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

[Pharo-project] Some code! (was Re: Date fromString: '6-Jan-10')

by GÃ¶ran Krampe

Hi guys! I admit not having tracked this thread in detail BUT attached is a changeset that I dug out from the Squeak beginners list where I posted it waaay back. The added Date class method included below shows what it offers. And of course, there are TONS of more cool stuff it could handle, but it is meant to be "smart enough" for a reasonable set of cases. ...And it was an "hour hack" ... :) regards, GÃ¶ran ----------------------------- readFrom: inputStream pattern: pattern "Read a Date from the stream based on the pattern which can include the tokens: y = A year with 1-n digits yy = A year with 2 digits yyyy = A year with 4 digits m = A month with 1-n digits mm = A month with 2 digits d = A day with 1-n digits dd = A day with 2 digits ...and any other Strings inbetween. Representing $y, $m and $d is done using \y, \m and \d and slash itself with \\. Simple example patterns: 'yyyy-mm-dd' 'yyyymmdd' 'yy.mm.dd' 'y-m-d' A year given using only two decimals is considered to be >2000." | day month year patternStream char | patternStream := pattern readStream. [patternStream atEnd] whileFalse: [ inputStream atEnd ifTrue: [^nil]. char := patternStream next. char = $\ ifTrue: [inputStream next = patternStream next ifFalse: [^nil]] ifFalse: [ char = $y ifTrue: [ (patternStream nextMatchAll: 'yyy') ifTrue: [year := (inputStream next: 4) asInteger] ifFalse: [ (patternStream peekFor: $y) ifTrue: [ year := (inputStream next: 2) asInteger] ifFalse: [ year := Integer readFrom: inputStream]]] ifFalse: [ char = $m ifTrue: [ (patternStream peekFor: $m) ifTrue: [ month := (inputStream next: 2) asInteger] ifFalse: [ month := Integer readFrom: inputStream]] ifFalse: [ char = $d ifTrue: [ (patternStream peekFor: $d) ifTrue: [ day := (inputStream next: 2) asInteger] ifFalse: [ day := Integer readFrom: inputStream]] ifFalse: [ inputStream next = char ifFalse: [^nil]]]]]]. (year isNil | month isNil | day isNil) ifTrue: [^nil]. ^self year: year month: month day: day! !

Jan. 21, 2010

Re: [Pharo-project] A dog-slow browser

by Levente Uzonyi

On Thu, 21 Jan 2010, Mariano Martinez Peck wrote: > Thanks for the info Dave. I just built a new vm on the same machine from >> the svn sources + VMMaker-dtl.156 (3.11.12-2148). I saw the line about >> gnuification during make (which is absent with 3.11-3.2135). Here are the >> #tinyBenchmarks results: >> >> #( >> '388172858 bytecodes/sec; 10541379 sends/sec' >> '385542168 bytecodes/sec; 10590318 sends/sec' >> '385832705 bytecodes/sec; 10521930 sends/sec') >> >> So it's true, the binary on squeakvm.org is not gnuified. >> >> > If it is not too much effort, it would be really cool to have such binary in > the website. From the Pharo world, we would also need TrueType compiled :) I couldn't compile the FT2Plugin. CMake doesn't know the dependencies of that plugin. I tried installing libfreetype6-dev, but it didn't help. And if I (or someone else) build a vm, it will not be generally useful (this is built on Ubuntu 9.04, it may even break on Ubuntu 8.04), remember the UUIDPlugin issue? Btw, I'm sure Ian will build a new vm soon. Levente > > We are close to release 1.0, so having a fast VM would be REALLY appreciated > for all!! > > Thanks a lot, > > Mariano > > > > >> >> Levente >> >> >>> >>> Dave >>> >>> >>> _______________________________________________ >>> 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 >> >

Jan. 21, 2010

Re: [Pharo-project] A dog-slow browser

by Mariano Martinez Peck

Thanks for the info Dave. I just built a new vm on the same machine from > the svn sources + VMMaker-dtl.156 (3.11.12-2148). I saw the line about > gnuification during make (which is absent with 3.11-3.2135). Here are the > #tinyBenchmarks results: > > #( > '388172858 bytecodes/sec; 10541379 sends/sec' > '385542168 bytecodes/sec; 10590318 sends/sec' > '385832705 bytecodes/sec; 10521930 sends/sec') > > So it's true, the binary on squeakvm.org is not gnuified. > > If it is not too much effort, it would be really cool to have such binary in the website. From the Pharo world, we would also need TrueType compiled :) We are close to release 1.0, so having a fast VM would be REALLY appreciated for all!! Thanks a lot, Mariano > > Levente > > > > > > Dave > > > > > > _______________________________________________ > > 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 >

Jan. 21, 2010

Re: [Pharo-project] A dog-slow browser

by Levente Uzonyi

On Wed, 20 Jan 2010, David T. Lewis wrote: > On Wed, Jan 20, 2010 at 11:54:03PM +0100, Levente Uzonyi wrote: >> On Wed, 20 Jan 2010, Igor Stasenko wrote: >> >>>> 2010/1/20 Levente Uzonyi <leves(a)elte.hu>: >>>> >>>> I guess that vm is not gnuified. >>>> >>> >>> I installed one that was avail for linux, latest one. Didn't built it >> myself. >> >> I just checked on a "not-that-fast" machine with the binary versions from >> squeakvm.org: >> >> 3.11-3.2135 -> #( >> '204309656 bytecodes/sec; 7660022 sends/sec' >> '202852614 bytecodes/sec; 7722256 sends/sec' >> '205457463 bytecodes/sec; 7785510 sends/sec') >> >> 3.10.6 -> #( >> '380386329 bytecodes/sec; 9512738 sends/sec' >> '377025036 bytecodes/sec; 10174641 sends/sec' >> '378978534 bytecodes/sec; 10183725 sends/sec') >> >> The 3.11-3.2135 is not gnuified and I couldn't force cmake to gnuify it >> (though I didn't try really hard). I think this is an issue that should >> be solved. > > Levente, > > For a while after the unix VM build was converted to cmake, the gnuify > step was missing. This was fixed some time ago (I don't remember the date > exactly). If you get the latest code from the Subversion repository, it > should produce a gnuified VM when you do the build. > > I'm not sure if the 3.11-3.2135 is properly gnuified or not. If not, > then it's probably time for a new build on the web site. Thanks for the info Dave. I just built a new vm on the same machine from the svn sources + VMMaker-dtl.156 (3.11.12-2148). I saw the line about gnuification during make (which is absent with 3.11-3.2135). Here are the #tinyBenchmarks results: #( '388172858 bytecodes/sec; 10541379 sends/sec' '385542168 bytecodes/sec; 10590318 sends/sec' '385832705 bytecodes/sec; 10521930 sends/sec') So it's true, the binary on squeakvm.org is not gnuified. Levente > > Dave > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Jan. 21, 2010

Re: [Pharo-project] [ANN] ConfigurationOfRFB

by Miguel Enrique CobÃ¡ Martinez

El jue, 21-01-2010 a las 01:43 -0600, Miguel Enrique CobÃ¡ Martinez escribiÃ³: > I have uploade the Metacello configuration for RFB (a.k.a. Remote Frame > Buffer a.k.a. VNC) > > To load it in a 10506 PharoCore image: > > Gofer it > squeaksource: 'MetacelloRepository'; > package: 'ConfigurationOfRFB'; > load. > (Smalltalk at: #ConfigurationOfRFB) project latestVersion load. > > Enjoy. The correct invocation in a clean PharoCore 10506 image is: Gofer it squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfRFB'; load. ((Smalltalk at: #ConfigurationOfRFB) project perform: #latestVersion) load. because the latestVersion message doesn't exist in a clean image. Cheers -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 21, 2010

Re: [Pharo-project] about fast lineDo and...

by StÃ©phane Ducasse

> ok I will try like that. >> usually I check all the deltas in chronological order and skip what is not needed and wished. >> >> Stef > > Good old change set ? yes I use that from time to time. Stef

Jan. 21, 2010

Re: [Pharo-project] about fast lineDo and...

by Nicolas Cellier

2010/1/21 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: >> >> Yes but I do not work on >>>>> Â Â Â - new first class instance variable >>>>> Â Â Â - byecode loader >>>>> Â Â Â - new package to speed up everything >>>>> Â Â Â - scripting syntax >>>>> >>>> >>>> Never mind, when I said "you", it was not specifically you, but anyone >>>> else prefering to wait because of no free time. >>>> >>>>> so at the end I'm not sure what we lose but they are >>>>> simple actions that other people can do too to distribute the load. >>>>> So may be I will stop looking at squeak log. >>>>> >>>> >>>> Sure, the danger is to become a follower of random activity driven by others. >>>> If time is sparse, you might prefer to keep control. >>>> But some changes like file buffering really makes a difference. >>> >>> Yes but tracking changes you do not know is also taking time. >>> I started but so far failed two times because you have to go slowly not to break too much the system. >>> Tomorrow I will follow the advice of levente and check. >>> The problem also is that the level of the tools is not really good. >> >> That shouldn't be hard: in a updated Trunk image, file out MultiByteFileStream and StandardFileStream. Open a File Browser in your Pharo image select one of the files, right click and select "code-file browser". This will show you all the methods in the file and will diff it with the current versions (if it exists). The annotation pane shows the timestamp and information about the method (new, identical, modified, etc). > > ok I will try like that. > usually I check all the deltas in chronological order and skip what is not needed and wished. > > Stef Good old change set ? Nicolas > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Jan. 21, 2010

[Pharo-project] Mac VM 5.4b1 ships

by John M McIntosh

I've push the latest cocoa based 32/64bit & 64/64bit macintel/powerpc VM 5.4b1 This includes the logic for the microsecond clock, to call the clock the change set is attached. I've also built the 64/64bit VM for Intel that works with 64bit squeak images. Look for the folder experimental/64bits in the regular places found via http://smalltalkconsulting.com/squeak.html or ftp://ftp.smalltalkconsulting.com/experimental/64bit/ or http://homepage.mac.com/johnmci/FileSharing.html click to experimental, 64bits In a sub folder the file names are: Squeak 64-32 5.4b1.app.zip or Squeak 64-64 5.4b1.app.zip This *beta* VM runs in 32 or 64 bit mode on macIntel, and 32bit mode on powerpc to support Squeak 32bit images. Includes the freetype plugin, and requires os-x 10.5 or higher. The *beta* 64/64 VM runs only in 64 bit mode on macIntel to support Squeak 64bit images. It does not include FreeType at this time, and it requires os-x 10.5 or higher. -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

Jan. 21, 2010

Re: [Pharo-project] about fast lineDo and...

by StÃ©phane Ducasse

> > Yes but I do not work on >>>> - new first class instance variable >>>> - byecode loader >>>> - new package to speed up everything >>>> - scripting syntax >>>> >>> >>> Never mind, when I said "you", it was not specifically you, but anyone >>> else prefering to wait because of no free time. >>> >>>> so at the end I'm not sure what we lose but they are >>>> simple actions that other people can do too to distribute the load. >>>> So may be I will stop looking at squeak log. >>>> >>> >>> Sure, the danger is to become a follower of random activity driven by others. >>> If time is sparse, you might prefer to keep control. >>> But some changes like file buffering really makes a difference. >> >> Yes but tracking changes you do not know is also taking time. >> I started but so far failed two times because you have to go slowly not to break too much the system. >> Tomorrow I will follow the advice of levente and check. >> The problem also is that the level of the tools is not really good. > > That shouldn't be hard: in a updated Trunk image, file out MultiByteFileStream and StandardFileStream. Open a File Browser in your Pharo image select one of the files, right click and select "code-file browser". This will show you all the methods in the file and will diff it with the current versions (if it exists). The annotation pane shows the timestamp and information about the method (new, identical, modified, etc). ok I will try like that. usually I check all the deltas in chronological order and skip what is not needed and wished. Stef

Jan. 21, 2010

[Pharo-project] [ANN] ConfigurationOfRFB

by Miguel Enrique CobÃ¡ Martinez

I have uploade the Metacello configuration for RFB (a.k.a. Remote Frame Buffer a.k.a. VNC) To load it in a 10506 PharoCore image: Gofer it squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfRFB'; load. (Smalltalk at: #ConfigurationOfRFB) project latestVersion load. Enjoy. -- Miguel CobÃ¡ http://miguel.leugim.com.mx

Jan. 21, 2010