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

Re: [Pharo-project] object format of true, false, nil and others...

by Craig Latta

Hi-- I'll throw in a word here for "Dissolve", a feature of Spoon. It uses an alternate version of the garbage collector to treat as garbage all methods which haven't been run since a chosen point in time. With it I created the smallest Smalltalk object memory I've seen so far (1,337 bytes). For more info, see: http://netjam.org/spoon/smallest thanks, -C -- Craig Latta www.netjam.org/resume +31 6 2757 7177 + 1 415 287 3547

Feb. 18, 2012

Re: [Pharo-project] Now you can use Fuel with FileSystem

by Craig Latta

Hi Francois-- > Yep, Esteban's setup is a bit what I have now, rST and remote tools > are what I want. > > If we had a bounty system I would even pay for this. =) I'm available! :) I've written remote browsers and debuggers. -C -- Craig Latta www.netjam.org/resume +31 6 2757 7177 + 1 415 287 3547

Feb. 18, 2012

Re: [Pharo-project] HTTPS

by Sven Van Caekenberghe

Dale, On 18 Feb 2012, at 20:56, Dale Henrichs wrote: > Sven, et. al., > > I tried to use Zodiac the other day and I immediately had problem that the SqueakSSL plugin was not recognized (the primitiveSSLCreate primitive fails)... > > I am testing on a mac and I downloaded and copied the squeakssl bundle into the Resources folder of the Squeak vm (okay I tried doing this with three different vms and each time got the same primitive failure) ... with no luck ... > > I'd like to use Zodiac, but have fallen back to using libcurl (via OSProcess) because it just works.... > > I know that folks are able to get SqueakSSL to work, but the fact that someone has to manually monkey with their vm really means that it is not a simple matter to use Zodiac, unless one is "building their own vm" ... > > So I would second the notion to include SSL support in the distributed vms ... > > "https support is expected to work out of the box" these days ... > > Dale Of course it would be better to include the plugin by default, but even then, you would have to use a particular prepared vm. Here is mine (for Mac OS X): http://dl.dropbox.com/u/16235440/Pharo.app.zip I must say I don't understand: it is not monkeying with your VM, it is just adding a plugin, which is copying one file to the proper place. Using libcurl via OSProcess seems way more complicated for most people. YMMV, of course. BTW, did you try the one-click download, it was made just for this situation. HTH, Sven

Feb. 18, 2012

[Pharo-project] three PointerFinder failing tests fixed

by Pavel Krivanek

Hi, I created a patch that fixes three of currenty failing tests: http://code.google.com/p/pharo/issues/detail?id=5329 Cheers, -- Pavel

Feb. 18, 2012

Re: [Pharo-project] Putting source code in method trailer [WAS] Re: Speeding up Finder string search

by Craig Latta

Hi-- If you have some way of uniquely identifying a method, then you can put the source wherever you like (e.g., in some other live object memory), and look it up when you need it. For an example, see the MethodID class in Spoon[1]. -C [1] http://dl.dropbox.com/u/15188004/spoon/Spoon%203%20alpha%203.zip -- Craig Latta www.netjam.org/resume +31 6 2757 7177 + 1 415 287 3547

Feb. 18, 2012

Re: [Pharo-project] New IDE alternative (was Misc. newbie questions)

by Craig Latta

Hi-- > > (It makes me angry that I can sometimes feel more productive > > writing Ruby in Emacs than in Squeak/Pharo! Why can I not have BOTH > > a world-class text editor AND live in an image (when I choose to)?)) > > This is why I want Coral. Another approach is to just make Smalltalk appear as an entire filesystem. Then any editor can speak to it like any other hard disk. For an example (using WebDAV), see [1]. -C [1] http://tinyurl.com/7j26wk6 (thiscontext.wordpress.com) -- Craig Latta www.netjam.org/resume +31 6 2757 7177 + 1 415 287 3547

Feb. 18, 2012

Re: [Pharo-project] pharo vision

by Craig Latta

Nick writes: > At FOSDEM this weekend I was impressed with the progress Craig Latta > has made on spoon recently [1] [2]. Does his work on remote browsing > and micro kernels fit into a future Pharo? I hope so! By the way, for an inventory of the contents of the kernel I'm using now, see: http://thiscontext.wordpress.com/initial-spoon-modularization thanks, -C -- Craig Latta www.netjam.org/resume +31 6 2757 7177 + 1 415 287 3547

Feb. 18, 2012

Re: [Pharo-project] Nautilus with Refactorings

by StÃ©phane Ducasse

Impressive spirit! I like that when we push all in the same direction. Thanks again for this energy. Stef

Feb. 18, 2012

[Pharo-project] removing DataStreamTest?

by StÃ©phane Ducasse

Hi I was planning to fix the two tests so that the file is deleted in the teardown but I really wonder if there are testing anything. New issue 5321 by renggli: DataStreamTest creates files that are not deleted http://code.google.com/p/pharo/issues/detail?id=5321 DataStreamTest>>#testFileNamedtest DataStreamTest>>#testReadOnlyFileNamedtest testFileNamed | testString file | testString := 'testString'. file := ReferenceStream fileNamed: (self name, 'test'). file nextPut: testString. file close. file := ReferenceStream fileNamed: (self name, 'test'). self assert: file next = testString. file close. testReadOnlyFileNamed | testString file | testString := 'testString'. file := ReferenceStream fileNamed: (self name, 'test'). file nextPut: testString. file close. file := ReferenceStream readOnlyFileNamed: (self name, 'test'). self assert: file next = testString. file close.

Feb. 18, 2012

Re: [Pharo-project] Nautilus with Refactorings

by Benjamin

On Feb 18, 2012, at 9:44 PM, Mariano Martinez Peck wrote: > > > On Sat, Feb 18, 2012 at 9:20 PM, Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> wrote: > > On Feb 18, 2012, at 8:47 PM, Mariano Martinez Peck wrote: > >> >> >> On Sat, Feb 18, 2012 at 5:12 PM, Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> wrote: >> >> On Feb 18, 2012, at 11:22 AM, Mariano Martinez Peck wrote: >> >>> Hi Ben. What you did is very very cool. I was waiting for it since a lot of time. So...I had no excuses left to use it ;) You asked for feedback? Ok, I have a long pass. I hope you can get something useful from it. I would like to provide more help but this is all I can offer right now: >>> >>> >>> - When you are in a pane, say Class column, you press the right arow -> and the focus goes to the next column, so if you do down arrow now you start to move between methods. However, when you move from column, the first row (protocols in this case) of the right column is not highlighted so you don't really know if you are there or not until you start to move (like arrow down). So, I would autoselect/autohighlight first row of next column in that case. >> >> >> >> The selected list is surrounded by a light blue rectangle. I think the width of this border should be increase a bit, but it depends of the theme :) >> >> >> hahahahah ok...now I see it. But it is incredible complicated to see it ;) Indeed it would be much better if: a) the border is bigger b) why not to always auto select the first row? >> >>> >>> - file out in menu of protocols or methods does nothing -> if file out of single protocols is not supported it should be removed from the menu. >> >> On methods, it works for me :) >> >> true...I tried again and it works. Sorry ;) >> >> On protocols, I fixed it :) >> >> >> mmm did you commit? I have updated everything and still it doesn't work. MessageNotUnderstood: NautilusUI>>fileOutCategories:from: > > I added the method in AbstractTool and forget to post the new version :) > The new version has been proposed, now it has to be integrated > > > Excellent. > > >> >>> >>> - multiple select in protocols does not add methods in methods pane. The new protocol is just highlighted, its methods are shown in the method column, but we lost the previous one, even if the previous category is still selected. >> >> - Done :) >> >> >> Almost ;) >> Example: browse RBReadBeforeWrittenTester. select protocol 'accessing'. Now with ctrl and click select also 'initialize-release'. Ok. good now we have both selected. But now, click again in 'initialize-release', and you will see that instead of only letting 'accessing', instead you show ALL methods. > > Ok, I have seen the bug. But I have to finish the new Rayman before :) (or my gf will kill me ;o) ) > > > I don't know what Rayman is, but take care of your gf ;) http://www.youtube.com/watch?v=CsDKCPn7_Yk ;) Finished ^^ Ben > > >> >> >>> - when selecting several methods and try to remove them, do not ask each method individuallyâ¦ "it has xx senders.. what do you want to do?" â¦ give options according to the fact that there are several methods, but it only once. For example: "remove all and see all senders" "or just remove all". If you ask that for each methodâ¦then it is not a bulk remove but again, having to remove method by methodâ¦ >> >> I should, but it's some work :) >> >> >> Ok, no problem. Maybe we can open tickets/wishlists somewhere ? > > There is a Nautilus flag in the bug entry > >> >>> - Sometimes the icon for the tests appear even if those methods are not tests. Look attached file notTestMethods. >> >> Here, the test icon is displayed because there is a corresponding test. (In your example, it means that the method BehaviorTest>>#testAllMethods exists :) >> >> >> hehehehe ok. Funny. In my cases I think I have a really low percentage of mapping between methods names and the test for them. But its fine, maybe for someone is useful. Maybe a different color/icon from the one using for those which are real tests? > > If you use Nautilus to generate the test class and the test methods, the mapping is respected. > In a better world, instead of using bullshitty string comparison, we should be able to flag a method to bind it with its testing methods. > > > ok, I see. > >> >>> >>> - When you are showing extensions to a class, I would not show the whole class, but intead just the extensions from the package I am browsing. Look attached NotShowFullClass. >> >> First I thought about that too, but then I figured out that when you are writing an extension, it's really useful to be able to know the whole set of available methods in a glance :) >> >> >> Ok. I found it confusing at the beginning, but ok. Anyway, in that case, following to what happens in Classes, the protocolos which are not mine should be in gray and MINE in black. In my example, I mean to let black *Ast-Core and put gray to all the rest? > > Why not, I should experiment that ;) > >> >> >>> >>> - It would be very very cool if from the Class Search popup we have a shortcut to not change the current browse but instead open it in another one. >> >> This should be done on Class Search, but indeed, it will be cool ;) >> >> >> :) >> >>> >>> - When you are in the class column and you do right click and create a class, the class column where you are is not refreshed automatically and you don't see the just created class. You need to go to another package and then come back to see itâ¦.I guess you need to trigger a refresh or invalidate some cache. >> >> - Fixed :) >> >> >> Ok, I guess you didn't commit yet ;) > > I fucked up my source file, it was one of the lost changes. > > > hehehe > > >> >>> >>> - When viewing hierarchy package and class columns are swappedâ¦confusing. >> >> - Yes I know, but there are a set of good arguments behind ^^ >> >> >> hehee I am all ears > > As I said a couple of days ago: > > I know it's surprising, it's a tough discussion between Camillo and me :) > Basically, I didi it this way because when you browse the hierarchy of a class, the classes displayed do not belong to the same package. > So if you keep the same display, you will have a selected package in the first pane, and a set of classes not from this package in the second one. > And it's disturbing too. > > So in one hand it's surprising, but on the other hand, the other solution display wrong information. > In a nutshell, I am open to any solution :) > > > > What about using both panes (packages and classes) for the class hierarchy but then in each class show between parentesis the package they belong to. Example > > Object (Kernel) > TestCase (SUnit-Core-Kernel) > MyTest (MyPackage) > > > ? > > >> >>> >>> - Bug somehow using the left arrow "<". Find attached PharoDebugBug1.log >> >> Can't reproduce it :s >> >> >> I cannot either. It just happened once and I couldn't reproduce it :( >> >> >> >>> >>> - Bug: select a package. Right click, find package, cancel. DNU. Find attached PharoDebugBug2.log >> >> - Fixed >> >> >> Checked and works >> >>> >>> - Does it make sense to show "Run tests" when we are in a package where we have not even one single test? For example, go to the package 'Zinc-HTTP-Streaming' right click..and you will see Run testsâ¦ I would only let that option for the cases that you KNOW there are tests inside. >> >> - Done >> >> Checked nad works. >> >> One new problem. If I do "Run tests" to the package CompilerTests some of the classes are not colored as green, but if I click in the class, then I see inside that their test methods are green. > > Yep, I got the same. A problem of cache I guess > >> >> >>> >>> -In the code pane, the extended search doesn't work and instead it just opens the same menu again. THis is a petty because I use "method source with it" quite a lot ;) >> >> - A submenu has been added :) >> >> >> AWEOME :) >> >>> >>> - Bug: select the package 'Multilingual-Tests-TextConversion'. Find PharoDebug3.log attached. >> >> - Fixed too >> >> Tested and works. >> >> >>> >>> - When selecting a package (but no classes yet), wouldn't make sense to set the package name as the title in the windows rather than Nautilus? >> >> - Fixed :) >> >> >> >> Tested and works. >> >>> >>> >>> Well, that's all so far. It is a very cool browser and I plan to start using it as my default :) >>> Please continue!!! >>> >> >> >> >> Thanks A LOT for your feedback :) >> >> >> Thanks for your great work. Immediate answer and a bunch of fixes. Awesome. >> >> Continue! > > Til i get such good (and useful) feedback, for sure I will ;) > > > :) > > > > Ben > > >> >> >> >> >> Ben >> >> >> >> >> >> >>> >>> >>> >>> >>> >>> On Sat, Feb 18, 2012 at 10:54 AM, Andreas Wacknitz <a.wacknitz(a)gmx.de> wrote: >>> >>> Am 18.02.2012 um 01:40 schrieb Benjamin: >>> >>>> Hi guys, >>>> >>>> I am glad to announce that refactorings (RBEngine) are now integrated into Nautilus :) >>>> >>>> >>>> You can also found the latest image on Jenkins: https://ci.lille.inria.fr/pharo/job/Nautilus%20with%20Refactorings/ >>>> >>>> I hope you will enjoy it, and as usual, any feedback is welcome :) >>>> >>>> >>>> (the next step is to do a nice changes browser :) ) >>>> Ben >>> I downloaded the Jenkins build this morning. >>> When I select the "Announcement-Tests-Core" package I will get a DNU False>>ifEmpty: >>> >>> Nevertheless, keep on your impressive work! >>> >>> Regards >>> Andreas >>> >>> >>> >>> >>> -- >>> Mariano >>> http://marianopeck.wordpress.com >>> >>> <NautilusFeedback.zip> >> >> >> >> >> -- >> Mariano >> http://marianopeck.wordpress.com >> > > > > > -- > Mariano > http://marianopeck.wordpress.com >

Feb. 18, 2012