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

Re: [Pharo-project] [ANN] NativeBoost meets Linux

by Igor Stasenko

On 28 September 2010 14:27, Schwab,Wilhelm K <bschwab(a)anest.ufl.edu> wrote: > Sig, > > That's kinda scary coming from someone who thinks in assembly instructions :) Â I generally have resorted to tracing when debugging code in my libraries. Â So far, DebugView has been hard to beat for that. Â Not the same level, I realize, just a plug for anything that helps us see what the machine is doing, especially when it something goes wrong. > I tried many different debugging tools in recent months. I found that all of them still inferior to examine and analysis of code :) Some of them even refusing to show me anything userful, or making is so hard, that its easier to simply read source code, rather than debugging. And debugging VM is not the same as debugging usual C program.. Now add a native code into the soup, and you'll understand why all of these cool debuggers are incapable to help.. unless you spent 100++ hours with them and mastered everything about debugger. >From other side, there is not much complex things in code generator. I isolating things into small pieces of code (couple lines of smalltalk code per method), so there is nearly impossible to do a mistake. It is relatively easy to find a mistake at first run of the code. The bug which i fixed recently was in callback code, which is quite complex and is rather exception than common case (a method contains maybe 100 lines of code). And bug was a logical mistake, not something like code generator producing wrong instructions etc :) > Bill > > > ________________________________________ > From: pharo-project-bounces(a)lists.gforge.inria.fr [pharo-project-bounces(a)lists.gforge.inria.fr] On Behalf Of Igor Stasenko [siguctua(a)gmail.com] > Sent: Tuesday, September 28, 2010 1:12 AM > To: The general-purpose Squeak developers list; Pharo Development > Subject: Re: [Pharo-project] [ANN] NativeBoost meets Linux > > On 26 September 2010 21:09, Igor Stasenko <siguctua(a)gmail.com> wrote: >> Hello all, >> >> I just uploaded a fresh snapshots of NB plugin and image code, >> with necessary bits for using it under Linux. >> >> I don't know how to package Linux VM binaries, so you won't find a prebuilt VM >> and should build it yourself (anyone wants to help me with that, so i >> can add a prebuilt VM at downloads?). >> >> As usual, follow the installation instructions, located here: >> http://code.google.com/p/nativeboost/wiki/Installation >> >> Warning: recently, i found a bug in callback code. >> So, please don't try using smalltalk callbacks until i fix it :) >> >> Running tests should be safe (i disabled the test which causing VM crash). >> > > Update: bug was fixed after two days sitting with debugger and trying > to convince gdb to let me debug and find the > place with problem, which finished by looking to own code once again > and finding what it doing wrong. > Thanks DDD+gdb developers with not letting me simply debug my stuff, > but instead train > my brain's analytic capabilities. :) > > >> >> -- >> Best regards, >> Igor Stasenko AKA sig. >> > > > > -- > Best regards, > Igor Stasenko AKA sig. > > _______________________________________________ > 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 > -- Best regards, Igor Stasenko AKA sig.

Sept. 28, 2010

[Pharo-project] Metacello Configuration Browser?

by Guillermo Polito

I'm always reading people saying that metacello configurations are some kind of... overdesign. So, maybe we need a tool (or to modify monticello browser) wich allow us to: - know wich packages are core and dev. Sometimes using a dev image, I write things depending on Dev packages and I don't realize it. - create automatically or propose configurations depending on the actual state of the packages. - browse configurations nicely :). It's a new concept, we need a nice tool to support it. Maybe if the configs are easily created, we can have more of them, and we will be no longer so lazy. Cheers, Guille

Sept. 28, 2010

Re: [Pharo-project] About installing NativeBoost in Linux

by Igor Stasenko

2010/9/28 Fernando olivero <fernando.olivero(a)usi.ch>: > > I can provide an image with VMMaker + NB plugin, if you want. > > Not necesary! > I found out that using the new ConfigurationOfVMMaker its pretty > straightforward, to have a "VMMaker image". > > > I running linux in virtual box, and had a problems with OpenGL, so i > had to disable it & exclude everything > which using it (B3DAcceleratorPlugin) from build, as well as some > other plugins, which for one or another > reason didn't worked for me. > So, it could be not the same VM you wanna use on your box. :) > > > > I ran into some problems when running make install, could you please send me > the vmmaker configuration file you are using? > So i can mimic the plugins configuration that worked for you, and > successfully compile the vm. > Will send in private mail. > Just try to build a VM with plugin. And once it can run a native code, > the rest is a piece of cake :) > I can give you detailed explanation, what you should implement/override > to complete NB bootstrap on Mac, once you will be ready. > > Ok, will let you know. > > > pd: > Maybe this could be useful to put in the native boost installation page , > Gofer it > squeaksource: 'MetacelloRepository' ; > package: 'ConfigurationOfVMMaker' ; > load > (Smalltalk globals classNamed: 'ConfigurationOfVMMaker' ) load. i added this to wiki page. > pd: what about doing a ConfigurationOfNativeBoostInstaller? I'm still didn't came to the point where i need to use metacello. I did the simplest possible thing: wrote installer script which works ok in Squeak & Pharo. If someone provide a Metacello config , i can add it to repository. One more way of installing it doesn't hurts :) > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- Best regards, Igor Stasenko AKA sig.

Sept. 28, 2010

Re: [Pharo-project] Smalltalks 2010 - Declared as Important Province Event

by Alexandre Bergel

Hernan, this is really cool. I hope Smalltalk will bootstrap soon in Chile. Cheers, Alexandre On 27 Sep 2010, at 20:46, Hernan Wilkinson wrote: > Hi, > we just wanted to share with you that Smalltalks 2010 (the 4th Argentinean Smalltalk conference) has been declare of "public interest" at the Entre Rios province (where the city of Concepcion del Uruguay is and where the conference is going to be held). > We are very happy and proud of such recognition to Smalltalk and our community!! > We hope to see you there! > > Hernan. > > PS: Attached is the statement... although, it is is Spanish :-) > > <Nota ACTIER Smalltalk declaraciÃ³n de InterÃ©s.pdf>_______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Sept. 28, 2010

Re: [Pharo-project] Feature request for Opal

by Alexandre Bergel

Cool idea Alexandre On 28 Sep 2010, at 03:30, Igor Stasenko wrote: > I want an API for easy plugging in an arbitrary pragma handler, > which could then either completely intercept parser/compiler output or > modify output. > > This is userful for things like clean implementation of FFI pragmas, > without patching the Parser > as well as for scenarios like: > > myMethod: myArg > <myLanguage> > > foo|<bar**baz > > > so, once pragma <myLanguage> parsed, and handled, an input source > stream redirected to completely different parser/compier > pair which, of course, should produce a compiled method as a result, > but may use any syntax. > > > > -- > Best regards, > Igor Stasenko AKA sig. > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Sept. 28, 2010

[Pharo-project] [update 1.2] #12151

by StÃ©phane Ducasse

12151 ----- - Issue 3007: user of generate: in debugit code - Issue 3008: Delete Graphics working copy

Sept. 28, 2010

[Pharo-project] [update 1.2] #12150

by StÃ©phane Ducasse

12150 ----- - Issue 2991: PharoKernel Part 15. Part 3

Sept. 28, 2010

[Pharo-project] Category rename doesn't show up in Monticello Merge/Changes

by Mariano Martinez Peck

Hi. If you take a class, rename a method category, and see "Changes" with Monticello, you will see its methods like "revision changed" but you are not able to see what really happened, like for example, a method category rename. This should be fixed and we should be able to see all changes, included categories names. Can someone reproduce this so that I can create a ticket? thanks Mariano

Sept. 28, 2010

Re: [Pharo-project] ECompletion/OCompletion tests takes A LIFE!!

by Levente Uzonyi

On Tue, 28 Sep 2010, Henrik Sperre Johansen wrote: > On 28.09.2010 01:43, Levente Uzonyi wrote: >> On Tue, 28 Sep 2010, Henrik Sperre Johansen wrote: >> >>> On 28.09.2010 00:41, Levente Uzonyi wrote: >>>> On Tue, 28 Sep 2010, Henrik Sperre Johansen wrote: >>>> >>>>> On 28.09.2010 00:07, Levente Uzonyi wrote: >>>>>> On Mon, 27 Sep 2010, Mariano Martinez Peck wrote: >>>>>> >>>>>>> 2010/9/27 Levente Uzonyi <leves(a)elte.hu> >>>>>>> >>>>>>>> On Mon, 27 Sep 2010, Mariano Martinez Peck wrote: >>>>>>>> >>>>>>>> IF you take a 1.1 core and just load OCompletion: >>>>>>>>> >>>>>>>>> MessageTally time: [ECContextTest new testUntypedVarsOnly] 20907 >>>>>>>>> >>>>>>>>> If you load OCompletion and the rest of the dev image: >>>>>>>>> >>>>>>>>> MessageTally time: [ECContextTest new testUntypedVarsOnly] 100137 >>>>>>>>> >>>>>>>>> >>>>>>>>> However, IN pharo dev 1.0 it is >>>>>>>>> >>>>>>>>> MessageTally time: [ECContextTest new testUntypedVarsOnly] 7405 >>>>>>>>> >>>>>>>>> >>>>>>>>> sooo.... ???? >>>>>>>>> >>>>>>>>> something has changed? >>>>>>>>> >>>>>>>> >>>>>>>> As I said, SortedCollection is not suitable for this kind of usage. >>>>>>> >>>>>>> >>>>>>> But in Pharo 1.0 OCompletion was doing the same....so? did >>>>>>> SortedCollection >>>>>>> chang between 1.0 and 1.1? >>>>>> >>>>>> No, something else changed. The same test is still 5x faster in 1.0 dev >>>>>> with the new code. >>>>> TimeProfiler to the rescue: >>>>> Preferences class >> ecompletionCaseSensitive >>>>> 1.0: (1725ms) >>>>> 1.1 (19724ms) >>>>> >>>>> Preference>>settingValue which is there for backwards-compatability in >>>>> 1.1 does both symbol creation and respondsTo:, so yeah, it's much slower >>>>> than the old dictionary lookup. >>>> >>>> Nice find. If you check my code, you'll find that I optimized that away. >>>> :) >>>> >>>> >>>> Levente >>> Yes ,you moved it out of the actual sort loop :) >>> Rewrite the methods in ECPreferences to use the new syntax: >>> <preference: category: description: type:>, and you shouldn't even have to >>> do that to get ok performance. >>> >>> IMHO, it's rather disturbing to see things like EC, where its preferences >>> are actually separated tucked away in an ECPreferences class, only to be >>> delegated from there over to Preferences... >> >> The preferences should be updated, but sending a message once is still >> better than 2k-30k times. >> >> >> Levente > Sure, on the order of a couple of milliseconds difference. > I keep forgetting the fact we generally have different definitions of what > constitutes "ok performance" ;) 30k was a bit optimistic assumption. For example the Pharo 1.1 OneClick image has 3220 globals. The sortBlock asks for the preference twice per evaluation. SortedCollection uses a binary search during #add:. If we are optimistic, then the preference is asked 3220 * (3220 ln / 2 ln) * 2 times which is ~75k. In the same image the "old" code asked for the preference 2282899 times during the tests, while the new code did it only 398 times. If the preference is changed to return the value of a variable, then the difference is only 80 ms, otherwise it's 36.6 seconds. 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 >

Sept. 28, 2010

Re: [Pharo-project] [ANN] NativeBoost meets Linux

by Schwab,Wilhelm K

Sig, That's kinda scary coming from someone who thinks in assembly instructions :) I generally have resorted to tracing when debugging code in my libraries. So far, DebugView has been hard to beat for that. Not the same level, I realize, just a plug for anything that helps us see what the machine is doing, especially when it something goes wrong. Bill ________________________________________ From: pharo-project-bounces(a)lists.gforge.inria.fr [pharo-project-bounces(a)lists.gforge.inria.fr] On Behalf Of Igor Stasenko [siguctua(a)gmail.com] Sent: Tuesday, September 28, 2010 1:12 AM To: The general-purpose Squeak developers list; Pharo Development Subject: Re: [Pharo-project] [ANN] NativeBoost meets Linux On 26 September 2010 21:09, Igor Stasenko <siguctua(a)gmail.com> wrote: > Hello all, > > I just uploaded a fresh snapshots of NB plugin and image code, > with necessary bits for using it under Linux. > > I don't know how to package Linux VM binaries, so you won't find a prebuilt VM > and should build it yourself (anyone wants to help me with that, so i > can add a prebuilt VM at downloads?). > > As usual, follow the installation instructions, located here: > http://code.google.com/p/nativeboost/wiki/Installation > > Warning: recently, i found a bug in callback code. > So, please don't try using smalltalk callbacks until i fix it :) > > Running tests should be safe (i disabled the test which causing VM crash). > Update: bug was fixed after two days sitting with debugger and trying to convince gdb to let me debug and find the place with problem, which finished by looking to own code once again and finding what it doing wrong. Thanks DDD+gdb developers with not letting me simply debug my stuff, but instead train my brain's analytic capabilities. :) > > -- > Best regards, > Igor Stasenko AKA sig. > -- Best regards, Igor Stasenko AKA sig. _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Sept. 28, 2010