March 2015 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Bloc Example Errors

by Sean P. DeNigris

Pharo Smalltalk Developers mailing list wrote > thanks for your feedbacks ! Sure! Thanks for tackling this critical project :) I am excited to see what users will do with all the power of Morphic without being held back by our current implementation. I'll take a video of the clock... ----- Cheers, Sean -- View this message in context: http://forum.world.st/Bloc-Example-Errors-tp4810365p4810537.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

March 9, 2015

Re: [Pharo-dev] Canonical ConfigurationOfGitFileTree

by Sean P. DeNigris

> just let me take care of this one Great, thanks! ----- Cheers, Sean -- View this message in context: http://forum.world.st/Canonical-ConfigurationOfGitFileTree-tp4810448p481053… Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

March 9, 2015

Re: [Pharo-dev] [pharo-project/pharo-core] 842399: 40525

by Thierry Goubier

Le 08/03/2015 16:06, Ben Coman a Ã©crit : > > > On Mon, Mar 2, 2015 at 11:45 PM, Thierry Goubier > <thierry.goubier(a)gmail.com <mailto:thierry.goubier@gmail.com>> wrote: > > > > 2015-03-02 16:33 GMT+01:00 GitHub <noreply(a)github.com > <mailto:noreply@github.com>>: > > > Log Message: > ----------- > 40525 > 14997 Lazy MCVersionInfo to reduce MC memory use > https://pharo.fogbugz.com/f/cases/14997 > > > Please consider this one if you see things changing in the > Repository Inspector. > > Thanks, > > Thierry > > > Previous to build 40525, when you have a slice (and presumably any > package) that does not have any version loaded, all package versions > would show up as bold, meaning that they were not installed. > > Following build 40525, such unloaded packages no show without bold. > > For example, open the Pharo4Inbox and search for 14353 in both builds > 40524 and 40525. Loading Slice 14997 on top of build 40524 reproduces > the problem > > I find the new behaviour disconcerting :) > https://pharo.fogbugz.com/default.asp?15097 > > cheers -ben > > Thanks Ben for reporting that. From your description, it should be easy to fix. I'll have a look (it should be something in versionHighlight:). Thierry

March 8, 2015

Re: [Pharo-dev] Canonical ConfigurationOfGitFileTree

by Thierry Goubier

Le 08/03/2015 17:25, Sean P. DeNigris a Ã©crit : > The copy at MetaRepoFor40 is ahead of that at > http://smalltalkhub.com/mc/ThierryGoubier/Alt30/main/. It makes it hard to > give general loading instructions that depend on GitFileTree. Assuming it's > not hacked for Pharo 4.0 specifically, would someone copy it over or give me > write access? Thanks. Hi Sean, just let me take care of this one. The lazy version info in the latest pharo 4 has some significant impact on GitFileTree, and, given that github:// urls work correctly on all platforms on Pharo 4, I'll rewrite the configuration to call on the filetree repository. Thierry

March 8, 2015

Re: [Pharo-dev] Bloc Example Errors

by Alain Plantec

Hello Sean, thanks for your feedbacks ! Regarding the font errors (1, 4, 5 and 6), these are TxText related. TxText does no work with strike fonts. As a fix, load FT fonts and change you source code font setting. Then these examples should run. Regarding 2, it is fixed now. Regarding the clock, all seems ok here, I do not get what you means in fact. Thanks Alain > Le 8 mars 2015 Ã 15:00, Sean P. DeNigris <sean(a)clipperadams.com> a Ã©crit : > > On Mar 8, 2015, at 3:15 AM, stepharo [via Smalltalk] <[hidden email] <x-msg://3/user/SendEmail.jtp?type=node&node=4810414&i=0>> wrote: > > Load the latest version. > > Ah, yes! Much better :) Now I will play a bit! > > btw: > 1. exampleMorphWithDoubleClick -> asFreetypeFont 'not supported yet. and ever' > 2. exampleMorphWithStep -> MessageNotUnderstood: BlNullBorder>>width: > 3. exampleSimpleClock + Wiggling mouse over = strange drawing distortion > 4. exampleBevelBorder -> MessageNotUnderstood: StrikeFont>>getGlyphWidth: > 5. exampleMiterBorder -> same as #4 > 6. First 4 BlTxTextViewExamples fail with #1 or #4 > Cheers, > Sean > > View this message in context: Re: Bloc Example Errors <http://forum.world.st/Bloc-Example-Errors-tp4810365p4810414.html> > Sent from the Pharo Smalltalk Developers mailing list archive <http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html> at Nabble.com.

March 8, 2015

Re: [Pharo-dev] [Pharo4] Image size now 23.2MB

by stepharo

Hi eliot It is a cool extension I would really like to get startup more exposed. Stef Le 5/3/15 22:53, Eliot Miranda a Ã©crit : > > > On Thu, Mar 5, 2015 at 4:45 AM, phil(a)highoctane.be > <mailto:phil@highoctane.be> <phil(a)highoctane.be > <mailto:phil@highoctane.be>> wrote: > > On Thu, Mar 5, 2015 at 12:05 PM, Sven Van Caekenberghe > <sven(a)stfx.eu <mailto:sven@stfx.eu>> wrote: > > Phil, > > > >> On 05 Mar 2015, at 11:44, phil(a)highoctane.be > <mailto:phil@highoctane.be> wrote: > >> > >> Images are starting slow, that's my experience. Here, a Pharo3 on > >> Windows with quite a bunch of things in it takes six seconds to > load > >> on first launch and three after that. > >> > >> And that's on a i7 4770K clocked 3.85GHz with SSD drives. > >> > >> It takes less time to revive a VMWare VM :-( > > > > Maybe it is Windows ? Or certain extra stuff that you have in > your image that does work at startup ? Because you should see sub > second startup times. > > > > (This is on the smallest Digital Ocean instance, 1 virtual CPU, > 512 Mb RAM, SSD) > > > > root@stfx:~/experiments/pharo4# curl get.pharo.org/40+vm > <http://get.pharo.org/40+vm> | bash > > % Total % Received % Xferd Average Speed Time Time > Time Current > > Dload Upload Total Spent > Left Speed > > 100 2885 100 2885 0 0 17974 0 --:--:-- --:--:-- > --:--:-- 18031 > > Downloading the latest 40 Image: > > http://files.pharo.org/image/40/latest.zip > > Pharo.image > > Downloading the latest pharoVM: > > http://files.pharo.org/vm/pharo/linux/stable.zip > > pharo-vm/pharo > > Downloading PharoV10.sources: > > http://files.pharo.org/sources//PharoV10.sources.zip > > Downloading PharoV20.sources: > > http://files.pharo.org/sources//PharoV20.sources.zip > > Downloading PharoV30.sources: > > http://files.pharo.org/sources//PharoV30.sources.zip > > Creating starter scripts pharo and pharo-ui > > > > root@stfx:~/experiments/pharo4# ./pharo Pharo.image printVersion > > [version] 4.0 #40535 > > > > root@stfx:~/experiments/pharo4# ls -lah Pharo.image > > -rw-rw-r-- 1 root root 22M Mar 4 14:51 Pharo.image > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval > '100 factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.386s > > user 0m0.184s > > sys 0m0.028s > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval > '100 factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.422s > > user 0m0.192s > > sys 0m0.032s > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval > '100 factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.863s > > user 0m0.404s > > sys 0m0.064s > > > > It is the latest (smaller) image, but that should not make much > difference. > > On my CentOS Linux, headless tests are fast as well. > > Granted, my Windows image was super large. 300Megs or something. > I cleaned the Monticello things with flush cache, which removed 41 > megs. > > I can't test the CLI version on Windows, as I do get a window opening. > > There are a ton of windows and morphs open in there and I've run a > couple of experiments. > > Now, the Linux image was large too. 200Megs or so. > And the window showed immediately, then it took a second and a half or > so to show the images back. > There is a black background before. > Maybe the system is processing its startUp list, unhibernating forms, > I don't know what's going on but it takes that 1.5 secs. > > The image gets large because of Monticello caches, and TWM (Tiling > Window Manager) keeping a handle on browsers etc over time. So, I > cleaned the caches, closed TWM, gc'ed a time or two and then got the > image back to 90 megs. > > Then the image opens in like 0.5 sec. It is hard to measure as what I > want is the time between issuing the pharo-ui XYZ.image and the fact > that I can actually see the UI. > > How would one do that? > > Sorry for the rant but generally speaking, the Pharo UI seems slower > than the Squeak 4.5 ui. > Or the EToys UI for that matter. > I am using those for little educational programs and they feel faster. > I'd like to know why. Both systems have morphic. Is it because of > aliased fonts? Squeak 4.5 also has that. > > Is there any way to monitor what's going on during image startup? > > > I modifier the AndreasSystemProfiler in the Squeak code base at > Cadence to allow profiling across snapshot; you see both save and > startup profiles. That's not showing e.g. disc activity from VM > reads, but it does help track down tike spent in the SMalltalk > start-up. IIRC this is in te Saueak code base. But if you're > interested I can dig it out. The changes are minor; to do with > correctly offsetting delays across snapshot. > > > Phil > > > > > > > > > Sven > > > > > > -- > > Sven Van Caekenberghe > > Proudly supporting Pharo > > http://pharo.org > > http://association.pharo.org > > http://consortium.pharo.org > > > > > > > > > > > > > -- > best, > Eliot

March 8, 2015

Re: [Pharo-dev] [Pharo4][Slots] Small progress of the day: Example 1st class Globals

by stepharo

marcus I would love to see how we can use slots to have first class caches as described in guillermo paper. Why do we need to have a reflective call when it is not needed? Could not we late bound such behavior? Stef Le 5/3/15 18:20, Marcus Denker a Ã©crit : > Hi, > > The Slots project contains not only Slots, but in addition âFirst classâ Globals, too, > > Pharo4 now contains an example: > > ExampleClassVariable, this class defines the ivar (slot) âstateâ. > > read > ^state > > write: anObject > state := anObject > > we can create a class defining a class var like that: > > Object subclass: #TT > slots: { } > classVariables: { #ClassVar => ExampleClassVariable } > category: âPlayGround' > > > Like with slots, the variable reads and write generate code to call the reflective > read and write methods: > > 21 <20> pushConstant: ##ClassVar > 22 <D1> send: read > 23 <7C> returnTop > > and the write: > > 21 <10> pushTemp: 0 > 22 <69> popIntoTemp: 1 > 23 <20> pushConstant: ##ClassVar > 24 <11> pushTemp: 1 > 25 <E1> send: write: > 26 <87> pop > 27 <78> returnSelf > > Full list of changes: > > - cleanup: moving all Reflectivity related code from Slot package to Reflectivity package > - fix code generation to not use Object>>#writeToGlobal: but instead the idea with the temp as seen above > - add example for Class Variable > > > Marcus >

March 8, 2015

Re: [Pharo-dev] Update goes havoc

by stepharo

This is why I would love that we pay attetion that the debugger is hyper stable. We should bullet proof it. Do you know how we can reproduce it? Stef Le 4/3/15 14:18, Norbert Hartl a Ã©crit : > I got this emergency evaluator while doing "Software update" from the > menu. Pretty harsh, no? :) >

March 8, 2015

Re: [Pharo-dev] Suggestions with Smart hovering

by stepharo

Me too :) It should not be that difficult but we are busy with other dev. Now alain did some fun handler in blocks changing the hand. Stef Le 4/3/15 12:26, Torsten Bergmann a Ã©crit : > Would like to see something like this in future Pharo > for suggestions as in this animated gif: > > http://forum.world.st/file/n4809356/smart-hover-1.gif > > >

March 8, 2015

Re: [Pharo-dev] [OT] Swift: Madness of Generic Integer

by stepharo

> > That said, I once considered the possibility of having a > PositiveSmallInteger and a NegativeSmallInteger just for the sake of > unifying the sign/magnitude handling. I would really like to see if this is better, because I have the impression that it could take advantage of dispatch. Stef > > But, let's imagine that it can pay.. > We can also transpose this to collections - ByteArray > ShortIntegerArray WordArray with signed/unsigned variants - > without resorting to VM considerations... > Are we sure that Smalltalk implementation (image side) would be > simpler than Swift with all the double dispatching machinery involved? > > So, if the goal of having such space/time optimization is founded > (which we can't know without context), I would say Swift does not > do it that bad... > > Some languages are much worse, like say: > > https://www.securecoding.cert.org/confluence/display/c/INT32-C.+Ensure+that… > > > ;) > > Thx for the share. Enlightment doesn't come easy. > > Remember that a lot of code in this world is still COBOL... > > Phil > > > > > Sven > > > > >

March 8, 2015