Squeak-dev March 1999

squeak-dev@lists.squeakfoundation.org

153 participants
338 discussions

Re: HP Journada?
by Markus Kohler 24 Mar '99

24 Mar '99

Hi, Just a note, there are already three Jornada devices. See http://www.hp.com/jornada/ for more information. The one you all mean is the 820 a StrongArm based subnotebook sized devive. There's also the 420 palm size device and the 680 a device that looks a bit like a Psion 5. I'm probably going yo get the 420 ... Markus -- Markus Kohler mailto:markus_kohler@hp.com

1 0

PDAgui
by Alexander Peuchert 24 Mar '99

24 Mar '99

Hi out there, maybe this question has been asked before, but ... I'm developing a GUI for the PDA ports of Linux right now( actually I started it two weeks ago). Now someone posted your Squeak as a good and stable alternativ. I'd also like to use something complete and stable, but have some questions about it ... - how big is a minimal Squeak system with gui support ? - how FAST is it, compared to a system written entirly in C ? - as Squeak being a object-oriented system it would be easy to implement some apps like a calendar or a web browser above it. But it would be better if somebody already had done it ... ? - as PDA don't have a big display things like overlapping windows and window handles are not needed( but could be supported for later use), is the Squeak windowing system ready for that ? I'm looking forward to recieving your answers. alex P.S.: Now that Linux is reaching to the PDA territory Squeak could help it get it usable on that system for end-users. And maybe this could fight WinCE like Linux does on servers right now ... ( maybe, in the far future, .... ) Alexander Peuchert mailto:peuc@comnets.rwth-aachen.de

1 0

problem with UIUC Download directory
by mdr＠scn.org 24 Mar '99

24 Mar '99

When I try to open ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/win32/ from the Squeak.org home page, I get the message FTP Error Could not login to FTP server User anonymous access denied. Could somebody please take a look at this. Thanks! Mike p.s. On an unrelated note, I notice that starting Squeak on my PC is significantly(!) faster than starting any of my other core applications (Word, Excel, Powerpoint, Netscape Browser/EMail). Squeak will be up and running before I even get a spash screen from some of the others.

1 0

Class change records
by Tim Rowledge 24 Mar '99

24 Mar '99

Does anyone else find it annoying that when you add an instvar, the entire class defn is included in the changeset? It means that such changes become utterly order and version dependant, even if it doesn't really need to be. For example, I've been doing new compiled method stuff and input event stuff and primitive error returning stuff. The instvar additions are (almost) entirely independant of each other and would be best recorded in changeset files as Interpreter addInstVarName: 'fooble' instead of the whole def. Has anyone either done work to make this happen, or been fired by enthusiasm to try it? I'm bored with hand editting files to try and keep'em clean. Oh yes, it ought to be a bit clever with multiple additions in one file. Nothing like recompiling ObjectMemory four times over to remind you of _that_ little optimisation! tim -- Useful Latin Phrases:- Furnulum pani nolo. I don't want a toaster. Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim>

1 0

[ENH] Saving image to some other directory
by Tim Rowledge 24 Mar '99

24 Mar '99

Some time ago we had a discussion about using default images when people (I think mainly for unix users) try to start squeak without specifying an image name. Various suggestions were made about how to provide the default image, but one of the problems seemed to be the saving of the image in the users' home (or wherever) directory. Typically we wouldn't want to have them save their 'used' image over, or even near, the 'virgin' image. One option would be to copy the default image and changelog to their default directory and then start it. Another would be to make it possible to save the image using not just a new name, but a new directory. This small changeset allows just that. The dialogue has the full path of the image name, and you can edit it to any reasonable path for your system. I think that we now need to make the image detect that it is a virgin (no rude jokes please - perhaps it could check for write permission on the changelog as a test) and if so, alert the user that the image should be saved to their local directory. Thus 'squeak' would run, discover no specified image name, run the default and immediately get the user to saveAs in their local place. On Acorn, Mac and Windows, the virgin image/changelog and sources could feasibly be bound to the VM, making it seem like a single file to end users. I dare say some equivalent trick can be done on unix. Of course, if only we all had gigabit/s net links, the default files could be off on a server... tim -- Strange OpCodes: FLR: Flash Lights Randomly Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim> 'From Squeak 2.0 of May 22, 1998 on 7 June 1998 at 10:20:51 pm'! "Change Set: saveElsewhere Date: 7 June 1998 Author: Tim Rowledge <tim(a)sumeru.stanford.edu> Make it possible to save your image/changes into a different directory than the current one - the entire image name is presented in the save as dialogue and you can change any or all parts of the pathname. This would be particularly useful for saving a user image into the users directoy after startinga default image from some default directory"! !FileDirectory methodsFor: 'file operations' stamp: 'TPR 6/7/1998 21:39'! copyFileNamed: fileName1 toFileNamed: fileName2 "Copy the contents of the existing file with the first name into a new file with the second name. Both files are assumed to be in this directory." "FileDirectory default copyFileNamed: 'todo.txt' toFileNamed: 'todocopy.txt'" | file1 file2 | file1 _ self readOnlyFileNamed: fileName1. file2 _ self newFileNamed: fileName2. file1 copyToFileStream: file2 ! ! !FileStream methodsFor: 'accessing' stamp: 'TPR 6/7/1998 21:37'! copyToFileStream: fileStream2 "Copy my contents into fileStream2" | buffer | buffer _ String new: 50000. [self atEnd] whileFalse: [fileStream2 nextPutAll: (self nextInto: buffer)]. self close. fileStream2 close. ! ! !SystemDictionary methodsFor: 'snapshot and quit' stamp: 'TPR 6/7/1998 22:10'! saveAs | dir newName newImageName newChangesName newDir | dir _ FileDirectory default. newName _ FillInTheBlank request: 'New File Name?' initialAnswer: Smalltalk imageName. newName isEmpty ifTrue:[^self]. newName = Smalltalk imageName ifTrue:[^self snapshot: true andQuit: false]. FileDirectory splitName: newName to:[ :path :fn| newDir _ FileDirectory on: path. newImageName _ (FileDirectory baseNameFor: fn), FileDirectory dot, FileDirectory imageSuffix. newChangesName _ (FileDirectory baseNameFor: fn), FileDirectory dot, FileDirectory changeSuffix]. ((newDir includesKey: newImageName) or: [newDir includesKey: newChangesName]) ifTrue: [ ^ self notify: (newDir fullNameFor: newImageName), ' is already in use. Please choose another name.']. (dir readOnlyFileNamed: self changesName) copyToFileStream: (newDir newFileNamed: newChangesName). self logChange: '----SAVEAS ', (newDir fullNameFor: newImageName), '----', Date dateAndTimeNow printString. self imageName: (newDir fullNameFor: newImageName). LastImageName _ self imageName. self closeSourceFiles; openSourceFiles. "so SNAPSHOT appears in new changes file" self snapshot: true andQuit: false. ! !

1 0

Re: [ENH] That crazy dot idea II
by Bob Arning 24 Mar '99

24 Mar '99

On Tue, 23 Mar 1999 20:38:06 +0100 Stefan Matthias Aust <sma(a)mail.netsurf.de> wrote: >Actually, I'm surpised that everybody seems to dislike the idea. Stefan, I had a reaction not unsimilar to others who responded to your suggestion, but I wanted to spend some time to try to understand why it bothered me so. Part of my reaction was of the "If it aint broke, don't fix it" variety. This has been expressed by others who championed the simplicity of Smalltalk syntax. Smalltalk is _so_ easy to read at the syntactic level that one can proceed to the semantic level in a manner of seconds in most cases. But more than the simple purist's objection, I wondered what practical effect it would have. While you could, and still can, make such changes to your image and distribute those changes to others who appreciate the clarity you seek, what happens when, and if, it enters the mainstream? What would be the next change proposed? Would it be admitted as well? Most of us have programmed in a fair number of languages and one common aspect, I suppose, is the ability to recognize what the language is when glancing at a bit of code. I think what bothered me the most was that some day I might be browsing code in Squeak and come across a method where I couldn't immediately identify the dialect or even the language (I was beginning to think your extensions looked a lot like Object Pascal). The time spent to stop and puzzle out what messages are being sent to what objects is time lost when I what I really need to do is to understand what those messages mean. The current simple and consistent syntax enables me to put my energies where they are most needed. I think it would be more palatable to accept the inclusion of totally different languages in the Squeak environment rather than to have a slow creep of Smalltalk syntax. At least one would have a clear signal that it was time to switch gears (and possibly dig out the manual). I wonder if your changes might meet with greater approval if they were a part of something larger and distinctly different that addressed some set of problems better than the standard Smalltalk. Cheers, Bob

1 0

Re: Dot notation and a crazy idea
by stan＠stanheckman.com 24 Mar '99

24 Mar '99

"Donald T. Major, II" <sasdtm(a)unx.sas.com> writes: > Once upon a time, Actor used to use the same shorthand syntax for > dictionaries, and all indexed instances; instead of just a number, > it could be other objects, including symbols, strings, etc. (literal > or otherwise). It was available for for both a[i] and a[#fooey] := > 'This message brought to you by the letter "N", and the number 4.' > As syntax went, it was actually darned convenient (as Larry Wall > seems to have more or less discovered for Perl, though he actively > differentiates between arrays and "hashes".) Python uses the same. It *is* darned convenient. Compact too. But while "just one more rule" doesn't subtract much from Perl's simplicity, I think Squeak has more simplicity to lose. If we someday dispose of the source file completely, saving only comments and byte codes, then we could experiment with syntax changes like this by just patching the interpreter and the pretty printer. Turn on new syntax. Read existing methods in new syntax. Write new methods in that syntax. Tire of new syntax. Turn off the patches. The great advantage of losing the source file is that we could go back to reading and editing in the old format, with no damage caused by months of work in the alternate syntax, because the bytecodes never changed. Only the interfaces between human beings and bytecodes did. Maybe playing with a new syntax then reverting to an old is already easy? I'm afraid to directly edit the sources file, but maybe there's another way? I think this sort of switchable syntax would largely work for testing the proposed brackets. I suspect it wouldn't work for more complicated changes, because those changes would cause us to write slightly different programs. These would look good when displayed in the syntax we wrote them in, but not so good when displayed in another syntax. Whoever is working on the "one pretty printing to rule them all" probably already knows more about this problem than I ever shall. :-) -- Stan

2 1

Re: Why so few binary method selectors?
by Colin Sarsfield 23 Mar '99

23 Mar '99

>Iannis Zannos wrote: >> >> I wrote: >> > I like it. (Incidentally, I know of only one other language that >> > exploits @ being pronounced as "at".) >> >> Which language do you have in mind? > >Eiffel. Though the usage of @ for indexing is just a minor aspect of it. Actually Pascal employs this @ at as well. Although the meaning is more like "stored at" or "addressed at" (i.e. it returns a pointer to the variable).

1 0

Re: HP Journada?
by Jeffrey Chapman 23 Mar '99

23 Mar '99

> p.s. Is there a set of benchmark results somewhere for various processors? I > am specifically interested in relative performance of Alpha, PC, PowerPC, and > palmtops. The only site I'm aware of with such comparisons is: http://www.urban.ne.jp/home/nishis/SmalltalkAgents/goodies/STones80Resul t.html

1 0

Re: Dot notation and a crazy idea
by Doug Way 23 Mar '99

23 Mar '99

On Fri, 19 Mar 1999, Marcel Weiher wrote: > I also am reminded of C-- etc. when I see all those dots, so are > there ideas for another equivalent path-separator or do we just have > to overcome our collective phobia? (Spaces don't cut it because > then you don't have the close binding of the path expression. Well, > at least my eyes have difficulties picking it out). While I think that Travis Grigg's description of the various Smalltalk symbols on his webpage was excellent, I think it can be argued that the dot "." (period) separator is actually slightly better for namespaces than a slash "/" symbol, for a few reasons: - The dot is vaguely familiar as a separator for several different reasons, whereas the slash separator in a series is really only familiar for one reason, as a pathname separator. I think this might tend to mislead people into thinking that these slashed namespaces were actual file pathnames (although I suppose they would get over it). Dot are used as separators for a variety of different things, not just in C/Java, but also they're familiar from filename extensions (Squeak2.3.image), internet domain name addresses (www.squeak.org), etc. So I don't think there would be an immediate confusion with any particular convention outside of Squeak. - The dot seems like a less "severe" separator than the slash, and thus names including dots read better as a single entity... probably simply because dots are smaller. The filename extension comparison works here too... Squeak2.3.image still feels like a single name, whereas Squeak2/3/image is more broken up into parts. (Or could it be argued that being broken up is better?) It's true that the dot is already used for two other purposes in Smalltalk, as a statement ender and as a decimal point... definitely a reason for caution. But these two other uses are unrelated enough so that at least there wouldn't be confusion between a namespace separator dot and a statement ending dot when looking at actual code. You get into real trouble (with C++ etc.) when a symbol has a few related but still different meanings. > However, the '.' was proposed for another purpose, namespace > support. My idea was that once you have it, why not make it general > and use it to simplify another part of the syntax that can get really > ugly really quickly? Not a bad idea... any syntax change in Smalltalk needs to "carry its weight", so it might as well be as general as possible. Probably an idea to keep on the back burner for awhile (blue plane) before replacing all of the calls to select:/collect: in the Squeak image, though. :-) I've also always thought that the @ sign was underutilized in Smalltalk... changing at: to a binary @ would be great also for the higher precedence of a binary operator. Although replacing at:put: is more of a challenge... I wouldn't go so far as adding new syntax (such as the suggested a[1] := b[3] etc.). - Doug Way EAI/Transom Technogies, Ann Arbor, MI dway(a)eai.com, dway(a)mat.net http://www.transom.com Smalltalk: Guaranteed Y2T Compliant

3 2

← Newer
1
...
5
6
7
8
9
10
11
...
34
Older →

Jump to page: