Squeak-dev June 2000

squeak-dev@lists.squeakfoundation.org

134 participants
686 discussions

[ENH] No contentsOfEntireFile
by Andres Valloud 02 Jun '00

02 Jun '00

Hi. This changeset removes the above mentioned method, fixes all senders, fixes the implementors and the MethodFinder. Andres.

1 0

Re: [ENH] Proposal to drop contentsOfEntireFile
by Andres Valloud 02 Jun '00

02 Jun '00

Hi Andres. > Proposed solution: rename the method contentsOfEntireFile in FileStream > to contents, and fix its 48 senders. > I can do this in a while. But... do you think it's ok? Ahhh... I think it's a good idea. But don't forget to fix the other 3 implementors, ok? Andres.

1 0

RE: Windows CE and minimal image
by D. Whitehorn-Umphres 02 Jun '00

02 Jun '00

I have had the same problem getting Squeak running on a Vadem Clio w/32MB RAM and a 64MB FlashRam card. In my case, I have all of Squeak in "\Storage Card1\Squeak\". Putting everything on the main memory (root) drive works fine as mentioned before. I know very little yet about Squeak but have been able to determine (I think) that the attempt to load the .image file when it is anyplace *other than* the root dir of the main memory results in a "\S" (literally, I'm not using C conventions) being prepended to the filename of the image file. Now I just need to learn enough to *fix* it. -D Whitehorn-Umphres

1 0

[BUG][FIX] Accept lowercase hex in URL escapes
by Duane Maxwell 02 Jun '00

02 Jun '00

'From Squeak2.8alpha of 26 April 2000 [latest update: #2210] on 1 June 2000 at 6:33:03 pm'! "Change Set: StringUnescape Date: 1 June 2000 Author: Duane Maxwell Lower case hex digits in percent escapes are legal in URLs. This changeset coerces escapes to uppercase before sending them off to Integer>>readFrom:base"! !String methodsFor: 'internet' stamp: 'DSM 6/1/2000 18:15'! unescapePercents "change each %XY substring to the character with ASCII value XY in hex. This is the opposite of #encodeForHTTP" | inStream ans c tok asciiVal | inStream _ ReadStream on: self. ans _ WriteStream on: String new. [ c _ inStream next. c = nil ] whileFalse: [ c = $+ ifTrue: [ ans nextPut: $ ] ifFalse: [ (c = $%) ifTrue: [ tok _ (inStream next: 2) asUppercase. asciiVal _ Integer readFrom: (ReadStream on: tok) base: 16. ans nextPut: (Character value: asciiVal). ] ifFalse: [ ans nextPut: c ] ] ]. ^ans contents ! !

1 0

Re: [ENH] More Faster Senders
by Lex Spoon 02 Jun '00

02 Jun '00

Bert Freudenberg <bert(a)isgnw.CS.Uni-Magdeburg.De> wrote: > On Thu, 1 Jun 2000, Andres Valloud wrote: > > > Hey! I never meant NOT to use attachments!!! This Netscape shows me my > > own attachments attached... sorry about that!!! > > It's almost a correct attachment. It is a separate MIME entity, but marked > as inline rather than attachment. The sqfixes archive gets it right. > Probably Stefan's Eudora just displays the inlines inline and doesn't > offer to save :-) > Celeste has this general problem, as well, and fixing up a good UI for this is a little tricky. After all, you want this thing to be fairly easy to use! -Lex

1 0

Re: eden
by JArchibald＠aol.com 02 Jun '00

02 Jun '00

=> 6/1/00 5:55:33 PM EDT, tim(a)sumeru.stanford.edu => << Not if you've ever used an Apple. >> If we're going to type this properly [and folk-lore-ickly], Tim, I think we're going to have to insist that the [verb] use of the Apple _must_ be 'to eat' the Apple. Cheers, Jerry. ____________________________ Jerry L. Archibald systemObjectivesIncorporated ____________________________ "If you're not scared, you're not pushing your limit, no matter what you may think." Observation of rock and snow climbing leaders. :-o

1 0

Re: Feature request :-)
by JArchibald＠aol.com 01 Jun '00

01 Jun '00

=> 6/1/00 4:45:38 PM EDT, sma(a)3plus4.de => << ...come on, kiss the frog! >> [I _must_ paraphrase!] ...come on, kiss the C (owhy)! Jerry. ____________________________ Jerry L. Archibald systemObjectivesIncorporated ____________________________

1 0

RE: Large array slows down Squeak
by Parker, Mike 01 Jun '00

01 Jun '00

> => 6/1/00 4:54:04 PM EDT, mparker(a)CaseServices.com => > << The memory system reserves 4000 words at the end of eden >> > > What is 'eden'? Sorry. Generational collectors generally segregate objects into two different areas of memory. All new objects are created in a special area, commonly called either "newspace" or "eden", that is the focus of ordinary (minor) collections. If they survive long enough they get moved into another area commonly called "oldspace", that is collected much more infrequently (like only at midnight, shutdown, or some other time when collection overhead isn't too noticeable). Which objects get promoted to oldspace (or "tenured") is determined by the tenuring policy. Some collectors have multiple oldspaces and unintelligent tenuring policies (Lisp machines and the Tektronix Smalltalk systems had five to seven). Most modern collectors seem to use only two, but have much more sophisticated tenuring (and detenuring) policies than early generational collectors. Squeak uses only two spaces that are contiguous with each other, so that the end of oldspace is the start of eden. After each minor collection, all objects are packed into the front of eden, which is in turn butted against the end of oldspace. In the current collector all of eden is tenured at once by the simple expedient of moving the "start-of-eden" pointer to the freemem ptr (i.e. the end of eden ptr), and clearing out the rootlist.

1 0

Re: Feature request :-)
by Bob Arning 01 Jun '00

01 Jun '00

On Thu, 01 Jun 2000 19:10:29 +0200 Karl Ramberg <karl.ramberg(a)chello.se> wrote: >I've never ever typed as much as I do now in Squeak and I type >quite slow so I would love to have the tabFinishTheWordThing from >unix in Squeak. (only thing I miss from unix) >In which class should one look to implement such a thing ? Karl, Type the beginning of a selector like openI in a workspace/browser. Then press cmd-q (yes, that doesn't mean quit!) and see if this gets you what you want. Repeating the cmd-q will cycle through possible alternatives. If you want to hack around with it, start in ParagraphEditor>>querySymbol:. Cheers, Bob

1 0

RE: [BUG?EndRequest?DocoFix?] listBuiltinModules
by Raab, Andreas 01 Jun '00

01 Jun '00

Andrew, > The documentation in Smalltalk listBuiltinModules suggests that this > function will list all plugins. The comment in Smalltalk listBuiltinModules says: "Return a list of all builtin modules (e.g., plugins)." What's wrong with this comment?! Being not a native speaker I might miss something here but given the selector and comment never even mention any other than builtin modules I don't see a problem ;-) - Andreas > -----Original Message----- > From: Andrew C. Greenberg [mailto:werdna@gate.net] > Sent: Wednesday, May 31, 2000 7:02 PM > To: squeak(a)cs.uiuc.edu > Cc: recipient list not shown > Subject: [BUG?EndRequest?DocoFix?] listBuiltinModules > > > >This VM... > > IS built from all the latest plugin sources > > IS "fat" (ie it will run on a 68K) > > DOES NOT have the new LargeInteger primitives > > (but it rocks with the plugin!) > > DOES have fast recovery from missing primitives > > (so it still runs fine without the plugin) > > DOES support 'Smalltalk listBuiltinModules' and related > functions > > The documentation in Smalltalk listBuiltinModules suggests that this > function will list all plugins. > > In fact, it lists only the "internal" plugins "built into" the VM, > and NOT subsequently loaded "external plugins." > > The documentation should reflect this fact, so that folks (like > myself) won't chase their tails trying to figure out how to load a > plugin that is already loaded. > > On the other hand, the "loaded" or "loadable" status of an "external" > plugin is a highly desirable piece of information. Presently, the > only real solution is to write some special purpose procedure in the > plugin to that end. > -- > Andrew C. Greenberg acg(a)netwolves.com > V.P. Eng., R&D, 813.885.2779 (office) > Netwolves Corporation 813.885.2380 (facsimile) > www.netwolves.com >

3 2

← Newer
1
...
62
63
64
65
66
67
68
69
Older →

Jump to page: