- Pharo-dev - lists.pharo.org

Re: [Pharo-project] open AtomBouncingMorphs in 3.7 and 3.9 is amazing...
by Bill Schwab May 31, 2008

May 31, 2008

Stef, Let me guess: 3.7 is MUCH faster, right??? Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> stephane.ducasse(a)inria.fr 05/30/08 11:25 AM >>> Hi guys it would be really interesting to understand why there is a so large difference between the two. Stef _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

2 1

[Pharo-project] Fwd: patch for Squeak 3.9.2 / whatever
by StÃ©phane Ducasse May 31, 2008

May 31, 2008

Begin forwarded message: > From: "Philippe Marschall" <philippe.marschall(a)gmail.com> > Date: May 30, 2008 10:52:50 PM CEDT > To: "StÃ©phane Ducasse" <stephane.ducasse(a)inria.fr>, "Adrian > Lienhard" <alienhard(a)netstyle.ch> > Subject: patch for Squeak 3.9.2 / whatever > > http://bugs.squeak.org/view.php?id=6834 >

1 0

[Pharo-project] Fwd: Patch should be applied: Delay class>>timeoutSemaphore:afterMSecs:
by StÃ©phane Ducasse May 31, 2008

May 31, 2008

Begin forwarded message: > From: "Masashi UMEZAWA" <masashi.umezawa(a)gmail.com> > Date: May 31, 2008 4:56:00 AM CEDT > To: "stephane ducasse" <stephane.ducasse(a)free.fr> > Subject: Fwd: Patch should be applied: Delay > class>>timeoutSemaphore:afterMSecs: > > Hi Stef, > > This is just FYI (maybe you do not read v3dot10 list). > > 3.9.1 and Pharo have the same bug also. Networking application would > stop suddenly for this bug... > > ---------- Forwarded message ---------- > From: Masashi UMEZAWA <masashi.umezawa(a)gmail.com> > Date: 2008/5/29 > Subject: Patch should be applied: Delay > class>>timeoutSemaphore:afterMSecs: > To: "Discussion about development of Squeak 3.10" > <v3dot10(a)lists.squeakfoundation.org> > > > Hi all, > > I've downloaded 3.10.1, and found that an important patch on Delay > seems not applied. > > Please see the details on Mantis: > http://bugs.squeak.org/view.php?id=6834 > > Strangely, the patch is not applied on 3.9 series also. I think it is > clearly a bug and it should be fixed. > Socket timeout relies on this method, so the bug is fatal. > > Cheers, > -- > [:masashi | ^umezawa] > > Cheers, > -- > [:masashi | ^umezawa] >

1 0

Re: [Pharo-project] Replacing 'ReadStream on smth:' by 'smth readStream'
by Bill Schwab May 31, 2008

May 31, 2008

Adrian, Fair enough. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> adi(a)netstyle.ch 05/31/08 4:55 AM >>> Hi Bill, Damien, I'm not against such refactorings at all. And I also know that stuff will break in the process. The only point I wanted to make is that those changes, especially if they touch a large code base, need to be done carefully. In the proposed change of #readStream, I assume Damien has thought about the consequences. As he noted in another mail, my example with Matrix indeed is not a problem. I just picked an implementor of #readStream to illustrate that such a change may not necessarily be behavior preserving. So, go ahead! Cheers, Adrian On May 30, 2008, at 22:46 , Bill Schwab wrote: > Stef, > > Whether or not I agree depends on what you mean. While I think I > see where Damien is going, my read is that #on: is actually > negotiable, given that we should be able to either fix (or simply > replace and than alias) ReadStream - two different ways of solving > the same problem. But if you concerned about breaking some things > that Damien and I are committed to fixing as part of the change, > then I submit that we will be helpless when we move from the > relatively calm world of streams to the chaos of morphic. > > Which part bothers you? > > The plan is to rennovate (sometimes messy) and *then* slowly evolve > the resulting stable system, right? > > Bill > > > > > Wilhelm K. Schwab, Ph.D. > University of Florida > Department of Anesthesiology > PO Box 100254 > Gainesville, FL 32610-0254 > > Email: bschwab(a)anest.ufl.edu > Tel: (352) 846-1285 > FAX: (352) 392-7029 > > >>>> stephane.ducasse(a)inria.fr 5/30/2008 3:27:20 PM >>> > I agree with adrian. > For pharo we should move but we should also make sure it is really > stable. > > Stef > > On May 30, 2008, at 2:18 PM, Adrian Lienhard wrote: > >> Have you verified that your change does not inadvertently modify any >> behavior? Its not likely, but changes like this can introduce subtle >> bugs. For instance, Matrix overrides readStream. So if there exists >> any use of "ReadStream on: aMatrix" that is there to bypass the >> overridden behavior, the code will break. >> >> Cheers, >> Adrian >> >> On May 30, 2008, at 10:47 , Damien Cassou wrote: >> >>> Hi, >>> >>> I've just published in the pharo inbox a package for the System >>> category. This change has been generated automatically through the >>> rewriter: >>> >>> ParseTreeRewriter new >>> replace: 'ReadStream on: ``@object' with: '``@object readStream'; >>> yourself >>> >>> The formatting of the changed method is completely different due to >>> the use of the refactoring browser. This makes the diffs not >>> readable. >>> Even if it is a real problem, I don't see any simple solution. What >>> do >>> you think? >>> >>> -- >>> Damien Cassou >>> Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good >>> luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu) >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

Re: [Pharo-project] Replacing 'ReadStream on smth:' by 'smth readStream'
by Bill Schwab May 30, 2008

May 30, 2008

Damien, I don't think I care how we do it (Nile should be fine), so long as we end up with good stream behavior. In my mind, that means breaking some things (on an informal survey, I find a lot of stuff that should not care because it never hits the nil return value), and I am volunteering to help fix what I help break, provided we agree up front to use the transformed code (assuming it works). I doubt the latter will a problem with this group, as long as we plan and then do the work. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> damien.cassou(a)gmail.com 05/30/08 12:02 PM >>> On Fri, May 30, 2008 at 5:54 PM, Bill Schwab <BSchwab(a)anest.ufl.edu> wrote: > Damien, > > The only potential problem is that you assume tests will catch everything. However, I think streams are broken, and I will certainly help fix what we break in getting them right. I guess the change I'm applying won't cause any problem but I can't be sure. I won't loose time fixing Streams as I would prefer installing Nile if everybody agree. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu) _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

Re: [Pharo-project] Replacing 'ReadStream on smth:' by 'smth readStream'
by Bill Schwab May 30, 2008

May 30, 2008

Adrian, Fair enough, but some house cleaning is also in order, and it *will* break some external code. If we fail to do the right things, we will have just another failed fork. Do the right things, and people will do right things to their code to get the benefits of a successful open Smalltalk. Just my 2 asCents. BTW, I *think* you will find that anyone wanting to fix external code can simply grab an image with it, use the what ever transformations we end up defining on their code (installing the RB might be a challenge??), export the fixed code, and file it into Pharo. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> adi(a)netstyle.ch 05/30/08 12:17 PM >>> On May 30, 2008, at 15:37 , Damien Cassou wrote: > On Fri, May 30, 2008 at 2:18 PM, Adrian Lienhard <adi(a)netstyle.ch> > wrote: >> Have you verified that your change does not inadvertently modify any >> behavior? Its not likely, but changes like this can introduce >> subtle bugs. >> For instance, Matrix overrides readStream. So if there exists any >> use of >> "ReadStream on: aMatrix" that is there to bypass the overridden >> behavior, >> the code will break. > > It should not break anything (if it does, the test will fail and I > will fix things :-)). In fact, ReadStream only deals with > SequenceableStreams so it should be ok. Do you agree? No I don't agree. Because it is not likely that there is a test for Matrix that will show a potential bug. I just wanted to point out that it is important to precisely look at the effect of such "global" transformations. (E.g. like Edgar tried to replace all #== with #= IIRC ;) With your proposed change, I think it is unlikely you will break anything, but you cannot say for sure without detailled *manual* checking. The difficulty is that you theoretically needed to verify that there is never passed a Matrix instance to "ReadStream on:". Also to consider with such changes is that you may break external code that you don't know of. To sum up, I'm not against such improvements at all, but they have to be done carefully, i.e., the argument that tests will show any problem does not hold. Cheers, Adrian > -- > Damien Cassou > Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good > luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu) > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

Re: [Pharo-project] Replacing 'ReadStream on smth:' by 'smth readStream'
by Bill Schwab May 30, 2008

May 30, 2008

Damien, That would be great. It does me little good in my current version of Dolphin, though I should look at the more recent versions. Making a long story short, D6 broke a lot of my code and reduced compatibility with win2k (the big show-stopper for me at the time), so I never upgraded. I am currently far more bothered by unwanted changes in Windows than I have ever been by deficiencies in Dolphin. IIRC, some of the breaking changes are now understood, so it might pay to have another look, though I still need to run on win2k, just (sadly IMHO) not as much as a few years ago. Another option would be to load code into Pharo and hit it with an up to date RB. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> damien.cassou(a)gmail.com 05/30/08 9:47 AM >>> On Fri, May 30, 2008 at 2:13 PM, Bill Schwab <BSchwab(a)anest.ufl.edu> wrote: > Interesting! As an aside, I have a love/hate relationship with the RB, and almost never let it rewrite any of my Dolphin code. The XP guys will recoil in horror at this, but I have too often been saved by 18 month old comments to have them disrespectfully scattered around my source. If that is fixed, I will become a believer. I *do* use pull up/down to great advantage, and will sometimes use other refactorings on new code that has not had time to accumulate valuable comments. This problem has been fixed some time ago IIRC. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu) _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

May 30, 2008

Hi guys it would be really interesting to understand why there is a so large difference between the two. Stef

1 0

[Pharo-project] Package Files has unsaved changes in latest Pharo
by Damien Cassou May 30, 2008

May 30, 2008

Hi, if you take latest released Pharo, open Monticello, select the package Files and the repository Pharo, you will see changes in the class declaration CrLfFileStream. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu)

1 1

Re: [Pharo-project] Replacing 'ReadStream on smth:' by 'smth readStream'
by Bill Schwab May 30, 2008

May 30, 2008

Damien, Interesting! As an aside, I have a love/hate relationship with the RB, and almost never let it rewrite any of my Dolphin code. The XP guys will recoil in horror at this, but I have too often been saved by 18 month old comments to have them disrespectfully scattered around my source. If that is fixed, I will become a believer. I *do* use pull up/down to great advantage, and will sometimes use other refactorings on new code that has not had time to accumulate valuable comments. The above said, I have considered doing something like your rewrite to fix stream method naming snags. Consider #next. As we have discussed, I think it should be named #nextOrNil, allowing #next to behave as it does in Dolphin. Yes, I have a bias toward Dolphin, but I also consider its streams to be elegantly/correctly built. One way to fix it would be to transform #next into #nextOrNil (which is what Squeak's #next does), and then add a #next that signals EndOfStream on exhaustion; EndOfStream would have no default action in this scenario. If we could further transform any idioms such as aStream next ifNotNil:[ :x | ... ] into aStream nextIfPresent:[ :x | ... ] etc., that could be great if we did it well. Even a clean change to #nextOrNil would allow us to make an otherwise breaking change as painlessly as possible. FWIW, I have no direct problem with #on:, though I see where you are going. We might get a similar benefit by simply changing ReadStream in the system dictionary, making it an alias for NSReadStream?? Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029 >>> damien.cassou(a)gmail.com 05/30/08 4:47 AM >>> Hi, I've just published in the pharo inbox a package for the System category. This change has been generated automatically through the rewriter: ParseTreeRewriter new replace: 'ReadStream on: ``@object' with: '``@object readStream'; yourself The formatting of the changed method is completely different due to the use of the refactoring browser. This makes the diffs not readable. Even if it is a real problem, I don't see any simple solution. What do you think? -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu) _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

2 2