Squeak-dev April 1998

squeak-dev@lists.squeakfoundation.org

107 participants
279 discussions

Problem with Transcript
by Reinier van Loon 09 Apr '98

09 Apr '98

Folks, has anyone noticed the following: The Transcript is collapsed but it is written to. You have a couple of debuggers. Then opening the Transcript makes Squeak loop. You can break out of the loop with the interrupt key. It turned out that Squeak didn't find the end of run while displaying. Any clues? Reinier. --- Want to express your thoughts? Use Smalltalk. --- Reinier van Loon (rvloon(a)inter.nl.net or reinier(a)2share.com) Squeak Download Area at http://web.inter.nl.net/users/R.L.J.M.W.van.Loon/Squeak/download.html

2 1

Re: printing HTML
by Bruce Boyer 09 Apr '98

09 Apr '98

The key is that HTML is not a page layout spec. Even in real publications tools, there's more than a little difference of opinion as to what's best to use. Some take more of a graphic artist's perspective and choose PageMaker, in which virtually everything is a graphical element to be placed and formatted to be visually appealing. Others are more content or writing oriented, want a decent layout but mostly don't want that to get in the way, and they choose FrameMaker. (Obviously, there are other choices; if you really don't care about layout, you choose Word.) HTML is really a minimal layout control format; it determines document structure only. Yes, the browser can even screw up the clear intention of the structural specification (numbering a <ul>, for instance), but then that browser should fall out of use for making a document really confusing or misleading, or generally useless. More subtle differences, such as how much leading follows a paragraph, can be annoying, especially to a layout oriented person, but could also be annoying to readers of the document, since some choices really can subvert the reading experience (decrease readability). Lex's recommendation to use PS or PDF is the right one if you really want to control layout in a web browser, and use your favorite page layout tool to create it. Bruce At 10:05 PM 4/8/98 -0400, Lex Spoon wrote: >Ted K. writes: > > > One problem with html is that you have no control over horizontal > > spacing. Tabs are not supported. Putting in more than one space > > character does not influence horizontal spacing. > > > > Other than that, html is a 'universal' printing format. I've > > noticed some gif support in the system. I would like to see us > > expand this capability to allow any window to dump an image of > > itself on the disk as a legal html file. An added benefit is that > > one can make web pages by 'printing' and moving the files to your > > web server. > > >I can't resist one of my favorite rants. > >With HTML it's even "worse" than this. Take any particular tag, and >you really don't know just how it will render on a random user's >browser. For instance, header tags might come out numbered, or >bold-faced, or enlarged, or centered, or maybe none of the above. ><ul> might number its "unordered" list, and <ol> might use letters >instead of numbers. You just don't know. > > >If you *really* wants to control layout, you should use something like >Postscript or RTF or TeX or Word, systems seriously designed for >describing the appearance of a document in addition to its textual >content. > >HTML is for *content* markup. > > >Lex > > > ************ Bruce Boyer Lead Technical Writer ObjectShare email: bboyer(a)objectshare.com ************

1 0

Blue plane?
by Reinier van Loon 09 Apr '98

09 Apr '98

Squeak is (amongst other things) meant to be a research tool. What about this crazy idea? To make Smalltalk a truly universal language with equal chances for anyone we might want to let go from English. But then we need a language which is not too difficult to learn. We might try... Esperanto! It has only 16 grammar rules and the rest is just learning words. Would be something different if we would starting programming Smalltalk a la Esperanto. We could even write these e-mails in Esperanto. Probably too hefty ;-). Reinier. --- Want to express your thoughts? Use Smalltalk. --- Reinier van Loon (rvloon(a)inter.nl.net or reinier(a)2share.com) Squeak Download Area at http://web.inter.nl.net/users/R.L.J.M.W.van.Loon/Squeak/download.html

3 3

Thread-Safe FileStreams
by Mark Guzdial 09 Apr '98

09 Apr '98

Over on the PWS mailing list (pws(a)cc.gatech.edu), we've been investigating Squeak's file system under thread access. If process 1 opens the file, then process 2 opens the file, then process 1 closes the file, then process 2 closes the file, what happens? The short answer is that you get a primitive failure, but if you do this long enough and ignore the errors (through exception handling), you end up leaking all the file handles. The code at the bottom of this message, run a couple times, uses up all the filehandles on a PowerMac 6100 running Jitter 1.31 under System 7.6.1. I am interested in what the solution is to this problem, but don't know enough about thread-safe file systems to know what the right level to attack the problem is. Is it: - An OS issue? Can UNIX handle this better than Macs can? - A VM or Image issue? Making file access, even for reads, mutually exclusive seems harsh. Maybe I can detect multiple opens to the same file for reading, then just buffer the reads? - An application issue? Clearly, for multiple writes, it's the application's job to make sure that they are serialized. Maybe the application has to serialize reads, too, since I can't know at open time if the access is read only or read-write. I'd be interested in hearing your suggestions! Mark --- CODE TO LEAK THE FILE HANDLES ---- | aFile | "Put something in the file" aFile _ FileStream fileNamed: 'fred.test'. 1 to: 100 do: [:i | aFile nextPutAll: (i printString); cr.]. aFile close. "Stress test" [1 to: 50 do: [:i | Transcript show: 'Process 1 ',i printString,'-'. [(Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile _ FileStream fileNamed: 'fred.test'. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile contentsOfEntireFile. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile close.] ifError: [:r :c | Transcript show: (c printString), (r printString); cr.]]] fork. [1 to: 50 do: [:j | Transcript show: 'Process 2 ',j printString,'-'. [(Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile _ FileStream fileNamed: 'fred.test'. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile contentsOfEntireFile. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile close.] ifError: [:r :c | Transcript show: (c printString), (r printString); cr.]]] fork. [1 to: 100 do: [:i | Transcript show: 'Process 3 ',i printString,'-'. [(Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile _ FileStream fileNamed: 'fred.test'. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile contentsOfEntireFile. (Delay forMilliseconds: (100 to: 500) atRandom) wait. aFile close.] ifError: [:r :c | Transcript show: (c printString), (r printString); cr.]]] fork. -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 (404) 894-5618 : Fax (404) 894-0673 : guzdial(a)cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html

4 10

Re: Pluggable views [was: Vers. 2.0 ...]
by stp＠limbo.create.ucsb.edu 09 Apr '98

09 Apr '98

For a good intro to pluggable MVC, take a look at the article "A Cookbook for using the Model-View-Controller User Interface Paradigm in Smalltalk-80." (by Glenn Krasner and myself) in Journal of O-O Programming 1(3) 26-49. stp _ Stephen Travis Pope _ Center for Research in Electronic Art Technology (CREATE) _ Dept. of Music, Univ. of California, Santa Barbara (UCSB) _ stp(a)create.ucsb.edu, http://www.create.ucsb.edu/~stp/

1 0

Re: Blue plane? (Magnitudes and units)
by stp＠limbo.create.ucsb.edu 09 Apr '98

09 Apr '98

In reference to Alan K's comment and Augustin Mrazik's response about "units," you might want to take a look at Siren's music magnitude classes (e.g., pitch, duration, loudness with units such as Hz, Sec or dB). There are a few interesting design patterns in there about interoperating with numbers and other magnitudes, mixed-mode arithmetic within a "species" (e.g., 1/4 beat + 30 msec), and the separation of representation (what do I represent? (I call this species)) from implementation (what "units" do I use? (I use classes for this)). There are many kinds of units in the world, but a smaller number of fundamental dimensions, and any model should take into account that the different measures, e.g., of distance or temperature, share some behaviors but also have some subtly different semantics (e.g., linear vs. angular distance). stp _ Stephen Travis Pope _ Center for Research in Electronic Art Technology (CREATE) _ Dept. of Music, Univ. of California, Santa Barbara (UCSB) _ stp(a)create.ucsb.edu, http://www.create.ucsb.edu/~stp/

1 0

Pluggable views [was: Vers. 2.0 ...]
by Dan Ingalls 09 Apr '98

09 Apr '98

jim benson <jb(a)speed.net> wrote... >When you say "pluggable views" what does this mean? Consider ListViews: Instances of the new ListView class have a set of selectors (their "plug") that get performed to read and write the model. Instead of a distinct ListView subclass and ListController subclass for EVERY listPane in the system, there is ONE ListView class (with controller) that gets instantiated with a different set of selectors for each listPane. The same thing is done for TextViews and SwitchViews. After some considerable code relocation (mainly menu actions), we get to throw away 20 or 30 classes, thus substantially simplifying the MVC world. More to the point though, having made this change, we need only to create a handful of compatible ViewMorphs in order to run all the existing MVC models (browser, debugger, fileList, etc) unchanged in a Morphic world. This will allow Squeak to support both UIs with very little duplication of code. The change should facilitate any widget experiments, including native interfaces, due to reducing the number of variant ListViews to one, and likewise for the other common pane types. - Dan

1 0

Re: Squeak needs printing capabilities
by Reinier van Loon 09 Apr '98

09 Apr '98

Lex Spoon wrote: .... > First, it would be nice to have some sort of simplified plain text > printing system, in addition to postscript--this would be faster on > many systems (eg, mine :)) and would be nicer on VM developpers for > wierd platforms. Agree. It needs to be solved at VM level because you don't want to know about system/platform specific details. > > Second, there don't necessarily have to be dedicated primitives for > printing to work. For instance, if there were some sort of system() > and/or popen() escape (as has been proposed previously), then Unix > users could just call lpr. And currently, if one has a network > printer available, one should be able to print via sockets from ANY > platform. But for some configurations (wimpy Windows 95 boxes maybe > :)), there would HAVE to be primitives added. Or, can in fact those > wimpy Windows 95 boxes handle network printing via some sort of > standard protocol? Disagree. It needs to be solved at VM level because you don't want to know about system/platform specific details. > We need at least the feature to send a string to a printer. If there isn't a printer available the output could be dumped into a file (like sqPrint<id>.prt where <id> is date/time of printing and/or a number). Reinier.

3 2

Re: Squeak needs printing capabilities
by Patrick Logan 09 Apr '98

09 Apr '98

> Seriously, I think we need to support printing from the VM (Yippee! > Platform independent printing). Would a connection between Squeak and Ghostscript be a step in the right direction? -- Patrick Logan mailto:patrickl@gemstone.com Voice 503-533-3365 Fax 503-629-8556 Gemstone Systems, Inc http://www.gemstone.com

7 8

Re: Squeak needs printing capabilities
by Reinier van Loon 09 Apr '98

09 Apr '98

> On Wed 08 Apr, Tim Rowledge wrote: > .... > There is no need for _users_ of the facility to have to worry about > the platform details, Agree. > but that does not mean that the implementation has to be buried in the VM. > In much the same way that filenaming > details are handled (albeit poorly at the moment) by having platform > specific subclasses, platform specific printing can be catered for > by having subclasses of some printing-interface class. I am not really a fan of this solution because people tend to use these specific subclasses which does not make code portable to other platforms. > Some parts may > involve platform specific primitives, or better yet, systemcall code > which avoids too much complication of the VM code. There need not be > any troubling image size problem if the code is dynamically loaded. Hmm. This systemcall code might look like (excerpt from printing.cs in KSCSQK.ZIP) Transcript show: 'sending to printer... '. "(ExternalProcess defaultClass forkJob: 'lpr' arguments: (Array with: TempFileName)) wait." Transcript show: ' ...cannot print directly; to really print (on a UNIX system), open a terminal window and type: lpr ' , TempFileName , '.'; cr! ! and like (excerpt from printers.cs in KSCSQK.ZIP) closeSendAndRemove "Close the file, send it to the printer, and remove it." self closeAndSend. self cshPort sendBuffer: ('rm ', self interface name, '\') withCRs.! This contains platformspecific commands like 'lpr' and 'rm'. I'd rather have an abstract interface for printing a document. Maybe the only thing we need at VM level is just 'lpr' (Unix) or 'copy lpt1:' (Windows/DOS) and the corresponding primitive is called 'print'. > > The VM is already _way_ too complicated and convoluted. Really? I think it is not that bad. It is quite clean actually. > -- > Useful Latin Phrases:- Utinam logica falsa tuam philosophiam totam > suffodiant! > May faulty logic undermine your entire philosophy! Was this generated or intentionally put here? I am unsure if I should feel offended now (my wife is good at Latin and when she heard/saw 'suffodiant' she frowned because it was insulting in her opinion). Anyway, it would be nice if we would have some rudimentary form of printing and as long as the solution delivers portable code I'm a fan. Best regards, Reinier.

2 1

← Newer
1
...
21
22
23
24
25
26
27
28
Older →

Jump to page: