Squeak-dev March 1999

squeak-dev@lists.squeakfoundation.org

153 participants
338 discussions

Re: Anyone making end-user programs with Squeak?
by Chris Reuter 04 Mar '99

04 Mar '99

On Thu, Mar 04, 1999 at 02:48:38PM -0600, William Barnett-Lewis wrote: > Chris, > > Unfortunately, I get an invalid header error from Winzip as well. > That's odd. I tried downloading the archive just now from a different site using lynx and it works fine, both under tar/gzip and WinZip. BTW, the URL I gave out references a page referencing the archive, not the archive itself. The archive can be be found at: http://www.csclub.uwaterloo.ca/~cgreuter/tower002.tgz If these problems persist, I'll try using zip instead. --Chris

1 0

Re: Size of changes file
by Bob Arning 04 Mar '99

04 Mar '99

On Sun, 28 Feb 1999 22:52:04 +0100 "Reinier van Loon" <R.L.J.M.W.van.Loon(a)inter.nl.net> wrote: >could it be that the changes file is constrained to 16M? It is. We had a discussion some time ago about options for increasing it (there were a couple of bits that could be used to increase the limit), but I think no resolution was reached. The trick, for now, is to compress your change file before it gets that big (unless, of course, you have 18M of unique changes). Cheers, Bob

6 8

Re: entao
by Lex Spoon 04 Mar '99

04 Mar '99

David Farber <dfarber(a)numenor.com> wrote: > > thoughts: > i'm certain i've seen both of these questions discussed if not answered on > the squeak list. the problem is that so much info flows through the list > that it takes quite a bit of time to find even messages that i've posted, > much less information that i only vaguely recall seeing. obviously, this is > where the swiki pages would be ideal--they are the perfect place for > information to accumulate. unfortunately (and this is just my impression, i > could be off here) it seems to me that the swiki pages aren't really living > up to their potential. i think the problem is very much a social one; the > technology is there, but it really hasn't been integrated into our > community. is anyone else feeling the same way? any ideas or thoughts on how > we might change this? > Okay, I just added the CrLfFileStream question to the FAQ. I'm not sure how to phrase the "question" part, but I took a stab at it. Anyway, maybe we should expand the "questioner fills in the FAQ policy" by making a habit of politely asking people to do so whenever someone responds to something that should be an FAQ. I know I did this when answering someone's question about proxy Web servers, and they didn't seem to mind at all doing the legwork. Lex

1 0

LinuxPPC Squeak
by Richard L. Peskin 04 Mar '99

04 Mar '99

Has anyone compiled the Linux 2.3 Squeak sources for LinuxPPC (4.1)? --dick peskin ================================= R. L. Peskin, Rutgers Univ. ; <peskin(a)caip.rutgers.edu>; <http://www.caip.rutgers.edu/~peskin> VT Phone (802) 824-4558 NJ Phone (732) 445-4208 "The corporate culture is concerned less with Occam's razor than his aftershave lotion."

1 0

Re: Linux 2.3 problems
by Richard L. Peskin 04 Mar '99

04 Mar '99

At 4:58 PM +0100 3/4/99, Ian Piumarta wrote: >If someone who has been having problems could download one of these >and see if the situation is improved, I'd be very grateful. FWIW: the >white outline on both regular cursors and CursorWithMasks works fine >for me at all bit depths on a variety of different platforms. I have downloaded the new Linux-i386 binary and it soes solve the curser problem; everything works fine. The cursor problem was strange, at least to me. It only occured on my Linux console, not when running Squeak from a remote X server (in my case XTen on the Mac). Anyone know why this might have been the case? (BTW, does this new version change the default memory?) --dick peskin ================================= R. L. Peskin, Rutgers Univ. ; <peskin(a)caip.rutgers.edu>; <http://www.caip.rutgers.edu/~peskin> VT Phone (802) 824-4558 NJ Phone (732) 445-4208 "The corporate culture is concerned less with Occam's razor than his aftershave lotion."

1 0

Launching Squeak Pages from a Web Browser
by Mark Guzdial 04 Mar '99

04 Mar '99

Thanks to a comment from Tim Rowledge on the PWS list referencing SystemDictionary>>readDocumentFile, I've got SqueakPages books opening from Internet Explorer. The code to do this is at http://guzdial.cc.gatech.edu/st/booksandWeb.cs This changeSet includes the code to save a book to an uploadSwiki (such as http://minnow.cc.gatech.edu/squeakPages.1) and the mods to Scamper to read a book that ends in .bo. To make clicking on a .bo URL in IE (or Navigator) open up the book in Squeak: - File in the booksandWeb.cs changeSet into your image. - Save your image as "squeak" (so that you get squeak.image and squeak.changes created) - Quit Squeak. - Open up your Web browser. Modify your "file helpers" (in IE, there is a Preferences item in the Edit menu.) - Create a new helper. I named mine "application/x-squeak-book". The extension must be ".bo". Set the application helper to your Squeak VM. (For this to work, your squeak.image must be in the same directory as your Squeak VM.) Now you're set! Go to the SqueakPages Swiki and click on some .bo link. The .bo file should download to your disk, and Squeak should open up, then load up the .bo file as a book. I found that the .bo file contains the URLs for its pages (property allTextURLs), so I read the original book URL out of that, and it seems to be working. Mark -------------------------- 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

2 1

Re: Linux 2.3 problems
by Lex Spoon 04 Mar '99

04 Mar '99

Bert Freudenberg <bert(a)isgnw.CS.Uni-Magdeburg.De> wrote: > On Wed, 3 Mar 1999, Marcus Denker wrote: > > > On Wed, Mar 03, 1999 at 03:50:21PM -0500, Richard L. Peskin wrote: > > > 1. The "Normal" cursor, the scroll bar between browser panes, the menu icon > > > (scroll region), adn other widgets are not clear. That is, details are not > > > showing up. For example if you move the cursor from the X window background > > > into Squeak, the "arrow" details get fuzzy. This occurs with X running on > > > the system console, but not with a remote X server. > > > > > Bert Freudenberg posted a fix to the list on Tue, 2 Feb 1999: > > > > |Subject: [BUG,FIX] CursorWithMask on X > > | > > |I really wanted to have that nice CursorWithMask but it showed only as a > > |black blob - stColorBlack and stColorWhite were not initialized with RGB > > |triplets (which are needed for 16/32 bpp visuals). A diff is attached. > > I'm not quite sure the symptoms described are the ones this fix is for > (it only sets the mouse pointer's color). My guess is that the cursor's > bits are reversed. In sqXWindow.c, in function ioSetCursorWithMask(), > there is a loop reversing bits: > > /* if (BitmapBitOrder(stDisplay) == LSBFirst)*/ > { > /* the bytes are always in the right order: swap only bits > within bytes */ > char *dp= (char *)data; > char *mp= (char *)mask; > for (i= 0; i < 32; i++) > { > dp[i]= swapBits(dp[i]); > mp[i]= swapBits(mp[i]); > } > } > To add more fuel to the fire, my Squeak cursor looks great (at least, most of them do). It has the desired white outline and everything. My X display is 16-bit, if that makes any difference. I'm not sure of the history of the sqXWindow.c file I'm using; I'll append the ioSetCursorWithMask and swapBits functions from it if anyone wants to try and compare. Lex #ifndef HEADLESS static unsigned char swapBits(unsigned char in) { unsigned char out= 0; int i; for (i= 0; i < 8; i++) { out= (out << 1) + (in & 1); in >>= 1; } return out; } #endif int ioSetCursorWithMask(int cursorBitsIndex, int cursorMaskIndex, int offsetX, int offsetY) { #ifndef HEADLESS unsigned char data[32], mask[32]; /* cursors are always 16x16 */ int i; Cursor cursor; Pixmap dataPixmap, maskPixmap; /* Use bits as mask if not specified */ if (cursorMaskIndex == null) cursorMaskIndex = cursorBitsIndex; for (i= 0; i < 16; i++) { data[i*2+0]= ((unsigned)checkedLongAt(cursorBitsIndex + (4 * i)) >> 24) & 0xFF; data[i*2+1]= ((unsigned)checkedLongAt(cursorBitsIndex + (4 * i)) >> 16) & 0xFF; mask[i*2+0]= ((unsigned)checkedLongAt(cursorMaskIndex + (4 * i)) >> 24) & 0xFF; mask[i*2+1]= ((unsigned)checkedLongAt(cursorMaskIndex + (4 * i)) >> 16) & 0xFF; } /* if (BitmapBitOrder(stDisplay) == LSBFirst)*/ { /* the bytes are always in the right order: swap only bits within bytes */ char *dp= (char *)data; char *mp= (char *)mask; for (i= 0; i < 32; i++) { dp[i]= swapBits(dp[i]); mp[i]= swapBits(mp[i]); } } dataPixmap= XCreateBitmapFromData(stDisplay, DefaultRootWindow(stDisplay), (char *)data, 16, 16); maskPixmap= XCreateBitmapFromData(stDisplay, DefaultRootWindow(stDisplay), (char *)mask, 16, 16); cursor= XCreatePixmapCursor(stDisplay, dataPixmap, maskPixmap, &stColorBlack, &stColorWhite, -offsetX, -offsetY); XFreePixmap(stDisplay, dataPixmap); XFreePixmap(stDisplay, maskPixmap); if (cursor != None) XDefineCursor(stDisplay, stWindow, cursor); #endif return 0; }

1 0

[FIX] BalloonMorph|#string:for:corner:
by Leandro Caniglia 04 Mar '99

04 Mar '99

On some circumstances (when aa > maxArea is false) the instance creation method of the subject fails because some temporaries (verts, mp) are nil. Here is a simple fix. Leandro Content-Type: application/octet-stream; name="BalloonMorph-stringforcorner.st" Content-Disposition: attachment; filename="BalloonMorph-stringforcorner.st" Attachment converted: Anon:BalloonMorph-stringforcorner.st (????/----) (000074E2)

1 0

re: exception handling, differences from ANSI
by Craig Latta 04 Mar '99

04 Mar '99

> I haven't seen the ANSI exceptions standard. How does your approach > differ? The ANSI standard exposes the situations in the programming interface. It requires a separate class for each situation, which I think is overkill. My approach uses instances of a single class to model situations, and refers to them in the programming interface only by name. The message interface is oriented toward blocks for handler setup (>>valueHandling:, etc.), and toward the current context for handling requests (>>handle:, etc.). In all, I think my approach is lighter and easier to use. thanks, -C -- Craig Latta composer and computer scientist craig.latta(a)netjam.org www.netjam.org latta(a)interval.com Smalltalkers do: [:it | All with: Class, (And love: it)]

2 1

[fixed FIX] Morphic Restore Display
by Bert Freudenberg 04 Mar '99

04 Mar '99

On Thu, 4 Mar 1999, Bert Freudenberg wrote: > "Change Set: RestoreDisplay-bf > Date: 4 March 1999 > Author: Bert Freudenberg > > "The Restore display menu item in Morphic did not update the screen size. > Fixed by sending restoreDisplay to a ScreenController even when in > Morphic." Ouch! How did this double quote get into the comment? You can't file in that. Next time I'll check before sending, promised. Attached again. /bert -- Bert Freudenberg Department of Simulation and mailto:bert@isg.cs.uni-magdeburg.de Computer Graphics http://isgwww.cs.uni-magdeburg.de/isg/bert.html Univ. of Magdeburg Content-Type: TEXT/PLAIN; charset=US-ASCII; name="RestoreDisplay-bf.cs" Content-ID: <Pine.LNX.3.96.990304093001.22895C(a)balloon.cs.uni-magdeburg.de> Content-Description: 'From Squeak 2.3 of January 14, 1999 on 4 March 1999 at 9:09:39 am'! "Change Set: RestoreDisplay-bf Date: 4 March 1999 Author: Bert Freudenberg The Restore display menu item in Morphic did not update the screen size. Fixed by sending restoreDisplay to a ScreenController even when in Morphic."! !HandMorph methodsFor: 'world menu' stamp: 'bf 3/4/1999 09:03'! buildWorldMenu "Build the meta menu for the world." | menu | menu _ MenuMorph new defaultTarget: self. menu addStayUpItem. menu add: 'previous project' target: owner action: #goBack. menu add: 'jump to project...' action: #jumpToProject. menu add: 'restore display' target: ScreenController new action: #restoreDisplay. menu addLine. menu add: 'new morph...' action: #newMorph. menu add: 'authoring tools...' action: #scriptingDo. menu add: 'remote...' action: #remoteDo. menu add: 'debug...' action: #debugDo. menu addLine. menu add: 'open...' action: #openWindow. menu add: 'windows...' action: #windowsDo. menu add: 'changes...' action: #changesDo. menu add: 'help...' action: #helpDo. menu add: 'do...' target: Utilities action: #offerCommonRequests. menu addLine. menu add: 'save' action: #saveSession. menu add: 'save as...' action: #saveAs. menu add: 'save and quit' action: #saveAndQuit. menu add: 'quit' action: #quitSession. ^ menu! !

1 0

Jump to page: