January 2010 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Nanosecond-level profiling in Pharo?

by David T. Lewis

On Fri, Jan 08, 2010 at 12:29:18AM +0200, Igor Stasenko wrote: > 2010/1/7 Alexandre Bergel <alexandre(a)bergel.eu>: > > Hi! > > > > I am currently working on a new profiler for Pharo. The idea is to > > graphically renders a profiling using Mondrian. I hope to release it > > soon. > > When discussing with an expert in the field (Walter Binder), the > > question about time grain was raised. I am partly based on > > MessageTally, which has a millisecond grain. Anyone is aware of a > > finer grain? At the level of nanosecond? Does it makes sense since > > there is no JIT (yet)? > > > > AFAIK, there's no primitives in VM, which provide timing with > nanosecond precision. Also, the accuracy of the #millisecondClockValue available to the image is different on different platforms. I haven't checked lately, but I think the Windows VM is good pretty much to the nearest millisecond, while Linux is more variable. Not sure about OS X or other unixy platforms. So even if you are dealing with times that are measured to millisecond precision, the accuracy may be less depending on the underlying platform and VM. Dave

Jan. 8, 2010

Re: [Pharo-project] [ANN] Metacello configurations for Pharo.

by Miguel Enrique CobÃ¡ Martinez

> > ps: There is another version about some tests...sorry, those are the > OB tests (BogusTestCase) and I don't know why it depends on soemthing > I don't know what it is...if someone can help me :) Yes I got this message on a 10504 rc1 PharoCore. Other than that, the process went ok. Good job. -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 8, 2010

Re: [Pharo-project] Notifications of windows opening/closing/focus

by Schwab,Wilhelm K

Henry, Then we need to fix the weak collections. Dolphin (of the mid-90's) already had thread-safe weak collections and a high-priority process to detect and clean up "damage" to them; it's time to catch up. The more I think about it, your idea is sounding a lot like the dependency days of MVC - remember #release? I'm trying to forget. I am not trying to be confrontational, but we can and should do better than this. Bill -----Original Message----- From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Henrik Sperre Johansen Sent: Thursday, January 07, 2010 8:27 PM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] Notifications of windows opening/closing/focus On 08.01.2010 02:07, Schwab,Wilhelm K wrote: > Henry, > > No argument, your solution is simpler. However, we do not have announcements like this now, and it appears that the poor state of weak collections in Squeak/Pharo is standing in the way of a clean implementation. My vote is to tackle the weaklings, at least until they work well enough to support what we should have in this case. Otherwise, we kick the weak can down the road as has been done all along rather than fixing it, AND we end up with a sub-standard announcments-based framework. > > Let's do it right. We will then have an evolving/improving weak collection and something in the image that puts some stress on same. If I'm missing something, please enlighten me. > > Bill > Weak collections does not stand in the way of making the use case of unregistering dynamic registrations manually easier. They stand in the way of the use case of unregistrating registeration automatically. Here's what you're missing: No matter how you implement the automatic unregistering, it will be nondetermenistic, for the exact same reasons finalization is nondeterministic. Thus, if you require deterministic unregistration, you have to do it manually, and the use case of making that easier to do have value in its own right. Cheers, Henry _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Jan. 8, 2010

Re: [Pharo-project] Notifications of windows opening/closing/focus

by Henrik Sperre Johansen

On 08.01.2010 02:07, Schwab,Wilhelm K wrote: > Henry, > > No argument, your solution is simpler. However, we do not have announcements like this now, and it appears that the poor state of weak collections in Squeak/Pharo is standing in the way of a clean implementation. My vote is to tackle the weaklings, at least until they work well enough to support what we should have in this case. Otherwise, we kick the weak can down the road as has been done all along rather than fixing it, AND we end up with a sub-standard announcments-based framework. > > Let's do it right. We will then have an evolving/improving weak collection and something in the image that puts some stress on same. If I'm missing something, please enlighten me. > > Bill > Weak collections does not stand in the way of making the use case of unregistering dynamic registrations manually easier. They stand in the way of the use case of unregistrating registeration automatically. Here's what you're missing: No matter how you implement the automatic unregistering, it will be nondetermenistic, for the exact same reasons finalization is nondeterministic. Thus, if you require deterministic unregistration, you have to do it manually, and the use case of making that easier to do have value in its own right. Cheers, Henry

Jan. 8, 2010

Re: [Pharo-project] [ANN] Metacello configurations for Pharo.

by Miguel Enrique CobÃ¡ Martinez

> > Loader new load: 'Pharo'. and get a full PharoDev image :) > Wonderful! -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 8, 2010

Re: [Pharo-project] [ANN] Metacello configurations for Pharo.

by Miguel Enrique CobÃ¡ Martinez

Wonderful news Mariano! This will ease the usage of Pharo in ways unseen before. Thanks. -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 8, 2010

Re: [Pharo-project] Notifications of windows opening/closing/focus

by Levente Uzonyi

On Fri, 8 Jan 2010, Henrik Sperre Johansen wrote: > On 08.01.2010 01:32, Schwab,Wilhelm K wrote: >> One should always be *able* to unsubscribe, but it should not be required. >> >> Squeak's weak collections do not measure up to the competition. Please take a look at my 1/2/2010 post "RE: [Pharo-project] Can't debug in 1.1". I doubt SharedLookupTable could be called efficient, but it solved some problems I was having due to the current implementation. I welcome concrete examples of how horrible it is, and offer it up as a starting point or simple inspiration for something more useful. >> >> Bill >> > I guess I should have quoted Gary's question of whether the approach > would be hard to unregister instead. > Originally, I (like you) thought not. > After actually looking at the implementation, I do. > > What I posted is one way to "solve" this, weak subscriptions is another. > On a scale of "what's harder to get right", I'd bet there's a larger > chance of having 2 arg blocks implemented first. I think there won't be 2 argument blocks, because that would make the current minimal implementation more complex or slower. (AXAnnouncements have this feature btw.) Levente > > Cheers, > Henry > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Jan. 8, 2010

Re: [Pharo-project] Pharo testers.... WE NEED YOU!!!

by Miguel Enrique CobÃ¡ Martinez

El jue, 07-01-2010 a las 22:02 +0100, Mariano Martinez Peck escribiÃ³: > Do you think we need a mailing list ? Or just copying each person is > enough ? -1 to a new list. A mail to the list with the tag proposed is enough. Thanks, by the way for the effort. -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 8, 2010

Re: [Pharo-project] Notifications of windows opening/closing/focus

by Schwab,Wilhelm K

Henry, No argument, your solution is simpler. However, we do not have announcements like this now, and it appears that the poor state of weak collections in Squeak/Pharo is standing in the way of a clean implementation. My vote is to tackle the weaklings, at least until they work well enough to support what we should have in this case. Otherwise, we kick the weak can down the road as has been done all along rather than fixing it, AND we end up with a sub-standard announcments-based framework. Let's do it right. We will then have an evolving/improving weak collection and something in the image that puts some stress on same. If I'm missing something, please enlighten me. Bill -----Original Message----- From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Henrik Sperre Johansen Sent: Thursday, January 07, 2010 7:45 PM To: pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] Notifications of windows opening/closing/focus On 08.01.2010 01:32, Schwab,Wilhelm K wrote: > One should always be *able* to unsubscribe, but it should not be required. > > Squeak's weak collections do not measure up to the competition. Please take a look at my 1/2/2010 post "RE: [Pharo-project] Can't debug in 1.1". I doubt SharedLookupTable could be called efficient, but it solved some problems I was having due to the current implementation. I welcome concrete examples of how horrible it is, and offer it up as a starting point or simple inspiration for something more useful. > > Bill > I guess I should have quoted Gary's question of whether the approach would be hard to unregister instead. Originally, I (like you) thought not. After actually looking at the implementation, I do. What I posted is one way to "solve" this, weak subscriptions is another. On a scale of "what's harder to get right", I'd bet there's a larger chance of having 2 arg blocks implemented first. Cheers, Henry _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Jan. 8, 2010

Re: [Pharo-project] Notifications of windows opening/closing/focus

by Henrik Sperre Johansen

On 08.01.2010 01:32, Schwab,Wilhelm K wrote: > One should always be *able* to unsubscribe, but it should not be required. > > Squeak's weak collections do not measure up to the competition. Please take a look at my 1/2/2010 post "RE: [Pharo-project] Can't debug in 1.1". I doubt SharedLookupTable could be called efficient, but it solved some problems I was having due to the current implementation. I welcome concrete examples of how horrible it is, and offer it up as a starting point or simple inspiration for something more useful. > > Bill > I guess I should have quoted Gary's question of whether the approach would be hard to unregister instead. Originally, I (like you) thought not. After actually looking at the implementation, I do. What I posted is one way to "solve" this, weak subscriptions is another. On a scale of "what's harder to get right", I'd bet there's a larger chance of having 2 arg blocks implemented first. Cheers, Henry

Jan. 8, 2010