Squeak-dev November 2004

squeak-dev@lists.squeakfoundation.org

164 participants
373 discussions

Question: What is the prefered tool for mailing list reading?
by Ruben Bakker 29 Nov '04

29 Nov '04

This mailing list is rather active. What tool do you use for reading mailing lists? Is it an normal e-mail client or is something more specialized to handle mailing list threads and filtering? Do you use Celeste or another squeak package? Thanks Ruben

5 5

Mac OSX display depth colour oddness
by Peace Jerome 29 Nov '04

29 Nov '04

Tim wrote: Bert Freudenberg <bert at impara.de> wrote: > Am 05.11.2004 um 22:15 schrieb Tim Rowledge: > > > I've just been testing out something in a 3.8b image on my pMac and > > noticed a strange colour effect. The default for the vanilla image is > > 16bpp and it looked just awful on a 32bpp display - very visible > > dithering on the gradient backgrounds of the menus, change in desktop > > colour etc. Setting Squeak to 32bpp made it looked ok. > > On Acorn it looks the same in 16 or 32bpp settings. :-) > > John only recently fixed this, in 3.7.5b1. SO it did. Looks much better now. tim -- addendum: I noticed this problem too and did a little looking into it. ON a 9.1 MacOS, I'm running sq 3.9a (seems to be in all of them) on mostly the Squeak 3.7.4Beta1.app. Gradient fills would have a sort of stepping effect if squeaks screen depth setting did not match my macs hdwe screen depth and go away if it did. When I tried this with a Squeak 3.7.5Beta1.app vm. The same thing happened! ! ! I tried grabbing a rectange of the funny gradient and exporting it as a jpg but when I later viewed it with picture viewer it was perfectly normal. So my suspicion was not the vm but the stuff inside squeak that converts between screen depths on the way to the display screen. I know there are other things it doesn't do right so its high on my list of the usual suspects. My earlier squeaks wouldn't recognize a 32bit screen depth only 16bit and I believe I turned my hardware settings down to get better looking gradients. So that the bug has now appeared going the other way I consider progress. __________________________________ Do you Yahoo!? All your favorites on one personal page � Try My Yahoo! http://my.yahoo.com

2 1

RE: ODBC connection
by Pablo Iaria 29 Nov '04

29 Nov '04

Hi Ronald, mmmmmm...ROE only works with pgsql I have tried tantalus (another small mapping tool) and it worked fine with ODBC. Cheers! Pablo.- > -----Original Message----- > From: squeak-dev-bounces(a)lists.squeakfoundation.org [mailto:squeak-dev- > bounces(a)lists.squeakfoundation.org] On Behalf Of Welch, Ronald P (US SSA) > Sent: Lunes, 29 de Noviembre de 2004 03:29 p.m. > To: The general-purpose Squeak developers list > Subject: RE: ODBC connection > > Ooops sorry I meant ODBC not JDBC. > > ----------=-=-=-=-=-=-=-=-========oOo========-=-=-=-=-=-=-=-=---------- > mailto:Ronald.P.Welch@baesystems.com Phone:(607)770-3701 > BAE SYSTEMS Controls 600 Main St Johnson City, NY 13790-1888 > ----------=-=-=-=-=-=-=-=-===================-=-=-=-=-=-=-=-=---------- > > > -----Original Message----- > From: Welch, Ronald P (US SSA) > Sent: Monday, November 29, 2004 1:28 PM > To: The general-purpose Squeak developers list > Subject: RE: ODBC connection > > > Taking this one step futher, is there a way to get > ROE to talk to Oracle via JDBC? > > ----------=-=-=-=-=-=-=-=-========oOo========-=-=-=-=-=-=-=-=---------- > mailto:Ronald.P.Welch@baesystems.com Phone:(607)770-3701 > BAE SYSTEMS Controls 600 Main St Johnson City, NY 13790-1888 > ----------=-=-=-=-=-=-=-=-===================-=-=-=-=-=-=-=-=---------- > > -----Original Message----- > From: Pablo Iaria [mailto:iariap@hotpop.com] > Sent: Monday, November 29, 2004 1:19 PM > To: 'The general-purpose Squeak developers list' > Subject: RE: ODBC connection > > > Hi Alejandro, > > Take a look at http://minnow.cc.gatech.edu/squeak/2480 > > Cheers, > Pablo.- > > > -----Original Message----- > From: squeak-dev-bounces(a)lists.squeakfoundation.org > [mailto:squeak-dev-bounces@lists.squeakfoundation.org] On Behalf Of > Alejandro Torres > Sent: Lunes, 29 de Noviembre de 2004 03:16 p.m. > To: squeak-dev(a)lists.squeakfoundation.org > Subject: ODBC connection > > There is any way to connect to a database like Oracle using Squeak ? > > > > Thanks > -AlxTorres

1 0

RE: ODBC connection
by Welch, Ronald P (US SSA) 29 Nov '04

29 Nov '04

Ooops sorry I meant ODBC not JDBC. ----------=-=-=-=-=-=-=-=-========oOo========-=-=-=-=-=-=-=-=---------- mailto:Ronald.P.Welch@baesystems.com Phone:(607)770-3701 BAE SYSTEMS Controls 600 Main St Johnson City, NY 13790-1888 ----------=-=-=-=-=-=-=-=-===================-=-=-=-=-=-=-=-=---------- -----Original Message----- From: Welch, Ronald P (US SSA) Sent: Monday, November 29, 2004 1:28 PM To: The general-purpose Squeak developers list Subject: RE: ODBC connection Taking this one step futher, is there a way to get ROE to talk to Oracle via JDBC? ----------=-=-=-=-=-=-=-=-========oOo========-=-=-=-=-=-=-=-=---------- mailto:Ronald.P.Welch@baesystems.com Phone:(607)770-3701 BAE SYSTEMS Controls 600 Main St Johnson City, NY 13790-1888 ----------=-=-=-=-=-=-=-=-===================-=-=-=-=-=-=-=-=---------- -----Original Message----- From: Pablo Iaria [mailto:iariap@hotpop.com] Sent: Monday, November 29, 2004 1:19 PM To: 'The general-purpose Squeak developers list' Subject: RE: ODBC connection Hi Alejandro, Take a look at http://minnow.cc.gatech.edu/squeak/2480 Cheers, Pablo.- -----Original Message----- From: squeak-dev-bounces(a)lists.squeakfoundation.org [mailto:squeak-dev-bounces@lists.squeakfoundation.org] On Behalf Of Alejandro Torres Sent: Lunes, 29 de Noviembre de 2004 03:16 p.m. To: squeak-dev(a)lists.squeakfoundation.org Subject: ODBC connection There is any way to connect to a database like Oracle using Squeak ? Thanks -AlxTorres

1 0

RE: ODBC connection
by Welch, Ronald P (US SSA) 29 Nov '04

29 Nov '04

Taking this one step futher, is there a way to get ROE to talk to Oracle via JDBC? ----------=-=-=-=-=-=-=-=-========oOo========-=-=-=-=-=-=-=-=---------- mailto:Ronald.P.Welch@baesystems.com Phone:(607)770-3701 BAE SYSTEMS Controls 600 Main St Johnson City, NY 13790-1888 ----------=-=-=-=-=-=-=-=-===================-=-=-=-=-=-=-=-=---------- -----Original Message----- From: Pablo Iaria [mailto:iariap@hotpop.com] Sent: Monday, November 29, 2004 1:19 PM To: 'The general-purpose Squeak developers list' Subject: RE: ODBC connection Hi Alejandro, Take a look at http://minnow.cc.gatech.edu/squeak/2480 Cheers, Pablo.- -----Original Message----- From: squeak-dev-bounces(a)lists.squeakfoundation.org [mailto:squeak-dev-bounces@lists.squeakfoundation.org] On Behalf Of Alejandro Torres Sent: Lunes, 29 de Noviembre de 2004 03:16 p.m. To: squeak-dev(a)lists.squeakfoundation.org Subject: ODBC connection There is any way to connect to a database like Oracle using Squeak ? Thanks -AlxTorres

1 0

RE: ODBC connection
by Pablo Iaria 29 Nov '04

29 Nov '04

Hi Alejandro, Take a look at http://minnow.cc.gatech.edu/squeak/2480 Cheers, Pablo.- -----Original Message----- From: squeak-dev-bounces(a)lists.squeakfoundation.org [mailto:squeak-dev-bounces@lists.squeakfoundation.org] On Behalf Of Alejandro Torres Sent: Lunes, 29 de Noviembre de 2004 03:16 p.m. To: squeak-dev(a)lists.squeakfoundation.org Subject: ODBC connection There is any way to connect to a database like Oracle using Squeak ? Thanks -AlxTorres

1 0

ODBC connection
by Alejandro Torres 29 Nov '04

29 Nov '04

There is any way to connect to a database like Oracle using Squeak ? Thanks -AlxTorres

1 0

Set::atRandom is extremely nonuniform / suggested fixes
by Joshua Scholar 29 Nov '04

29 Nov '04

Oops there was a typo in my suggested Set::AtRandom Here's the one I'm actually using: atRandom: aGenerator "Answer a random element of the receiver. Uses aGenerator which should be kept by the user in a variable and used every time. Use this instead of #atRandom for better uniformity of random numbers because only you use the generator. Causes an error if self has no elements." | ind entry | entry _ nil. [entry == nil] whileTrue: [ self emptyCheck. ind _ aGenerator nextInt: array size. entry _ array at: ind. ]. ^entry.

1 0

Set::atRandom is extremely nonuniform / suggested fixes
by Joshua Scholar 29 Nov '04

29 Nov '04

Set::atRandom and therefore everything derived from Set (like Dictionary) has a terrible atRandom. An entry with 3 nils in front of it is 4 times as likely to be chosen as one with no nils in front of it. I recommend changing from atRandom: aGenerator "Answer a random element of the receiver. Uses aGenerator which should be kept by the user in a variable and used every time. Use this instead of #atRandom for better uniformity of random numbers because only you use the generator. Causes an error if self has no elements." | ind | self emptyCheck. ind _ aGenerator nextInt: array size. ind to: array size do:[:i| (array at: i) == nil ifFalse:[^array at: i]]. 1 to: ind do:[:i| (array at: i) == nil ifFalse:[^array at: i]]. self errorEmptyCollection. To atRandom: aGenerator "Answer a random element of the receiver. Uses aGenerator which should be kept by the user in a variable and used every time. Use this instead of #atRandom for better uniformity of random numbers because only you use the generator. Causes an error if self has no elements." | ind entry | entry _ nil. [entry == nil] whileTrue: [ self emptyCheck. ind _ aGenerator nextInt: array size. entry _ array at: ind. ]. ^ array at: entry. .... Note I put the emptyCheck in the loop just in case there's a possiblilty of another thread changing the object while it's being scanned.. A somewhat faster alternative in the bad case where you have a large array with only a few entries is to do a few checks at random and then to settle down to scanning the rest with a randomly chosen, relatively-prime span size. For that you need something like: atRandom: aGenerator | ind entry i span foundrelativeprime| self emptyCheck. entry _ nil. i_1; "try 10 times with random samples" [ entry == nil and: [i<10] ] whileTrue: [ ind _ aGenerator nextInt: array size. entry _ array at: ind. i _ i+1; ]. entry == nil ifTrue: [ "failed 10 random samples - the array must be almost empty" "pick a random, relatively prime span and scan every entry foundrelativeprime _ false. [ foundrelativeprime] whileFalse: [ span _ 1 + (aGenerator nextInt: ((array size)-1)). foundrelativeprime _ (1 = ((array size) gcd: span)). ]. i _ array size; [i > 1] whileTrue: [ ind _ (ind+span) \\ ((array size). entry _ array at: ind. entry == nil ifFalse:[^entry]]. i _ i-1. ] self errorEmptyCollection. ] ^entry.

1 0

Re: [ANN] eCompletion - code completion inspired by eclipse
by Ruben Bakker 29 Nov '04

29 Nov '04

Hi Stef, >> The instance variable types are only detected in the Debugger, as a >> context is provided. In the case of Browser, Workspace etc. only "type >> suggesting" names do the job. But for instance variables this is not >> something people use. >strange because each time you send a message to an instance variable you >would like to have a smart completion too. Your right of course. What I was trying to say, is that people do not choose the name of instance variables to be "type suggesting". >For local variables I guess that a good heuristics to find the class is >to look >for what is in front of a new method :). Are you doing that? Yes, If the type can't be found out I try two things. Examples: 1) a variable named "rectange" is made uppercase to "Rectangle" 2) or the first uppercase letter is searched. So aRectangle, asRectangle, myRectangle and so on would result in "Rectangle" again. For Debugger and Workspace the type guessing is more accurate, as the context or the bound variables can be accessed. >on cincom store you have a package called roel typer ro something like >that. All the code is available (ie one class and its tests :)) Ok, I'll have a look at it, thanks for the tip. Ruben On Monday, November 29, 2004, at 08:35AM, stéphane ducasse <ducasse(a)iam.unibe.ch> wrote: >hi ruben > >I hope that we will have soon the simple typer of roel so we could get >already some >nice type information for instance variables. > >> Stef, Thanks for your feedback. >> >> The instance variable types are only detected in the Debugger, as a >> context is provided. In the case of Browser, Workspace etc. only "type >> suggesting" names do the job. But for instance variables this is not >> something people use. > >strange because each time you send a message to an instance variable you >would like to have a smart completion too. >For local variables I guess that a good heuristics to find the class is >to look >for what is in front of a new method :). Are you doing that? > >> I planed to add the guessing of instance variables types for the next >> version in the following way: Analyze the initialize method with the >> RBParser. You see a solution in analyzing the bytecode, is this what >> is working in Dolphin Completion? > >I do not know. roel coded what he did in SOUL (kind of prolog) >and hard-coded it in VisualWorks and his demoe last week was nice >so I want to have that in squeak. > >> If you pass some pointers, I would very happy to try it out and >> eventually integrate it in eCompletion. > >on cincom store you have a package called roel typer ro something like >that. All the code is available (ie one class and its tests :)) >> >> Ruben >> >> >> >> On Sun, 28 Nov 2004 22:19:28 +0100, you wrote: >> >>> this sounds quite cool >>> Now I would like to know how do you sort the symbols. >>> (I know that romain completion package is smarter that the default >>> alt-q since he is paying attention to instance variables and also >>> methods defined in the class (if I'm not wrong). >>> >>> Now roel showed use a simple bytecode analyser that guess quite well >>> the type of instance variables and this is >>> on my to do list to port it to Squeak. ) >>> >>> So can you tell us more. >>> >>> Stef >>> >>> >>> >>> On 28 nov. 04, at 21:32, Ruben Bakker wrote: >>> >>>> >>>> >>>> The package is available on the SqueakMap with the name eCompletion. >>>> More information can be found on this page: >>>> http://homepage.mac.com/monique_bakker/squeak/eCompletion.html. >>>> >>>> Have fun... >>>> Ruben >>>> >>>> >>> >> >> >> > > > >

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
38
Older →

Jump to page: