May 2016 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] the fastest way to load external code

by Eliot Miranda

Hi Doru, > On May 21, 2016, at 12:50 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > > Hi, > > I am looking for ways to make loading of code from an st file faster. To this end, I would like to find in-image mechanisms that are not mandatory for obtaining a functional image. > > For example, one place I look at is SystemAnnouncer uniqueInstance, and I know that I should have no Nautilus open and that I can disable Renraku and GTExample: > > ReRuleManager uniqueInstance unsubscribe. > GTExampleOrganizer instance stopSystemAnnouncements. > > What else do you see around? Strikes me that parsing, not compiling and either holding onto the source String or simply keeping a pointer into the source file makes sense. Compilation to methods can wait until installation. With that approach you need no notifications at all. HTH _,,,^..^,,,_ (phone) > > Cheers, > Doru > > > -- > www.tudorgirba.com > www.feenk.com > > "Don't give to get. Just give." > > > > > > >

May 21, 2016

Re: [Pharo-dev] how to swallow a syntax error?

by Tudor Girba

Hi, Thanks! I tried before with catching Error, but of course SyntaxErrorNotification is a Notification :). This means I am tired and I should go to sleep. Thanks again for unlocking me. Cheers, Doru > On May 21, 2016, at 11:30 PM, Henrik Nergaard <henrik.nergaard(a)uia.no> wrote: > > Catching the Notification stops it from popping up the syntaxErrorDebugger. > > 'AAA.st' asFileReference readStreamDo: [ :s | > importer := CodeImporter fileStream: s. > > [ importer evaluateDeclarations ] on: SyntaxErrorNotification do: [ :e | â¦] > > ]. > > From: Pharo-dev [mailto:pharo-dev-bounces@lists.pharo.org] On Behalf Of Tudor Girba > Sent: Saturday, May 21, 2016 11:25 PM > To: Pharo Development List <pharo-dev(a)lists.pharo.org> > Subject: [Pharo-dev] how to swallow a syntax error? > > Hi, > > I am having trouble making the compiler not be interactive when it encounters a syntax error. > > To reproduce the problem I attach here an .st file that has a funny character in the method name. If I file this in, like: > âAAA.stâ asFileReference fileIn. > > I get a Syntax Error dialog. I would like to avoid that. > > After diving in, I noticed that I can tell the requestor of the CodeImporter to not be interactive, but that does not seem to have an effect: > 'AAA.st' asFileReference readStreamDo: [ :s | > importer := CodeImporter fileStream: s. > importer requestor interactive: false. > importer evaluateDeclarations ] > > I also tried to play on the end of the SyntaxErrorDebugger, and replace it with something that logs the exception, but I was not successful because that one is expected to return a string that can be compiled. > > Can anyone point me in the right direction? > > Cheers, > Doru > > -- > www.tudorgirba.com > www.feenk.com > > "Every now and then stop and ask yourself if the war you're fighting is the right one." > -- www.tudorgirba.com www.feenk.com "Don't give to get. Just give."

May 21, 2016

Re: [Pharo-dev] how to swallow a syntax error?

by Henrik Nergaard

Catching the Notification stops it from popping up the syntaxErrorDebugger. 'AAA.st' asFileReference readStreamDo: [ :s | importer := CodeImporter fileStream: s. [ importer evaluateDeclarations ] on: SyntaxErrorNotification do: [ :e | â¦] ]. From: Pharo-dev [mailto:pharo-dev-bounces@lists.pharo.org] On Behalf Of Tudor Girba Sent: Saturday, May 21, 2016 11:25 PM To: Pharo Development List <pharo-dev(a)lists.pharo.org> Subject: [Pharo-dev] how to swallow a syntax error? Hi, I am having trouble making the compiler not be interactive when it encounters a syntax error. To reproduce the problem I attach here an .st file that has a funny character in the method name. If I file this in, like: âAAA.st<http://aaa.st>â asFileReference fileIn. I get a Syntax Error dialog. I would like to avoid that. After diving in, I noticed that I can tell the requestor of the CodeImporter to not be interactive, but that does not seem to have an effect: 'AAA.st<http://aaa.st>' asFileReference readStreamDo: [ :s | importer := CodeImporter fileStream: s. importer requestor interactive: false. importer evaluateDeclarations ] I also tried to play on the end of the SyntaxErrorDebugger, and replace it with something that logs the exception, but I was not successful because that one is expected to return a string that can be compiled. Can anyone point me in the right direction? Cheers, Doru -- www.tudorgirba.com<http://www.tudorgirba.com> www.feenk.com<http://www.feenk.com> "Every now and then stop and ask yourself if the war you're fighting is the right one."

May 21, 2016

how to swallow a syntax error?

by Tudor Girba

Hi, I am having trouble making the compiler not be interactive when it encounters a syntax error. To reproduce the problem I attach here an .st file that has a funny character in the method name. If I file this in, like: âAAA.stâ asFileReference fileIn. I get a Syntax Error dialog. I would like to avoid that. After diving in, I noticed that I can tell the requestor of the CodeImporter to not be interactive, but that does not seem to have an effect: 'AAA.st' asFileReference readStreamDo: [ :s | importer := CodeImporter fileStream: s. importer requestor interactive: false. importer evaluateDeclarations ] I also tried to play on the end of the SyntaxErrorDebugger, and replace it with something that logs the exception, but I was not successful because that one is expected to return a string that can be compiled. Can anyone point me in the right direction? Cheers, Doru -- www.tudorgirba.com www.feenk.com "Every now and then stop and ask yourself if the war you're fighting is the right one."

May 21, 2016

Re: [Pharo-dev] the fastest way to load external code

by Tudor Girba

Hi, > On May 21, 2016, at 10:47 PM, Peter UhnÃ¡k <i.uhnak(a)gmail.com> wrote: > > > For example, one place I look at is SystemAnnouncer uniqueInstance, and I know that I should have no Nautilus open and that I can disable Renraku and GTExample: > > You can just do this and not worry about subscribers > > SystemAnnouncer uniqueInstance suspendAllWhile: Indeed. The problem with this is that RPackage will get out of sync. > (I was trying to speed up code generation so I did some benchmarks last year http://forum.world.st/fast-subclassing-class-creation-in-tests-tp4857779p48… ) Thanks for the pointer. > Also couple weeks back I found out that adding instance variable to a class is extremely slow and it's best to add them during the initial setup, but maybe that isn't a problem here. Indeed, I noticed that, too. Cheers, Doru > Peter > > > On Sat, May 21, 2016 at 9:50 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > Hi, > > I am looking for ways to make loading of code from an st file faster. To this end, I would like to find in-image mechanisms that are not mandatory for obtaining a functional image. > > For example, one place I look at is SystemAnnouncer uniqueInstance, and I know that I should have no Nautilus open and that I can disable Renraku and GTExample: > > ReRuleManager uniqueInstance unsubscribe. > GTExampleOrganizer instance stopSystemAnnouncements. > > What else do you see around? > > Cheers, > Doru > > > -- > www.tudorgirba.com > www.feenk.com > > "Don't give to get. Just give." > > > > > > > > -- www.tudorgirba.com www.feenk.com "It's not what we do that matters most, it's how we do it."

May 21, 2016

Re: [Pharo-dev] Random is not random at startup

by stepharo

I checked and CMWC Marsaglia is part of polymath PMNumberGenerator allSubclasses an OrderedCollection(PMBernoulliGenerator PMBinomialGenerator PMConstantGenerator PMExponentialGenerator PMGaussianGenerator PMPoissonGenerator) PMRandomGenerator allSubclasses an OrderedCollection(PMMarsagliaKissRandom PMLehmerRandomGenerator PMLinearCongruentialRandomGenerator PMMersenneTwisterRandomGenerator PMParkMillerMinimumRandomGenerator) So we could pick one of them for Pharo. Stef

May 21, 2016

Re: [Pharo-dev] Random is not random at startup

by stepharo

> > Though this only works on Unix, so on Windows it uses the current time > as a seed. It might get better results using #microsecondClockValue > instead of #millisecondClockValue. And I'd think about just taking the > microsecond value and sending it #hashMultiply instead of the weird > things it's doing now. (bitXor with the identity hash of the > UUIDGenerator instance? That depends on the randomness of identity > hashes, which may or may not be very good.) > It seems likely that Windows has a built-in random number generator, > as well, which would probably be better. > > While we're at it, the class Random is a Park-Miller generator, which > has quite poor randomness by modern standards. Most other Smalltalks > have upgraded -- GemStone uses CMWC (pure Smalltalk, very simple and > quite fast; I wrote that one), VW I believe uses Lagged Fibonacci, and > Squeak has moved to Mersenne Twister last I heard. I've tested all > three against statistical tests of randomness -- generated about 650GB > of random bytes from each, ran them through the tests (which require > that much data to get good statistics). > > All three are pretty good, only failing a few tests. The only > generator I tested that passed *all* tests was Linux /dev/urandom. > Which is also fast. If you're on Linux, and don't need a repeatable > random sequence, I'd use /dev/urandom. If you do need to be able to > have a repeatable sequence with good randomness properties, I'd use > one of the three generators I mentioned above. Of the three, the > simplest and fastest is CMWC. In Polymath random there are several alternate random generators but I do not if there is CMWC > > Regards, > > -Martin > >

May 21, 2016

Re: [Pharo-dev] the fastest way to load external code

by Peter UhnÃ¡k

> For example, one place I look at is SystemAnnouncer uniqueInstance, and I know that I should have no Nautilus open and that I can disable Renraku and GTExample: You can just do this and not worry about subscribers SystemAnnouncer uniqueInstance suspendAllWhile: (I was trying to speed up code generation so I did some benchmarks last year http://forum.world.st/fast-subclassing-class-creation-in-tests-tp4857779p48… ) Also couple weeks back I found out that adding instance variable to a class is extremely slow and it's best to add them during the initial setup, but maybe that isn't a problem here. Peter On Sat, May 21, 2016 at 9:50 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > Hi, > > I am looking for ways to make loading of code from an st file faster. To > this end, I would like to find in-image mechanisms that are not mandatory > for obtaining a functional image. > > For example, one place I look at is SystemAnnouncer uniqueInstance, and I > know that I should have no Nautilus open and that I can disable Renraku and > GTExample: > > ReRuleManager uniqueInstance unsubscribe. > GTExampleOrganizer instance stopSystemAnnouncements. > > What else do you see around? > > Cheers, > Doru > > > -- > www.tudorgirba.com > www.feenk.com > > "Don't give to get. Just give." > > > > > > > >

May 21, 2016

Re: [Pharo-dev] TEXT drag and drop

by stepharo

Le 21/5/16 Ã 16:02, Thierry Goubier a Ã©crit : > Le 21/05/2016 15:58, stepharo a Ã©crit : >> How can I see the effects of your changes? > > Install both packages that are in that smalltalkhub archive, I did that even before asking. > and try dragging a text selection. I have no idea how to do this. > > At least, that process works for PluggableTextMorph. > > Nice demonstration of both the drag and drop api and the text api, > because it's even nicely undoable in the editor :) > > Thierry > >> >> Stef >> >> >> Le 18/5/16 Ã 19:09, lb a Ã©crit : >>> Hi, >>> I have made a little tool for draging and dropping text selection. >>> like windows Word. >>> It can work in Pharo 40 50 and 60. >>> >>> Have a fun! >>> >>> LiangBing >>> >>> MCHttpRepository >>> location: >>> 'http://www.smalltalkhub.com/mc/liangbing/Text-DragDrop/main' >>> user: '' >>> password: '' >>> >>> >>> >> > > >

May 21, 2016

Re: [Pharo-dev] Should we updated UpdatedPharoByExample for Pharo 5?

by stepharo

but the current version is not about pharo40 but 50. Le 21/5/16 Ã 18:29, Cyril Ferlicot D. a Ã©crit : > Le 21/05/2016 15:56, stepharo a Ã©crit : >> Hi cyril >> >> >> I think that we should find the right balance between details that are >> specific to a version and stable parts. >> >> Now I target Pharo 50. I think that the focus of Pharo 60 should be on >> external (new tools) but on stabilisation, speed up and >> >> cleanup (and internal things like bootstrap). >> >> Stef >> > Ok, so if nobody disagree I think we should tag the current version as > Pharo 4. Mark the current Jenkins build as "Always keep artefact" and > publish a link on the book's page. > > I can do the two first steps but not the last. >

May 21, 2016