Squeak-dev May 1999

squeak-dev@lists.squeakfoundation.org

126 participants
289 discussions

RE: Random Numbers
by Steve Wart 02 May '99

02 May '99

Thanks for the pointer. I thought about "stocking up", but I decided not to because I wanted something small that would more or less deal with the numbers "running out". The CORBA interface definitely sounds better than HTTP, but I like the fact that the SGI folks are crypto geeks (meaning the quality of the data is very good). However, I don't like the fact that the seed buffer I get is publicly available for a full minute before the next one comes along. Is anyone doing CORBA in Squeak? I used the Url class to read the HTTP stream; originally I wanted a Random class in GemStone, but there is nothing there except basic socket support. Steve Patrick Logan wrote: > > There are a few sources of true random numbers on the Internet, so > your system could have some redundancy. See http://www.random.org > > Also at this particular site is a source of random numbers in bulk via > HTTP or CORBA, so you could "stock up" on random numbers and use them > as you need them. >

2 1

Re: JPEGReadWriter, Differences between Squeak and VisualWorks
by Tim Olson 02 May '99

02 May '99

>Hi. > >> > These selectors, #>> and #<<, should be deleted from Squeak... >> I put these methods into Squeak > >Ouch!!!... :) oops, sorry! Hi, Andres: >Hmmmmm... is the JPEG reader converted to C too? Besides >> and <<, today I >found a method in the JPEG reader with something like > >(aNumber to: someOther by: yetAnother) do: aBlock > >... errr... now I don't know what to say! :)))... Yes, there still were a few uses of #<< and #>> still laying around in JPEGReadWriter; I've cleaned them up in this changeSet which improves performance by about 4%. Unfortunately, the Interval>>to:by:do: use that you point out in JPEGReadWriter>>idctBlockInt:component: is there because that method is on the edge of not compiling -- the required frame size is close to exceeding the maximum of 32 for a large frame. Changing it to use the compiler in-lined to:by:do: increases the requirements to over that limit. -- tim Attachment converted: Anon:JPEGReadWriter...s.2May949am.cs (TEXT/R*ch) (000098CE)

1 0

Question: Block and ST-76
by Marcel Weiher 02 May '99

02 May '99

Hi folks, just finished reading the 'Early History of Smalltalk' paper in HOPL-II a couple of weeks ago, and now I have some question regarding blocks and open-color keywords. >From the short snippets of code in the paper, it looks like the square-bracket/block notation was used generally in Smalltalk-76 to delimit chunks of code, instead of only for delayed evaluation as it is now. Closed-colon keywords evaluated the expression (be it a block or something else) immediately (pass-by-value), open-colon keywords simply passed the expression itself (generalized pass-by-name). Is this correct? It also seems that some of the problems described in 'Smalltalk-80: Bits of History, Words of Advice' are actually part of the bigger problem of dsitinguishing between names/references and the objects they refer to when both are first class object, the same going for proxies, collections etc. Have there been any attempts at a general solution? Marcel

1 0

Re: JPEGReadWriter, Differences between Squeak and VisualWorks
by Andres Valloud 02 May '99

02 May '99

Hi. > <<, >> et al. are done with #bitShift: and a positive or negative value, > depending on the direction These selectors, #>> and #<<, should be deleted from Squeak... or left but the compiler instructed to use a proper #bitShift: instead... Profiler measure: [4 bitShift: 2] 1mus 739ns Profiler measure: [4 << 2] 3mus 17ns Andres.

2 2

[PWS] Just a thank you
by Reinier van Loon 02 May '99

02 May '99

I just want to express my sincere thanks to everyone who made PWS to what it is today. Why? I just spent 5 hours with PWS and have accomplished the following: 1) a consistent look and feel of a web site (because pages are generated); 2) a way to maintain various data (members, addresses) in a very easy way (decoupling URL and action). Thanks again. Reinier.

1 0

2.4b image/changes mismatch
by Hans-Martin Mosner 01 May '99

01 May '99

Hi, I just downloaded 2.4b from the UIUC file server, and the image and changes file don't seem to correspond. I remember that there was some discussion on the list and that 2.4b was supposed to have fixed that... So, it is not fixed. Just open a fresh 2.4b and look at StringMorph>>drawOn: to see what I mean. Hans-Martin

2 1

Re: [Q] Sound on Linux 2.0.30 + libc5
by akubi 01 May '99

01 May '99

Thank you, Ninja. ninja wrote: >Before change I've used emulate app icon option >in window maker. This prevent me from using >sound in squeak and I do not know why. >But after turning off this option and is the default case, >I can use sound ability in squeak. I checked with my old kernels (2.1.130 & 2.2.1 ). Unfortunately, such errors as "Finding dependencies /lib/modules/2.2.1/misc/sound.o unresolved symbol(s)" was reported on booting, and I could not use OSS. c.f. cat /dev/sndstat PowerMac DMA sound driver: sound.format = 0x8 (unsigned 8 bit) sound.speed = 8000Hz (phys. 8820Hz) sound.stereo = 0x0 (mono) sq.block_size =32768 sq.max_count = 4 sq.max_active = 4 sq.count = 0 sq.rear_size = 0 sq.playing =0 sq.syncing = 0 p.s. When I compiled "OSS Sound support" with "y" (i.e. not as a loadable module), cat /dev/sndstat reported OSS/Free:3.8s2++-971130Load type: Driver compiled into kernelKernel: Linux localhost.localdomain 2.2.1 #6 Fri Apr 30 16:56:37 JST 1999 ppcConfig options:0Installed drivers: Type 36: SoftOSS Virtual Wave TableCard config: SoftOSS Virtual Wave TableAudio devices:Synth devices:0: SoftOSSMidi devices:Timers:0: Systemclock1: SoftOSSMixers: However, "AbstractSound stereoBachFugue play." reported ioctl(SNDCTL_DSP_GETOSPACE): Invalid argument on the xterm. p.p.s. When I compile either with "PowerMac DMA Sound" as a loadable module, or without it, cat /dev/sndstat reported OSS/Free:3.8s2++-971130Load type: Driver compiled into kernelKernel: Linux localhost.localdomain 2.2.1 #7 Fri Apr 30 17:17:28 JST 1999 ppcConfig options:0Installed drivers: Card config: Audio devices:Synth devices:Midi devices:Timers:0:System clockMixers: but "AbstractSound stereoBachFugue play." reported /dev/dsp: No such device on the xterm. So, I 'm wating for new release of LinuxPPC (R5). Thank you very much, Ninja. -- Kazuya SAKURADA akubi<a98s111(a)ex.media.osaka-cu.ac.jp>

1 0

[ENH] Singleton class
by Norton, Chris 01 May '99

01 May '99

Hi Folks. I needed an object with only one instance so I whipped up the ol' Singleton class. It's nothing special, but I thought you might like it too, so here it is... ---==> Chris PS> [For Newbies...] Singletons are classes that can only have a single instance of themselves. This is a fairly common situation in software engineering. I often use singletons to represent the single doorway to my database (e.g. DBManager). To make use of Singleton, simply create a subclass of it. You can access the single instance of your new class by calling its class method "current" (Singleton will not allow you to create new instances of itself via Singleton class>>new). For example: DBManager current saveSomeStuff: stuffToSave. Enjoy! <<Singleton.cs>> Content-Type: application/octet-stream; name="Singleton.cs" Content-Disposition: attachment; filename="Singleton.cs" Attachment converted: Anon:Singleton.cs (????/----) (000098C8)

2 1

RE: [ENH] Singleton class
by Norton, Chris 01 May '99

01 May '99

shaping(a)bigfoot.com [SMTP:shaping@bigfoot.com] said: >How about class method #only? Hi Shaping. The "Design Patterns Smalltalk Companion" book's example implementation of Singleton uses Singleton class>>current which is the same protocol that I've also used for years to represent the single instance. ---==> Chris PS> More on "Design Patterns Smalltalk Comparison" can be found at: http://www.awl.com/cseng/swpatterns

1 0

Jump to page: