRandom class methodsFor: 'instance creation' stamp: 'mwb 7/26/2002 10:54'! new "Create new instance of MTRandom" ^ super new initialize! ! !MTRandom class methodsFor: 'class initialization' stamp: 'mwb 7/27/2002 11:23'! initialize " MTRandom initialize " "Set period parameters" N _ 624. M _ 397. MatrixA _ 16r9908B0DF. LowerMask _ 16r7FFFFFFF. UpperMask _ 16r80000000. "Set tempering parameters" TemperingMaskB _ 16r9D2C5680. TemperingMaskC _ 16rEFC60000! ! MTRandom initialize!

1 0

Re: New Random Number Generator
by Mark4Flies＠aol.com 29 Jul '02

29 Jul '02

> The random number generator is a strong candidate for being a plugin. > It really is a fast one. A good way to use it is to fill a buffer at > a time and dole the numbers out using a stream. I think that your idea to make this generator a plugin is great. You are way past me, though. First I have to fix it before I speed it up! Thanks! -Mark

1 0

Another Swiki defence idea: better tools
by cg＠cdegroot.com 29 Jul '02

29 Jul '02

The premise of a Swiki's openness is that it is easier to rollback than to damage it. Why not simply add a button to the recent changes screen, so that you can roll back any damage done from a single IP during a given time frame with one push of the button? Some smart ideas like that, some distributed system maintenance, and script kiddies are likely to give up very quickly (especially if the FrontPage explains that you've got these buttons in place, yaddayadda). -- Cees de Groot http://www.cdegroot.com <cg(a)cdegroot.com> GnuPG 1024D/E0989E8B 0016 F679 F38D 5946 4ECD 1986 F303 937F E098 9E8B Cogito ergo evigilo

5 5

Re: Glorp (WAS Re: Squeak and SQlite Database)
by Bruce ONeel 29 Jul '02

29 Jul '02

There is an ODBC driver for Postgres SQL at odbc.postgresql.org cheers bruce Nevin Pratt <nevin(a)smalltalkpro.com> wrote: > > > yampa(a)mindspring.com wrote: > > > > nevin(a)smalltalkpro.com wrote: > >> > > >>In about a month, I plan on revisiting Glorp for Squeak (using > >>PostgreSQL for the database). I intend on using it in a big way for my > >>wife's e-commerce web site: > >> > > I am interested in using Glorp as well. There was a Squeak ODBC interface > > posted here about a month ago. Would it be better to implement Glorp with > > ODBC, rather than PostgreSQL? It seems like we could use just about any > > database with it then. > > > > Doug R. > > > My wife's site is deployed on a FreeBSD box. PostgreSQL is also > mandatory for the site for a variety of reasons. I'm not sure what, if > any, ODBC support is available under FreeBSD for PostgreSQL. > > So, I'm inclined to continue using the Squeak PostgreSQL driver, and > forego ODBC. > > Nevin

1 0

Re: Glorp (WAS Re: Squeak and SQlite Database)
by yampa＠mindspring.com 29 Jul '02

29 Jul '02

I am interested in using Glorp as well. There was a Squeak ODBC interface posted here about a month ago. Would it be better to implement Glorp with ODBC, rather than PostgreSQL? It seems like we could use just about any database with it then. Doug R. On Fri, 26 Jul 2002 15:36:19 -0400 Nevin Pratt <nevin(a)smalltalkpro.com> wrote: >In about a month, I plan on revisiting Glorp for >Squeak (using >PostgreSQL for the database). I intend on using >it in a big way for my >wife's e-commerce web site: >http://www.bountifulbaby.com Nevin

2 1

Re: Taking the world!
by Martin Drautzburg 29 Jul '02

29 Jul '02

Diego Gomez Deck <DiegoGomezDeck(a)ConsultAr.com> writes: > Hi, > > With the 3.2 version on the road we can publish in slashdot, freshmet, etc > and take the world for us! :-P Absolutely. What smalltalk needs the most is press coverage. A good scandal would be nice, sex and crime, security alerts, hot fixes, a new smalltalk virus, managers facing court. Nobody wants to listen to the "smalltalk is better" mantra, as true as it is. I believe for slashdot you need an original article to point to. I made two submissions about smalltalk so far but they posted neither of them.

2 1

← Newer
1
...
4
5
6
7
8
9
10
...
40
Older →

Jump to page: