October 2008 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] using this mailing list

by StÃ©phane Ducasse

This is strange we will check. Stef On Oct 9, 2008, at 10:44 PM, nicolas cellier wrote: > > Apologies for previous ugly formatted post... > I have problems using this mailing list because: > > 1) I do not receive the posts directly (ncellier at ifrance dot com) > I tried twice to subscribe, but did not receive any answer > > 2) I can not use news.gmane.org through thunderbird like other > squeak lists > Why? > > 3) I can not use http://blog.gmane.org/gmane.comp.lang.smalltalk.pharo.devel > It does not seem to be updated frequently enough. > > If someone could fix one of these problems, that would be super cool. > > Nicolas > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 10, 2008

Re: [Pharo-project] [ANN] 10091

by StÃ©phane Ducasse

Thanks this is great mail! Stef On Oct 9, 2008, at 10:18 PM, nicolas cellier wrote: > >> http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 > >> > >> Patches for Interval indexOf: and includes: > > > > I wonder about these fixes because of http://code.google.com/p/pharo/issues/detail?id=35 > > > > Did you reconsider the correctness of these fixes? They seem > > complicated and after reading through the bug reports again > > (especially http://bugs.squeak.org/view.php?id=7180) I concluded > > that they may introduce more problems than they fix. > I know I tried to find my way there too. > Now I trusted nicolas. > Nicolas are you strong on these fixes? > > Stef > > > > > > > Adrian > > _______________________________________________ > > Pharo-project mailing list > > Pharo-project at lists.gforge.inria.fr > > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > > > Thank you Stef. > Yes, the fix is safe. > I develop here after these 3 points > A)- The code was wrong > B)- Discussions were too long, but the fixes are very easy > C)- library could be enhanced > > Adrian, hope I will convince you (also, read the code). > > Cheers > > Nicolas > > > > > A) The code was wrong > > Prior this patch: > - #includes: implementation were WRONG. > - #indexOf: was missing, falling back to naive super slow > implementation > > Try these two tests in any Squeak image of your choice (3.7 to: 3.10): > > "There is no reason that the first is true while the second is false" > {(Float pi to: 4*Float pi by: Float pi) includes: Float pi*2 + > 1.0e-11. > (Float pi to: 4*Float pi by: Float pi) includes: Float pi*2 - > 1.0e-11}. > > "There is no reason for this timeToRun to grow" > (1 to: 7) collect: [:i | > [(1 to: 10000000) indexOf: (10 raisedTo: i)] timeToRun] > > > > B)- the fixes are very easy > > I simply implemented indexOf: with same algorithm as #includes: > (corrected!). > Then removed includes: because super includes: is implemented > sending indexOf: > Then I removed valueIncludes: because includes: was sole sender. > You see, two methods removed, one fix added: no revolution > > > > C)- library could be enhanced > > What I would prefer is: > - implement a method with an argument so that user can choose > comparison tolerance > - expand this fuzzy inclusion test to any other collection > Maybe something like: > includes: aNumber moreOrLess: aTolerance. > Then Interval includes: could eventually use a default tolerance. > But the obvious bugs must first be corrected... > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 10, 2008

[Pharo-project] Installer refactor

by Matthew Fulmer

I spent most of the day yesterday refactoring Installer. I split it up into 14 classes, so now it should be a bit easier to understand. I'll fix badly categorized methods and add some comments. -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

Oct. 10, 2008

[Pharo-project] using this mailing list

by nicolas cellier

Apologies for previous ugly formatted post... I have problems using this mailing list because: 1) I do not receive the posts directly (ncellier at ifrance dot com) I tried twice to subscribe, but did not receive any answer 2) I can not use news.gmane.org through thunderbird like other squeak lists Why? 3) I can not use http://blog.gmane.org/gmane.comp.lang.smalltalk.pharo.devel It does not seem to be updated frequently enough. If someone could fix one of these problems, that would be super cool. Nicolas

Oct. 9, 2008

[Pharo-project] [ANN] 10091

by nicolas cellier

>>/ http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 />>/ />>/ Patches for Interval indexOf: and includes: />/ />/ I wonder about these fixes because of http://code.google.com/p/pharo/issues/detail?id=35 />/ />/ Did you reconsider the correctness of these fixes? They seem />/ complicated and after reading through the bug reports again />/ (especially http://bugs.squeak.org/view.php?id=7180) I concluded />/ that they may introduce more problems than they fix. /I know I tried to find my way there too. Now I trusted nicolas. Nicolas are you strong on these fixes? Stef >/ />/ />/ Adrian />/ _______________________________________________ />/ Pharo-project mailing list />/ Pharo-project at lists.gforge.inria.fr <http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project> />/ http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project /> Thank you Stef. Yes/, the fix is safe. I develop here after these 3 points A)- The code was wrong ///B)- Discussions were too long, but the fixes are very easy //C)- library could be enhanced //Adrian, hope I will convince you (also, read the code). /// Cheers Nicolas A) ///The code was wrong// /Prior this patch: ///- #includes: implementation were WRONG. - #indexOf: was missing, falling back to naive super slow implementation /// Try these two tests in any Squeak image of your choice (3.7 to: 3.10): // "There is no reason that the first is true while the second is false" //{(Float pi to: 4*Float pi by: Float pi) includes: Float pi*2 + 1.0e-11. (Float pi to: 4*Float pi by: Float pi) includes: Float pi*2 - 1.0e-11}. / /"There is no reason for this timeToRun to grow" (1 to: 7) collect: [:i | [(1 to: 10000000) indexOf: (10 raisedTo: i)] timeToRun] ///// ////B)- the fixes are very easy//// I simply implemented indexOf: with same algorithm as #includes: (corrected!). Then removed includes: because super includes: is implemented sending indexOf: Then I removed valueIncludes: because includes: was sole sender. You see, two methods removed, one fix added: no revolution / /C)- library could be enhanced /What I would prefer is: - implement a method with an argument so that user can choose comparison tolerance - expand this fuzzy inclusion test to any other collection Maybe something like: includes: aNumber moreOrLess: aTolerance. Then Interval includes: could eventually use a default tolerance. But the obvious bugs must first be corrected... /

Oct. 9, 2008

Re: [Pharo-project] would be good to check the brekapoint behavior

by StÃ©phane Ducasse

do you know if this code in ok for pharo? Stef On Oct 9, 2008, at 8:40 PM, Marcus Denker wrote: > > On 09.10.2008, at 18:37, StÃ©phane Ducasse wrote: > >> Hi marcus >> >> you told me once that they was a breakpoint support in squeak. >> It would be great to have it because removing self halt in the code >> is a pain. >> > > There is a very simple breakpoint mechanism in Squeak. It allows to > set breakpoints > on method entry, and clear all at once. > > There is an enty in the menu in the browser to toggle the > breakpoint, attached is a > changeset that fixes some small things so that it actually works. > > > <BreakPointFixes.1.cs> > > > > The whole implementation is maybe a bit too simple, though. On the > other hand, it's a start. > > Marcus > > > -- > Marcus Denker -- denker(a)iam.unibe.ch > http://www.iam.unibe.ch/~denker > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 9, 2008

Re: [Pharo-project] [ANN] 10091

by StÃ©phane Ducasse

>> http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 >> >> Patches for Interval indexOf: and includes: > > I wonder about these fixes because of http://code.google.com/p/pharo/issues/detail?id=35 > > Did you reconsider the correctness of these fixes? They seem > complicated and after reading through the bug reports again > (especially http://bugs.squeak.org/view.php?id=7180) I concluded > that they may introduce more problems than they fix. I know I tried to find my way there too. Now I trusted nicolas. Nicolas are you strong on these fixes? Stef > > > Adrian > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 9, 2008

Re: [Pharo-project] would be good to check the brekapoint behavior

by Marcus Denker

On 09.10.2008, at 18:37, StÃ©phane Ducasse wrote: > Hi marcus > > you told me once that they was a breakpoint support in squeak. > It would be great to have it because removing self halt in the code > is a pain. > There is a very simple breakpoint mechanism in Squeak. It allows to set breakpoints on method entry, and clear all at once. There is an enty in the menu in the browser to toggle the breakpoint, attached is a changeset that fixes some small things so that it actually works. The whole implementation is maybe a bit too simple, though. On the other hand, it's a start. Marcus -- Marcus Denker -- denker(a)iam.unibe.ch http://www.iam.unibe.ch/~denker

Oct. 9, 2008

Re: [Pharo-project] [ANN] 10091

by Adrian Lienhard

Hi Stef, On Oct 9, 2008, at 19:05 , StÃ©phane Ducasse wrote: [..] > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 > > Patches for Interval indexOf: and includes: I wonder about these fixes because of http://code.google.com/p/pharo/issues/detail?id=35 Did you reconsider the correctness of these fixes? They seem complicated and after reading through the bug reports again (especially http://bugs.squeak.org/view.php?id=7180) I concluded that they may introduce more problems than they fix. Adrian

Oct. 9, 2008

Re: [Pharo-project] Monticello: an idealist question

by Matthew Fulmer

On Thu, Oct 09, 2008 at 09:22:53AM -0700, Matthew Fulmer wrote: > On Thu, Oct 09, 2008 at 10:44:34AM -0500, David Pennell wrote: > > VW runs has been on Linux since about 2001. > > Not for me. Both times I tried it (within the past year), it > gave an opaque error message while launching the VW installer, > and I couldn't figure out what to do, so I gave up. I tried it again, and it started installing stuff. I left before it finished. I'll see what happened tonight. -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

Oct. 9, 2008