- Pharo-dev - lists.pharo.org

Re: [Pharo-dev] CompiledMethod>>hash can produce clashes

by Max Leske

> On 17.10.2014, at 02:46, Ben Coman <btc(a)openInWorld.com> wrote: > > Richard Sargent wrote: >> Eliot Miranda-2 wrote >>> On Wed, Oct 15, 2014 at 10:50 AM, Richard Sargent < >>> richard.sargent@ >>>> wrote: >>>> One of the best things about Smalltalk is how easily we can say what we >>>> mean. I think you would be better off creating a method named something >>>> like >>>> #hasSameEffectAs: to answer what you are presently using #= to do, and >>>> change #= to answer the, in my opinion, more sensible "is the same as" >>>> that >>>> we conventionally think of #= meaning. >>>> >>> But that's the point. #= has to mean something and having it mean #== >>> isn't useful, so one has to choose some value-based semantic for >>> CompiledMethod>>#= and the one that's there is useful. Defining what #= >>> means for some value type is far easier than defining what it might mean >>> for something as complex as a CompiledMethod. The definition in >>> Squeak/Pharo has been useful to me in implementing a closure-based system, >>> so I'm unapologetic about the current definition. It is a good one but it >>> doesn't preclude defining others. >> An interesting response. You ignored the point that e.g. #hasSameEffectAs: >> provides greater clarity and add an argument against something I didn't say. >> I also don't think defining equality for a CompiledMethod is particularly >> difficult. If I were to recompile a method's source code, I would get a new >> instance of a CompiledMethod that would, in my opinion, be equal to the one >> already installed in the class (and perhaps cached in the VM's >> optimizations). So one would be able to say that we would not replace an >> existing CompiledMethod with an equal one. The current implementation of #= >> has no such characteristic, since it proclaims a CompiledMethod named #a to >> be equal to one named #z. > > @Richard > > That doesn't seem to be a good example for what your trying to say. > Given... > > [1] SomeClass>>a "original instance" > ^1 > > [2] SomeClass>>a "recompiled instance" > ^1 > > [3] SomeClass>>z > ^1 > > ...you seem to be saying that its useful to know if [1]=[2], > but imply that is invalidated by [2]=[3] ? > > But [1]=[2] remains true, and just as useful for your example. > > > @Max > > I guess to call it a bug, you bumped into a different use case > where [2]=[3] is problematic. Can you describe that? Well, not problematic. Once you accept that neither selector nor class are part of a CompiledMethod it is obvious that two instances with the same byte codes produce the same hash. The actual problem is more one of understanding and use. The following code answers a collection with the CompiledMethods Collection>>add:, Collection>>do: and Collection>>remove:ifAbsent: Collection methods select: #isAbstract. All three CompiledMethods are implemented as â^ self subclassResponsibilityâ, so they have the same byte codes. Now, if you take that collection and make a set out of it youâll lose Collection>>do: since #do: and #add: produce the same hash, but #remove:ifAbsent: doesnât because the number of arguments is calculated into the hash (actually the CompiledMethod header is). So, as long as you think of CompiledMethods as objects that have a name, it looks like a bug and in my opinion this behaviour is something that messes with the mind of newcomers. Just a (silly) idea: something like a CompiledMethodWrapper might solve the problem (at least from the user perspective; everything is slightly different from the VM perspective :) ), as it could hold on to the class and the selector independently of the actual CompiledMethod. In the end however, one doesnât work with compiled methods a lot and the hash situation is unlikely to cause a lot of problems (people working with CompiledMethod usually know what they are doing). Cheers, Max > > > cheers -ben > > >> The blue book say #= means "Answer whether the receiver and the argument >> represent the same component." The current implementation does so only for >> some, in my opinion, counter-intuitive definition of "same component". >> -- >> View this message in context: http://forum.world.st/CompiledMethod-hash-can-produce-clashes-tp4784722p478… >> Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

Oct. 17, 2014

Re: [Pharo-dev] Playground and its current usability in Pharo4.0 Latest update: #40307

by Esteban Lorenzano

> On 16 Oct 2014, at 22:58, Torsten Bergmann <astares(a)gmx.de> wrote: > > Tudor wrote: >> Indeed. This was a discussion I wanted to spawn as well. I would prefer to change the name from Workspace to Playground in the World Menu. The reason is >that the Playground is such a distinct departure from the Workspace that it deserves a different name. > > I do not think it is a departure from the Workspace concept. > > Therefore I would keep "Workspace" for the menu item name and the window name. It would be enough if the initial > tab is called "Playground" because one can play in the code pane (or do serious stuff). > > This way we would keep up with a known concept (also from other IDE's), it meets what people would expect and > no books would have to be rewritten ;) > > BTW: Squeak had a goodie with something similar back in the days (cant remember the name). It was neat as one > could assemble different objects and different object representations in one navigateable window. > > >>> I would suggest to use the well known order "Cut","Copy","Paste". >> Thanks for noticing. This will be fixed. > > Thanks! > >> Indeed, we had a debate about the name. I proposed the solution 3b you mention, but it was decided >> that it would be too confusing to change the Cmd+i action at this time, and we chose "Go" to be >> the name of the action that is mapped on Cmd+g. > > Here we agree - I would have choosen 3b as well. Decided by whom? Did I miss the discussion in the > list or was it offlist? basically, it was decided by me (and the course of the actionsâ¦ I did not forced anyone :P). My reasons remain but I have to say: 1) Survey reflected a clear position towards what now is called 3b 2) We are still making testsâ¦ nothing is written in stone and we are looking for better ways to do thingsâ¦ so that can change again (and again, and again, and like that until we are happy) > >>> ...icons... >> This is a known issue and we will look at it soon. > > Sounds good. > >> At the moment, GT is an external project and the contributions should happen directly in its repository. > > Then please add me to the repo. Still the process is unclear - will I still provide a slice with > issue number or will I change the config. > > It is not very lucky that this process was started (without discussion/announcement/description first > and only for specific packages). > > Thx > T. >

Oct. 17, 2014

Re: [Pharo-dev] Playground and its current usability in Pharo4.0 Latest update: #40307

by Esteban Lorenzano

> On 16 Oct 2014, at 21:01, phil(a)highoctane.be wrote: > > I'd like to have the old workspace as it is for quick commands and as a kind of CLI where I do not need all the noise of the right part of the playground. > > And the playground for what it is meant to do. > > These two things are *not* for the same use case. yes, they should. If they are not then we have an error somewhere. Could you elaborate a bit on why they are not fulfilling same usecase for you? cheers, Esteban ps: In the future (not in pharo4, certainly), old workspace and inspectors will be removed from environment, so we need to make sure the new ones can do the job correctly. > > > âPhil

Oct. 17, 2014

Re: [Pharo-dev] PharoNOS

by kilon alios

so the advantage of using this is that is a slim down linux distro with Pharo included ? Why not just download a small linux distro and install Pharo ? On Fri, Oct 17, 2014 at 10:14 AM, mikefilonov <mikefilonov(a)gmail.com> wrote: > Thank you Torsten. Great ideas :) I made a to-do list for this project not > to > forget anything > https://trello.com/b/JXtcf2Ye/pharonos > > >Yes, I have the following quick setup scenario in mind: downlod the ISO, > start a virtual computer > >with it (in virtualization software or on a cloud server). Use the Pharo > based setup wizard (see above) > >to setup keyboard layout, give a fixed IP, adjust the clock, ... > >and download a premade Seaside (or other webframework image) from CI. Now > run on port 80, configure >another > >port and give a name to the machine in the network. This way one could > easily setup virtualized > >network machines to provide network delivered applications. > > Isn't it what Pharocloud is all about? ;) > > > > > -- > View this message in context: > http://forum.world.st/PharoNOS-tp4784982p4785075.html > Sent from the Pharo Smalltalk Developers mailing list archive at > Nabble.com. > >

Oct. 17, 2014

Re: [Pharo-dev] PharoNOS

by mikefilonov

Thank you Torsten. Great ideas :) I made a to-do list for this project not to forget anything https://trello.com/b/JXtcf2Ye/pharonos >Yes, I have the following quick setup scenario in mind: downlod the ISO, start a virtual computer >with it (in virtualization software or on a cloud server). Use the Pharo based setup wizard (see above) >to setup keyboard layout, give a fixed IP, adjust the clock, ... >and download a premade Seaside (or other webframework image) from CI. Now run on port 80, configure >another >port and give a name to the machine in the network. This way one could easily setup virtualized >network machines to provide network delivered applications. Isn't it what Pharocloud is all about? ;) -- View this message in context: http://forum.world.st/PharoNOS-tp4784982p4785075.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

Oct. 17, 2014

Re: [Pharo-dev] Mysteries of the compiler

by Eliot Miranda

On Thu, Oct 16, 2014 at 11:53 PM, Marcus Denker <marcus.denker(a)gmail.com> wrote: > On 16 Oct 2014 at 22:07:03, Nicolai Hess (nicolaihess(a)web.de) wrote: > > 2014-10-16 17:40 GMT+02:00 Marcus Denker <marcus.denker(a)inria.fr>: > >> >> >> >> While in the write we need to mark: >> >> >> >> lookupVariableForWrite: aVariableNode >> >> >> >> | var | >> >> >> >> var := scope lookupVar: aVariableNode name. >> >> >> >> var ifNil: [^var]. >> >> var isSpecialVariable ifTrue: [ self storeIntoSpecialVariable: >> aVariableNode ]. >> >> var isWritable ifFalse: [ self storeIntoReadOnlyVariable: >> aVariableNode ]. >> >> >> >> var isTemp ifTrue: [ >> >> "when in a loop, all writes escape" >> >> scope outerScope isInsideOptimizedLoop ifTrue: [ var >> markEscapingWrite ]. >> >> "else only escaping when they will end up in different >> closures" >> >> (var scope outerNotOptimizedScope ~= scope >> outerNotOptimizedScope) >> >> ifTrue: [ var markEscapingWrite]]. >> >> ^var >> >> >> >> >> >> I checked that all Opal tests that are failing are actually testing >> wrong, and I double checked >> >> that all those methods are now compiled like with the old old compilerâ¦ >> >> Now recompilng the image. >> > >> > Hmmâ¦ nopeâ¦. something wrong. >> > >> >> the âouterScopeâ was wrong. >> >> So without that, I can recompile the image and your example is compiled >> without temp vectorsâ¦ >> >> Marcus >> > > good! > I'll open an issue on fogbugz and create a slice with your changes. > > about reordering of tempvars: > It would be nice if the order preserves, but I think we can live with it. > > Yes, we should order them. But keep in mind that this does not mean that > you > > can access them via âat: indexâ. The index is âvirtualâ, so var3 can be > actually var2, > > because the *real* var2 lives together with var1 in a tempVector (which is > now var1). > > Knowledge has to come from somewhereâ¦ e.g. the DebuggerMethodMap builds up > > a description, while we (for now) just delegate to the AST (+scopes + > semantic vars). > > this means that > > #tempAt: > > sees always the real world. e.g. a method with 3 vars where they are all > in the tempVector, > > tempAt:2 will raise an error. > > namedTempAt: does what the old one did. > Not quite. What namedTempAt:[put:] do is access the temps in the order they appear in the list answered by ContextPart>>tempNames which is self debuggerMap tempNamesForContext: self. This is a flattened list that includes both direct and indirect temps. This is also the list that the Squeak debugger uses to populate the bottom-right context inspector in the debugger. namedTempAt:[put:] matches tempAt:[put:] until you get to an indirect temp. In a method the Squeak compiler always puts the indirect temps after all the direct temps, so the two match until the first indirect temp. But in a closure the situation is potentially much more complex because indirection vectors can be part of a closure's copied values, so it may have any number of indirection vectors copied form outer scopes, its own local direct temps and its own local indirect temps. The copied values get pushed after the arguments and before the local temps. So the ordering potentially gets complicated. This now can get confused in the Opal case with the changing order, which > we should fix. > > (and yes, we should check how much slower the AST bases scheme is for > accessing > > temps by offset... we could cache a representation similar to what the old > DebuggerMethodMap > > does). > Or use DebuggerMethodMap, rewriting the code to make it more comprehensible... Why reinvent the wheel? > > Although there is an issue with the debugger: > 14058 <https://pharo.fogbugz.com/default.asp?14058> Inconsistent > information in debugger > I solved that one by using tempvar names instead if the index, but > this has another issue (sometimes the last (or most nested) tempvar is > always nil). > > The real error is, EyeDebuggerContextInspector does recognizes if the > index changes > (somewhere in EyeDebuggerContextInspector>>#updateList it > compares the new elements (with new indices) with the old elements list > and skips the update) > > Okâ¦ that could be fixed with ordered temp names? > > Marcus > > > > -- best, Eliot

Oct. 17, 2014

Re: [Pharo-dev] Mysteries of the compiler

by Marcus Denker

On 16 Oct 2014 at 22:07:03, Nicolai Hess (nicolaihess(a)web.de) wrote: 2014-10-16 17:40 GMT+02:00 Marcus DenkerÂ <marcus.denker(a)inria.fr>: >> >> While in the write we need to mark: >> >> lookupVariableForWrite: aVariableNode >> >>Â Â Â | var | >> >>Â Â Â var := scope lookupVar: aVariableNode name. >> >>Â Â Â var ifNil: [^var]. >>Â Â Â var isSpecialVariable ifTrue: [ self storeIntoSpecialVariable: aVariableNode ]. >>Â Â Â var isWritable ifFalse: [ self storeIntoReadOnlyVariable: aVariableNode ]. >> >>Â Â Â var isTemp ifTrue: [ >>Â Â Â Â Â Â Â "when in a loop, all writes escape" >>Â Â Â Â Â Â Â scope outerScope isInsideOptimizedLoop ifTrue: [ var markEscapingWrite ]. >>Â Â Â Â Â Â Â "else only escaping when they will end up in different closures" >>Â Â Â Â Â Â Â (var scope outerNotOptimizedScope ~= scope outerNotOptimizedScope) >>Â Â Â Â Â Â Â Â Â Â Â ifTrue: [ var markEscapingWrite]]. >>Â Â Â ^var >> >> >> I checked that all Opal tests that are failing are actually testing wrong, and I double checked >> that all those methods are now compiled like with the old old compilerâ¦ >> Now recompilng the image. > > Hmmâ¦ nopeâ¦. something wrong. > the âouterScopeâ was wrong. So without that, I can recompile the image and your example is compiled without temp vectorsâ¦ Â Â Â Â Marcus good! I'll open an issue on fogbugz and create a slice with your changes. about reordering of tempvars: It would be nice if the order preserves, but I think we can live with it. Yes, we should order them. But keep in mind that this does not mean that you can access them via âat: indexâ. The index is âvirtualâ, so var3 can be actually var2, because the *real* var2 lives together with var1 in a tempVector (which is now var1). Knowledge has to come from somewhereâ¦ e.g. the DebuggerMethodMap builds up a description, while we (for now) just delegate to the AST (+scopes + sematic vars). this means thatÂ #tempAt: sees always the real world. e.g. a method with 3 vars where they are all in the tempVector, tempAt:2 will raise an error. namedTempAt: does what the old one did. This now can get confused in the Opal case with the changing order, which we should fix. (and yes, we should check how much slower the AST bases scheme is for accessing temps by offset... we could cache a representation similar to what the old DebuggerMethodMap does). Although there is an issue with the debugger: 14058Â Inconsistent information in debugger I solved that one by using tempvar names instead if the index, but this has another issue (sometimes the last (or most nested) tempvar is always nil). The real error is, EyeDebuggerContextInspector does recognizes if the index changes (somewhere in EyeDebuggerContextInspector>>#updateList it compares the new elements (with new indices) with the old elements list and skips the update) Okâ¦ that could be fixed with ordered temp names? Â Marcus

Oct. 17, 2014

New Spur trunk image available

by Eliot Miranda

Hi All, finally the Spur Squeak trunk image is updateable. The image in http://www.mirandabanda.org/files/Cog/SpurImages/2014-10-16/ was created today and thanks to Bert Freudenberg's latest Monticello work can be updated independently of the non-Spur trunk. Spur VMs are available in http://www.mirandabanda.org/files/Cog/VM/VM.r3105/ (and later as they appear). Without wanting to appear too overconfident the Spur system looks to be ready for use apart from image segments (which I hope to have working some time next month). I'm really interested in having this stress tested by as many people as possible. Spur really does offer a significant performance and functionality improvement over the current system, but it needs testing to ensure its reliability. Esteban Lorenzano is hard at work on the Pharo bootstrap for Spur so I hope Pharo 4 Spur will be available soon. As far as trunk goes, using Spur alongside non-Spur trunk is going to be difficult to manage for the near future. Right now, Spur modifies the Collections, Compiler, Kernel and System packages, and this is done by auto-editing the non-Spur versions of those packages, something I do periodically as new versions arrive. I also auto-edit trunk configurations (the part of the image update scheme that ensures packages are loaded in the right order when there are dependencies between packages) from non-Spur "update" to Spur "update.spur" forms. This at east means that Spur can keep up with trunk. But it does /not/ provide a way of committing to Collections.spur, Compiler.spur, Kernel.spur or System.spur without getting out of sync with non-Spur trunk. Note that apart from these packages, one /can/ safely commit any other package from a Spur image to trunk. Right now the plan is to release both V3 (the pre-Spur format) and Spur versions of Squeak 4.6 (I hope it'll be called Squeak 5.0). This isn't my preference. I'd like to see just Spur released, once reliability is verified. But I understand the safety and backward-compatibility concerns (Spur won't be able to load V3 image segments, and vice verse). The issue is of course that we have this tricky package situation to manage where, to keep the two systems in sync, modifications to Collections, Compiler, Kernel and System need to be committed from V3 and auto-edited to Spur. I think that's too clumsy to be practicable. Perhaps allowing the two systems to fork and doing a manual merge will be acceptable, but it'll be work to keep them in sync. -- best, Eliot

Oct. 17, 2014

Re: [Pharo-dev] PharoNOS

by phil＠highoctane.be

Le 17 oct. 2014 04:02, "mikefilonov" <mikefilonov(a)gmail.com> a Ã©crit : > > Hi Phil, > > I would really happy to know if this works for you. > > Considering this use case in mind do you have an idea how PharoNOS should > handle the Image? > Should it be embedded in ISO or should it be configured from clean Pharo on > each device? > > For instance, if it is really useful we can have a server which would > re-build the ISO with a requested image. > > What do you think on this? I would have an ssh enabled box and updare things with ansible so that I could update at ton of devices at once. Phil > > > > > -- > View this message in context: http://forum.world.st/PharoNOS-tp4784982p4785039.html > Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com. >

Oct. 17, 2014

Re: [Pharo-dev] PharoNOS

by Torsten Bergmann

Hi Mike, > Yes sure I can. So this PharoNOS is basically TinyCore Linux with pure X > server and Pharo running in fullscreen. Yes I already found out. > I have made some auto-installation scripts which configure the disk so no > setup wizard is needed. All space available is mounted under /mnt/universe/ > so you may use that folder to store stuff. Everything else is stored in > memory. Nice side effect of this is an ease of update as you may just > replace the ISO and you'll get the new version of everything (except of the > image of course). Mmmhh ... maybe it should initially boot a downsized "Image selector" image (that allows to download images from CI like Pharo Launcher). Something similar to the usual Linux install wizards to initially setup a machine, configure keyboard layout, clock/timezone, IP/DHCP ... but done in Smalltalk. OSProcess commands could be used underneath to configure the underlying Linux. If this "PharoNOS GraphicalInstall" supports downloading premade image(s) from CI similar to PharoLauncher one could easily configure the machine with for instance with a premade Seaside, Bootstrap, Moose, ... image. This image could also be the fallback to repair, ... > Image is chosen on start by reading the "/mnt/universe/image" file. You may > upload other image and change this file so after reboot it will start with a > new one. This info should be included in the docu in the pillarhub article (if not already) > Jan Tomsa has pointed out that the disk setup is quite dangerous > (https://twitter.com/DetektivTomy/status/522860736155238401) I plan to > update it later today with more safe solution which will not erase all the > data. Nice > Also I think Pharo should be run under root user so you can bind > to any port. Yes, I have the following quick setup scenario in mind: downlod the ISO, start a virtual computer with it (in virtualization software or on a cloud server). Use the Pharo based setup wizard (see above) to setup keyboard layout, give a fixed IP, adjust the clock, ... and download a premade Seaside (or other webframework image) from CI. Now run on port 80, configure another port and give a name to the machine in the network. This way one could easily setup virtualized network machines to provide network delivered applications. > I can also put more 3d-party libs like sqlite3, so we can have some > interesting apps for the terminals run on embedded Pharo :) However I need > some feedback on what is needed. sQLite3 would be cool, also unqlite. Both are small shared libs and embeddable. This way one could load the Native Boost bindings for either SQLite3 or PUnqlite and use an embedded RDBMS or a NoSQL Database and also load/use Glorp/Voyage or other to access it. > So that is it, all comments are welcome. A specific small image demonstrating Pharo 3 or 4 lively in the web browser also comes up as an idea. Emulating and bootstrapping an x86 machine on top of JavaScript is possible these days. Look at http://copy.sh/v86/ to bootstrap a linux in your browser. I tried to run the ISO with it, but it hangs. I guess it is due to query for network (which is not available in browser). Maybe a special non-network image could be built. This way we could include a premade machine in the Pharo Project webpage so people can directly bootstrap and play with Pharo. Similar to SqueakJS (http://bertfreudenberg.github.io/SqueakJS/) but without having to port the Pharo VM to JS. >I'll take a look at AioPlugin issue. How I can test if it is loaded well? I loaded OSProcess from the config browser and used (PipeableOSProcess command: 'pwd') output. While doing this the system lamented about the missing AioPlugin. Bye T.

Oct. 17, 2014