February 2012 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] a fix to KomServer

by GÃ¶ran Krampe

On 02/14/2012 11:13 AM, Milan Mimica wrote: > Any feedback would make me feel cozy inside on valentine's :) > Tell me someting! Is it good? Completelly wrong? Am I supposed to fix > things? Do you accept patches? Wrong mailing list? It was a while since I helped with maintaining Kom and I am a bit short on time. The SS repo is open for global write - to make it easier to contribute :) And of course we accept and want patches :) If noone else steps up and help you out I might squeeze in some time during the week - you should probably cross post to squeak-dev and seaside lists. regards, GÃ¶ran

Feb. 14, 2012

Re: [Pharo-project] a fix to KomServer

by Milan Mimica

Any feedback would make me feel cozy inside on valentine's :) Tell me someting! Is it good? Completelly wrong? Am I supposed to fix things? Do you accept patches? Wrong mailing list? On 14 January 2012 13:47, Milan Mimica <milan.mimica(a)gmail.com> wrote: > oops, looks like I have committed something... > I have committed a patch to http://www.squeaksource.com/KomHttpServer > I wanted to have it checked by someone first, didn't really think it would > let me commit just like that. Sorry for the inconvenience. > > Anyway, now that it's there someone has to look at it. :) It's not much. > The problem was KomServer threw MNO when a request was made with an > absolute URL in the HTTP request, which is valid as I read it [1][2]. > > > [1] http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.2.1 > [2] http://www.w3.org/Protocols/rfc2616/rfc2616-sec5.html#sec5.2 > > -- > Milan Mimica > http://sparklet.sf.net > -- Milan Mimica http://sparklet.sf.net

Feb. 14, 2012

Re: [Pharo-project] [BUG]FileDialogWindow(Object)>>doesNotUnderstand: #directoriesFor:

by Alexis Parseghian

On Mon, Feb 13, 2012 at 8:30 PM, Alexis Parseghian <a.parseg(a)gmail.com> wrote: > At first, trying to load text into a workspace (from its menu). More > generally, any attempt at opening the file dialog. UITheme > class>>exampleDialogs fails with the same. > > Still worked in update #14284 (previous images I had). http://code.google.com/p/pharo/issues/detail?id=5292

Feb. 14, 2012

Re: [Pharo-project] [BUG]FSReference(Object)>>doesNotUnderstand: #asLowercase

by Alexis Parseghian

On Tue, Feb 14, 2012 at 3:04 AM, Camillo Bruni <camillobruni(a)gmail.com> wrote: > open a bug report, I'll fix it :) > > we're getting FSFilesystem slowly into the system... > the side effects are that some tools might not completely work :/ http://code.google.com/p/pharo/issues/detail?id=5291

Feb. 14, 2012

Re: [Pharo-project] bug? Lazy Class creation via method saving not stored in recent changes log

by Max Bareis

Hi Stef, additionally Renaming is also stored only partially in changes log. save your image rename the ClassB of the former example to ClassC via ctrl-click menu. check recover lost changes There you should see the change in the method but not the class rename. Regards Max Am 14.02.2012 um 10:08 schrieb Max Bareis: > Hi Stef, > > this is also easy to create with the same steps in 1.4a 14329 > > save your image > create ClassA > create ClassA>>methodA with: > ClassB methodB > save it choose define new class press OK on the Information Required Box and check the recover lost changes. > > There should be one entry regarding the method change but no entry for the class creation > > Regards > > Max > > Am 13.02.2012 um 17:02 schrieb StÃ©phane Ducasse: > >> thanks Max >> >> Yes this is probably a problem with the notification of class definition. >> Do you have a step by step reproduce able case for 1.4. >> >> Stef >> >> On Feb 13, 2012, at 4:34 PM, Max Bareis wrote: >> >>> Hi, >>> >>> I have encountered the issue, that the class creation is not stored in recent changes log when I process the following steps (ClassB does not exist): >>> >>> >>> ClassA>>methodA >>> >>> ClassB methodB. >>> >>> save. >>> >>> Now a popup appears and offers the creation of the ClassB. After accepting the class is created and visible in the System Browser, but there is no trace of the class creation in recent changes log. The method change is stored in the log. This affects Pharo 1.3 13315 OneClick >>> >>> Regards >>> >>> Max >>> >> >> > >

Feb. 14, 2012

Re: [Pharo-project] Are Objects really hard?

by Sven Van Caekenberghe

On 14 Feb 2012, at 09:41, Stefan Marr wrote: > While you are at it, add PHP and Ruby into your benchmarks. > You will see that the memory consumption of PHP is just outrageous. > And the speed of Ruby is legendary... to humor me, please use Ruby 1.8 and compare it to 1.9. > > Look at what people write about production setups for these languages. Your approach just doesn't scale. > The bottom line is, the quality of the language implementation for real-world applications just does not matter at all. People developed the right tools for the job... Stefan, you are absolutely 100% right: Ruby, Python, PHP, Perl, JavaScript are successful and scale well even though they often have less efficient runtimes than Smalltalk. Choosing a language in which you are productive is a better strategy. Sven

Feb. 14, 2012

Re: [Pharo-project] bug? Lazy Class creation via method saving not stored in recent changes log

by Max Bareis

Hi Stef, this is also easy to create with the same steps in 1.4a 14329 save your image create ClassA create ClassA>>methodA with: ClassB methodB save it choose define new class press OK on the Information Required Box and check the recover lost changes. There should be one entry regarding the method change but no entry for the class creation Regards Max Am 13.02.2012 um 17:02 schrieb StÃ©phane Ducasse: > thanks Max > > Yes this is probably a problem with the notification of class definition. > Do you have a step by step reproduce able case for 1.4. > > Stef > > On Feb 13, 2012, at 4:34 PM, Max Bareis wrote: > >> Hi, >> >> I have encountered the issue, that the class creation is not stored in recent changes log when I process the following steps (ClassB does not exist): >> >> >> ClassA>>methodA >> >> ClassB methodB. >> >> save. >> >> Now a popup appears and offers the creation of the ClassB. After accepting the class is created and visible in the System Browser, but there is no trace of the class creation in recent changes log. The method change is stored in the log. This affects Pharo 1.3 13315 OneClick >> >> Regards >> >> Max >> > >

Feb. 14, 2012

Re: [Pharo-project] Are Objects really hard?

by Stefan Marr

On 14 Feb 2012, at 01:57, Guido Stepken wrote: > Amazon Cloud even has widely installed "memcached", hat works also as httpd cache, preventing any Apache webserver and indirect Pharo Webserver from "seeing" any http/get request. > > So, i am NOT surprised, that any of those Pharoers doesn't want me here ... Pharo *is* dogsloooooow, Seaside nearly *unusable*. Unusable for any serious business application, unforseeable memory pumping effects, expecially in vservers with limited or even dynamically assigned memory. > > I will open Benchmarks later ... While you are at it, add PHP and Ruby into your benchmarks. You will see that the memory consumption of PHP is just outrageous. And the speed of Ruby is legendary... to humor me, please use Ruby 1.8 and compare it to 1.9. Look at what people write about production setups for these languages. Your approach just doesn't scale. The bottom line is, the quality of the language implementation for real-world applications just does not matter at all. People developed the right tools for the job... Regards Stefan -- Stefan Marr Software Languages Lab Vrije Universiteit Brussel Pleinlaan 2 / B-1050 Brussels / Belgium http://soft.vub.ac.be/~smarr Phone: +32 2 629 2974 Fax: +32 2 629 3525

Feb. 14, 2012

Re: [Pharo-project] [BUG]UndefinedObject(Object)>>doesNotUnderstand: #asSocketAddress

by Marcus Denker

On Feb 13, 2012, at 8:58 PM, Sven Van Caekenberghe wrote: > Thanks for the bug report, Alexis. > > Part of the solution is to change the last line of NetNameResolver class>>addressForName:timeout: to > > ^ result ifNotNil: [ result asSocketAddress ] > > But I don't like NameLookupFailure>>#defaultAction either. > > Apart from the fact that this is an extremely silly interactive action, it should call super when aborting. Retry feels very old school and ancient, it won't make a difference. > > Furthermore, the contract of NetNameResolver class>>addressForName:timeout: should better not include returning nil, signaling NameLookupFailure is enough, but that might break some older code. > > Comments ? > We should remove the "send bug report by mail" feature. We will later add a way to add bug reports to the bug tracker within pharo. Using email for this is not a good idea these days. Marcus -- Marcus Denker -- http://marcusdenker.de

Feb. 14, 2012

Re: [Pharo-project] Are Objects really hard?

by Marcus Denker

> > I've got best connections to top 500 enterprise and deciders in germany. Sure. :-) > I would be pleased to present them a Seaside Shop, running on Pharo, PostgreSql or Magma behind, ready to run from within one image. > > > Pharo is far from what we dream about but we are serious about it and we are concerned by business. > > > > Now your trolling is good because we are not an obscure community without a troll but we have one troll > > so this is good. Thanks. > > Ok. If you are not able to get Pharo development managed under consideration of business needs, means: benchmarks and stress tests equiv to sunit tests, please step back as project leader! > Now you are getting serious. My counter offer: We all decide to ban you from this mailinglist. > > If you are positive and provide information that is not just bullshitting then we will consider to reply and consider your request else *plonk* > > I am positive, still, although you cost me much money. You are *not* positive. You are a Troll. Please unsubscribe. Marcus -- Marcus Denker -- http://marcusdenker.de

Feb. 14, 2012