Squeak-dev September 2001

squeak-dev@lists.squeakfoundation.org

168 participants
408 discussions

Re: [ENH][Modules] Another version
by danielv＠netvision.net.il 08 Oct '01

08 Oct '01

[reordered for convinience] Henrik Gedenryd <Henrik.Gedenryd(a)lucs.lu.se> wrote: > danielv(a)netvision.net.il wrote: > Have you read the previous postings where I explained the basic design? Yes, but good idea. I just read them again, with my interest level higher. Quite interesting. > > OTGH, DMs are supposed to replace change sets, but when I make a change > > set, it usually touches a few logical modules, not one. At least > > eventually, it is the conjunction of deltas to several modules, not one. > > Yes, to your project module, you may add multiple deltas, one for each > module you "touch". This also automatically handles the equivalent of > grouping and managing multiple change sets, without having to add another ad > hoc mechanism for doing that. Ah. So something that's born as a change-set (side effect of a lazy afternoon's hacking) later evolves into a project module with several delta modules, each holding the extensions the author believes should be integrated next into various modules he may not be responsible for. Right? IIUC, I like this concept, it relates to how I'm working now on the RB module. BTW, would these DM be submodules or imports? can DMs be submodules? (I'm still trying to pin those concepts down...) > > On one hand, they represent the same thing as a Module, except stored as > > deltas from a specific module. > > OTOH, DMs can be extensions, but Modules can't. ?! > A Module can't be an extension since it isn't an extension of something. A > Delta is by definition an extension from a base (module). In your String>>asUrl example - String is probably a far more long-lived concept than the BasicTextHandling-V.26 module. Would I have to extend a specific Module and version? this seems to flow from your definitions, but practically speaking, asUrl probably doesn't depend on many features of a specific module version, and it would mighty tedious maintenance to be constantly rereleasing the same code with updates requirements. Might be a good idea to have looser definitions of requirements - like BasicTextHandling >= V.26 (this is done in the Debian package system, at least, I don't know about ENVY/Team). But specifically in the case of extending String - This class is so long lived, so basic, that probably even just specifying the name of the module relative to which extension must be defined, is limiting it's use unneccessarily. So IMHO, we should have module that include extensions of classes, even if no module is reference in which the class is defined. Maybe there's a better solution. I don't know if this is what you mean by the half implied reference to DeltaClasses in the DeltaModule comment... > If you look at the code I posted there is an example showing how it is > supposed to work. Sorry, I wasn't able to resolve that reference... what message, what code do you mean? > Having DMs, I've decided to leave change sets untouched. So people can > continue using them like now, but a major point in this discussion has been > the weaknesses of change sets, which is what DMs are meant to remedy. Well, good, that's what wasn't clear. > > What do you mean when you say your are/will be built for a > > "internet-distributed, collaborative open-source project"? > > Well, "delivering runtime (end-user) pieces of software" represents a very > small fraction of what Squeak is used for. Agreed totally. > There are numerous other aspects > that we need to deal with, as have been stated by others already. Handling > contributions, bug fixes, changes to system classes, handling conflicts, > shared repositories, and allowing multiple uncoordinated projects to proceed > independently are some such things. I like what I understand from your proposals so far. So far the idea of "no class extensions unless you specify baseline module and version" seems a bit tighter than my comfort level, otherwise very interesting. > Henrik

11 16

Re: [ANN] Squeak RB release
by danielv＠netvision.net.il 08 Oct '01

08 Oct '01

Thanks for the bug report! The fix will be integrated into the next version. About Lint/Code bugs - Well, that depends on your expectations, I guess... I inspected the results for Object, and it would appear that it hosts lots of messages that are "misplaced", in a sense. For example, this snippet of code out of one of those methods - class := self class. class isMeta ifTrue: [^self]. "a class" new := self clone. class isSystemDefined ifFalse: [uc := deepCopier uniClasses at: class ifAbsent: [nil]. uc ifNil: [deepCopier uniClasses at: class put: (uc := self copyUniClassWith: deepCopier). Objects don't implement copyUniClassWith:, but the tests will protect most objects from venturing there. That generally works (so it's not a bug), and I bet there are some reasonable tradeoffs involved (so maybe it's not even Lint). It would me or you some effort to decide either way, but I sure wouldn't expect it of Smalllint ;-) Generally speaking, Smalllint is more useful on young code (where what looks like a stupid mistake is more likely to be one), but you never know. From a quick inspection, there's lots of good work that can be done cleaning up the Kernel classes starting from a Lint report on it. If you want to start off picking on a weaker victim, just go after my code :-) One enhancement that I know of, and I'll integrate as soon as I'm told it's ready, is a better mechanism to tell Smalllint to ignore "infractions" (using comments in the code). Anyway, thanks for the feedback! Daniel "PhiHo Hoang" <phiho.hoang(a)home.com> wrote: > Hi Daniel, > > I got a walk back 'Error: subscript is out of bounds: 0' when clicking > in SmallLint Results window where there are no result messages. > > I am learning to use RB and found that SmallLint reports 39 'messages > sent but not implemented' bugs for Kernel-Objects and Kernel-Classes. Are > there really bugs for these two categories or it's SmallLint's bug ;-) > > > Cheers, > > PhiHo > > ----- Original Message ----- > From: <danielv(a)netvision.net.il> > To: <squeak-dev(a)lists.squeakfoundation.org> > Sent: Friday, September 28, 2001 8:51 PM > Subject: [ANN] Squeak RB release > > > > Hey guys, another release is up. > > A couple of fixes, and one main new feature - a search feature based on > > parse tree matching. > > > > This lets you find just about any pattern in the code that you can > > describe. You can look in a class, a class category, or all the system > > (that'll take a while!). > > > > The way the search work is like this - > > You give it a pattern that looks like the code you want to match > > `@exp halt > > > > will match "3 halt" and also "(3 + 4) halt". > > > > You can also specify a test clause. For example, > > > > The pattern - > > `@Expression > > with the test clause - > > aNode isVariable and: > > [(class _ Smalltalk classNamed: aNode name) isBehavior and: > > [class allSuperclasses includes: Morph]] > > > > Will look for every expression, that's a variable, that's really a > > global, and a class, and a subclass of Morph (In short, you'll see all > > the methods that reference some Morph class explicitly). > > > > Note that the clause is really compiled as the body of a block, that > > recieves aNode as it's only parameters. > > > > To invoke the search, go to the "RB - Class" menu, and poke the "ind > > code like..." item. > > Invoking it with no class will apply the search to the category, if no > > category either, then all categories. > > > > To load the RB, do (btw, tests show you must delete any Squeak%20RB > > projects from your Squeaklets directory) - > > ProjectLoading > > installRemoteNamed: 'Squeak%20RB.pr' > > from: > > 'http://squeak.heeg.de:8080/TFODC69UQ0MFRIHMG40DRML1CU7VYZ9-4/Squeak%20R > > B.pr' > > named: 'SqueakRB' > > in: Project current. > > > > Daniel > > PS - anyone know the rationale for Project>>fromUrl: working only for > > the predefined servers? > > > >

3 2

[ENH][GOODIE] TwoModeButtonMorph
by Jim Benson 06 Oct '01

06 Oct '01

Part of Project Zurgle This is a modified version of the previously released TwoModeButtonMorph that complies to the specification outlined in the WindowFrames paper. As such, some of the protocols and structures have changed. There are two examples: TwoModeButtonMorph example Clicking the button opens a browser window in the demo. This code expects the file 'close.bmp' to be in the current Squeak directory for the demonstration. and: TwoModeButtonMorph example1 Clicking the button opens a browser window in the demo. This code uses a renderCode block to draw the button, rather than have images associated with it. At this point, I'm thinking that I will not be releasing code to the list until I'm quite a bit further along in an attempt to keep list traffic down. Instead, I'll start gathering it and placing it on a web page, and post when I make significant changes. I post this changeset because I posted a previous version in the past. This code was constructed in Squeak version 3.1a-4173 Jim "Dance like it hurts. Love like you need money" -- Dogbert

6 17

Cache lines and branch prediction.
by Scott A Crosby 04 Oct '01

04 Oct '01

Whoops. Looks like I am obsolete with my memory of CPU's.. Oh, for the days when cache lines were 16 bytes. I feel old already. It looks like you can now assume them to be 32 (Pentium 2) or even 64 bytes. (AMD Athalon) So, given the bigger cache line, which we can expect to be big enough to contain a full entry, my idea of using two tables is no longer useful; it will lead to using more cache lines because entries will be split up between lines. (And the other portions of those cache lines will most likely be wasted.) Given this, my 'cache conscious' design is no longer relevant, and a variant of the current scheme will work well: 1. Make the cache hold 2k-16k entries. 2. Upon a cache hit on the first slot, use it as-is. 3. Upon a cache miss on the first probe, but a hit upon a later probe Replace the first slot probed with the entry of the hit. 4. If all probes miss, lookup the correct method. If there is an empty slot use it, otherwise replace either the 2nd or 3rd (chosen randomly) probe with the lookedup information. Advantages of this: It is still pretty similar to the current scheme and only requires very minimal changes. Also, we never add to the second slot. Reason: #1: Morphic is very heavyweight and uses a lot of selectors, classes, and methods. #3: If it is used a second time before being removed, put it in the first slot so we notice it quickly. #4: We do not want to ever erase an entry. Note that this scheme still wastes 1/3 of the slots, but it will probably increase the hit rate, of the full cache (because its a larger cache) and increase the hit rate of the first probe... Until Jitter comes along and makes this all irrelevant. -- Now, for branch prediction: http://developer.intel.com/design/intarch/techinfo/Pentium/operatn.htm#1004… A branch prediction hit on the intel has a cost of one clock cycle, and may be run in parallel with any other integer instruction, even a comparison!!!! A miss costs a full pipeline flush. (I do not know the cost of this, but I'd assume at least 6 ticks, maybe over 10 clock cycles.) and another 3 to 4 clock cycles. The example, taken from the above: for(k=i+prime;k<=SIZE;k+=prime) flags[k]=FALSE; Can run in 2 clocks/iteration. (plus any cache miss overhead) It runs the store&add in one tick, and the compare&branch in the second tick. So a branch prediction miss costs what? About 10x that of a hit. So, you can see why a branch prediction hit rate of 60% at that critical if-then is so ugly. Increasing it to (say) 95% theoretically (which is our hit rate after 3 probes) would be a 3x improvement. And having a method cache that cannot fit in the CPU cache is not much of a problem either; the CPU cache will figure out what entries are important and hold them. Scott

2 2

[BUG] [ENH] [FIX] Misc small features and bugfixes
by Scott A Crosby 03 Oct '01

03 Oct '01

A variety of small changes to the system. - Small documentation fixes, - Fixing a buglet in creating ReadWriteStreams with on:. - Giving Writestream an understanding of lf/crlf messages. - Giving Socket a lowlevel try-to-send-data function - Giving SharedQueue the ability to safetly check its size with critical sections. - More flexible CharacterSet generation, and more predefined charactersets. - Giving String a 'indexOf:startingAt:endingAt:' function. [*] [*] The corresponding indexOf:startingAt function is automatically built as a plugin. This should be added into the string utilities plugin. Scott -- No DVD movie will ever enter the public domain, nor will any CD. The last CD and the last DVD will have moldered away decades before they leave copyright. This is not encouraging the creation of knowledge in the public domain.

1 1

[Q][Security] in web based squeaking?
by Torge Husfeldt 02 Oct '01

02 Oct '01

Hi there, A while ago I heard of the (theoretical) ability of swikis to support real smalltalk scripting inside the edited pages. Is this secure? Is there any swiki active that makes use of this feature? (if not the first than most probably not the second ;-) Is anybody interested in developmentor actively developing in this direction? On the other hand what about security in the Squeak browser plugin? I gather it is very safe against malicious code but only by being very restrictive - is this still true? If so, I'd like to improve it to give the user the choice between restrictiveness and power/security and insecurity. I'd like to hear from every effort that has been made recently in this direction, any active projects or any thoughts you have on this topic. Looks like this is _the_ thing I will be working on for the next few months if it proves to be worth(not commercially though) it. Thanks in advance Torge P.S.: Even replys like "wrong list, post it there:..." welcome

4 4

Why we should remove {} from Squeak
by ducasse stephane 02 Oct '01

02 Oct '01

Hi daniel I send it into the mailing list too becuase this is important >From a language design perspective #() is necessary because it is compiled statically and cannot not be simulated by other construct. Storing #() in Stack frame is not really goo but this is an implementation aspect. #{} is not necessary because this is shortcut macro-expansion for Array new add: .... So why having only this one. There is two ways: or you have a way to describe macro-expansion like in Scheme where you can define any macro in a clear way (you know the semantic of the stuff or we remove it because why not having {{}} for Set new add: {#[]} for OrderedCollection new....Because I use more OrderedCollection than array Removing would simply the parser, compiler, Object simpler the printer printer the VM, the bytes code emitter, then Squeak would be a bit more standard but this is not the most important point. Why can I send the message caseOf: to ___any Object___ berk!!!! I'm pretty sure that I can destroy the semantics prevservation of certain refactoring (I did it for ObjectStudio RB because OS has this &(*&*&( construct too). The parser of Rb in Squeak has to e more complex than the one of VW because of the expressions inside {}.... And it would avoid to have people writing, (I saw that in indsutrial code people wrote __huge__ { {,.n,mnm, } {knlkj lk jjkl with block everywhere} {,.n,mnm, } {knlkj lk jjkl with block everywhere} {,.n,mnm, } {knlkj lk jjkl with block everywhere} {,.n,mnm, } {knlkj lk jjkl with block everywhere} {,.n,mnm, } {knlkj lk jjkl with block everywhere} } So I hope that I convince you. I will try to came up with a change ste that REMOVE this anthrax from Squeak!!! Stef

8 9

Help with sound
by Daniel Joyce 02 Oct '01

02 Oct '01

The Midi works great. But none of the sound playing demos work. The wave editor, etc. Is sound that bad on Unix? NB, normal window/open/close sound effects and others work fine too. Where is that code handled? Perhaps I can look there? Or is it just that all the sampled sound demos are broken? -Daniel

4 3

Re: Skip List in #4308
by danielv＠netvision.net.il 02 Oct '01

02 Oct '01

Well, I really cannot say who said what, or who I'm answering to, but - In Celeste, this could make recovery after image crashes faster. With SortedCollections re-adding the n logged actions done before the crash takes n^2 time, with SkipLists, presumably, n*log(n). Or we could tweak Celeste to not update this list during recovery, and resort it once later. Still nicer to just have nice behaving data structures... Anyone know what this thing eats in memory usage? Daniel Karl Ramberg <karl.ramberg(a)chello.se> wrote: > On Tue, 25 Sep 2001, Andreas Raab wrote: > > > Folks, > > > > I noticed them coming into Squeak a while ago and always since I wanted to > > ask this question: What the heck are they good for?! ;-) I don't see them > Basically, Skiplists (and/or balanced tree implementations of > SortedCollection) should be used anywhere where you have a lot of > inserts/removals from a sorted collection, and profiling shows that that > is a problem. > > Scott

3 3

Theoretical texts
by Maarten Maartensz 01 Oct '01

01 Oct '01

L.S. I am new to Squeak (since July) and should like to recieve references to texts that address foundational issues of - Smalltalk - Self - Blockclosures (and Lambda-calculus) on a fairly sophisticated level, especially if these texts were written with reference to and knowledge of mathematical logic. I have read the Blue Book and other introductions, but these do not answer most of the questions I'd like to see discussed, such as e.g. the relation between lambda-calculus and blockclosures. Probably, the sort of stuff I am most interested in would be on line Ph.D.-theses that concern aspects of smalltalk and have been written by mathematical people. Incidentally, on my site I have been uploading over the last months some of my impressions of Squeak (section Computing). Regards, Maarten Maartensz. Website in Holland about philosophy, logic, M.E. (Myalgic Encephalomyelitis) and much more: http://www.xs4all.nl/~maartens/

3 2

Jump to page: