Squeak-dev December 2008

squeak-dev@lists.squeakfoundation.org

91 participants
114 discussions

Re: [ANN] ConflictFinder (help needed!) (Andreas Raab)
by stephan＠stack.nl 18 Dec '08

18 Dec '08

> You miss KomHttpServer, KomServices and DynamicBindings. Yep, I noticed. But SUTreeReport misses uids in renderContentOn and uid:on: and WAResponseTest misses response in testForbidden and testNotFound Stephan

2 1

Re: [ANN] ConflictFinder (help needed!) (Andreas Raab)
by stephan＠stack.nl 18 Dec '08

18 Dec '08

Quoting squeak-dev-request(a)lists.squeakfoundation.org: > 20. Re: [ANN] ConflictFinder (help needed!) (Andreas Raab) > > Thanks. Anyone who is interested can try running this first: So, how do you want to receive results? First experiment (no Universe): ConflictFinder findConflictsIn:{ 'Seaside' -> [MczInstaller installFileNamed: 'D:\conflicttest\Pharo0.1-10074-oneclick.app\Seaside2.8a1-pmm.581.mcz']. 'Scriptaculous' -> [MczInstaller installFileNamed: 'D:\conflicttest\Pharo0.1-10074-oneclick.app\Scriptaculous-lr.243.mcz']. 'ShoreComponents' -> [MczInstaller installFileNamed: 'D:\conflicttest\Pharo0.1-10074-oneclick.app\ShoreComponents-3-pk.39.mcz']. }. Results in: Analysing base system ... done. Loading Seaside ... (response is Undeclared) (response is Undeclared) (response is Undeclared) (response is Undeclared) (response is Undeclared) (response is Undeclared) (HttpService is Undeclared) (HttpResponse is Undeclared) (HttpResponse is Undeclared) (HttpResponse is Undeclared) (HttpResponse is Undeclared) (HttpRequest is Undeclared) (HttpRequest is Undeclared) done. Loading Scriptaculous ... (uids is Undeclared) (uids is Undeclared) done. Loading ShoreComponents ... done. Analysis complete. Stephan

3 2

3.11 deliverable no.1
by Keith Hodges 18 Dec '08

18 Dec '08

Hello All, The first result of the 3.11 effort is going to be an image that I have called 3.10bc. Where bc stands for "Build Candidate". The purpose of this image is to provide a base starting point for everyone who would like to contribute to 3.11, and do image building activities of any kind. 3.10bc is the first "fixed point", being essentially the stable 3.10 release, against which small projects can be carried out safely. To if you were to test a bug fix, and I want to reproduce it, it will be a case of taking 3.10bc and repeating your test. 3.10bc consists of, 3.10.2 + LevelPlayingField (Monticello, PackageInfo, MonticelloConfigurations, Monticello-Files, Installer-Launcher), and the latest of AtomicLoading, Universes, SqueakMap, Sake/Packages, SUnit, and Tasks, together with some essential fixes which support the above packages. Full details of the mantis bug reports for those fixes are included in the image. This "base" is generated and should be easily reproducible for any of the images supported by LevelPlayingField. Some tools like SUnitGUI are not backwardly compatible, BUT SUnit includes a non GUI TestReporter which is. Projects which can start based upon 3.10bc A) Bug Harvesting All the outstanding bugs on mantis can be assessed, in their respective categories. B) Image Tidying C) Image wide adjustments renaming etc. e.g. _ to := revisited. == to = Transcript show -> self logCR: D) Module Definition and Removal Move classes into a category, make them into a monticello package, move methods to become extension methods, and publish the package. E) Validation of external packages, and getting more stuff to work And many more I am sure. 3.10bc+fixes, from A above, will be a continuously moving target, for those intending to build on top of 3.11, so instead we provide a 3.11bc this will be 3.10bc + tidying, reorg and removals. But Not Bug Fixes. This is an image that may be used as a second fixed point if preferred. back to the day job for a few days for me I am afraid Keith

2 2

alien os-x plugin goes to 1.1.0
by John M McIntosh 18 Dec '08

18 Dec '08

I've updated the os-x alien plugin for powerpc and macintel, it is found on the usual sites via http://www.smalltalkconsulting.com/squeak.html in the Experimental directory as IA32ABI.1.1.0.bundle.zip The macintel flavour supports the ability to call foreign entry points with any type of data and the ability to callback to smalltalk code from the entered routine, plus the ability to get/set bytes, shorts, longs, float, and doubles. The powerpc flavour supports the ability to call foreign entry points with any type of data, and the ability to get/set bytes, shorts, longs, float, and doubles. Hopefully tomorrow I can look into the callback logic and propose some code/ assembler to enable that feature. It does btw run on 10.4 powerpc G3 machines now. There was a number of structural changes required enable the typeless Alien api to understand it must work with a floating point signature array that indicates how to deal with Float or Double data. I also added the ability to pass not only Alien and Integer data but also to pass Float/ Double objects as parm data. I note Squeak maintains floating point numbers as a Double. If you pass that as a Float by indicating that in the signature data (4 bytes versus 8 bytes) we convert the number from a double to a float before passing to the routine. On return the Alien api makes you explicitly indicate if the value is a double, integer, or float. Users should be aware it's possible to call a routine that returns a double but ask for a Float back. This might lead to unexpected numerical issues. Please ensure the size of the floating point value you want to return matches the function's actual return type. Lastly passing the wrong sized floating point values as parms (say a Float) versus an expected Double, will result in confused behaviour, or VM crashes as we are building the parm stack area based on information from the float signature array which must then match what the routine expects the stack to look like. -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

1 0

[OT] If programming languages were religions...
by Klaus D. Witzel 17 Dec '08

17 Dec '08

[... then] "Smalltalk is the ancient Egyptian religion. The Initiated know it already had all the important concepts working long ago and most popular modern languages are incomplete subsets of it that obsess over artificial restrictions of their own creation while entirely missing the reason their code exists in the first place. And the heart of your code will be judged against the Feather of Truth before it can join the message passing afterlife." ----------- This was slashdotted, originated here - http://www.aegisub.net/2008/12/if-programming-languages-were-religions.html Could nor resist mentioning for the squeak-list records ;) /Klaus -- "If at first, the idea is not absurd, then there is no hope for it". Albert Einstein

1 0

I would like to create a new team (actually "old" project because I'm already working on it :) )
by Eliot Miranda 17 Dec '08

17 Dec '08

Hi All, in response to Janko message about teams I would like to create two new new teams called: Faster Execution Engine and Faster Interoperability Engine (suggestions as to what FUM could stand for gratefully received). Yes, you smell the blood of an Englishman. I would like to be team leader. The tasks are: Faster Execution Engine: produce an execution engine that supports closures and provides performance broadly comparable with other Smalltalk JIT virtual machines, of the order of 10 to 20 times faster than the current VM for benchFib. Faster Interoperability Engine: produce a more flexible and faster FFI based on and integrated with the execution technology in FEE. best Eliot

8 10

Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2....
by Jerome Peace 17 Dec '08

17 Dec '08

[squeak-dev] Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2.... >Keith Hodges keith_hodges at yahoo.co.uk >Wed Dec 17 05:05:32 UTC 2008 > Hi Keith, Puck: "Shazaam, he's up early." Me: Thanks for the quick reply. >Jerome Peace wrote: >> [squeak-dev] Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2.... >> Keith Hodges keith_hodges at yahoo.co.uk >> >>> Wed Dec 17 02:17:56 UTC 2008 >>> >>> Dear Greg, >>> >>> some of us have spent a lot of time trying to improve things in work >>> that has been made available SINCE 3.10's release. >>> >>> However you are basically ignoring what people are trying to tell you, >>> and insist on being >>> negative and non-constructive. >>> >> >> There is something that Shaw said about progress and the unreasonable man: >> http://www.elise.com/quotes/quotes/shawquotes.htm >> >> >>> So I wont be replying any further. >>> >> >> Puck points out that Greg is no more unreasonable than you or I. Just a little more ornery. >> >> Don't take this as a call to action for you personally. I figure you are doing quite a lot and maybe are at even your energetic limit of service. >> >> What's being called for here is a recruitment of resources (i.e. others to do the testing). And the call to conscience to listen to the result and live up to high standards. All of these will improve the community. >> >> And Greg is starting to come up with realistic suggestions to get to his unrealistic ideals. >> >> Can we find a box big enough and strong enough to hold an entire universe or two? >> Can you or someone else write a script to load the guru's hot-dog?* >> Can we get someone different, new, fresh, to do the periodic and episodic testing on a regular enough basis? > Can we get someone in authority to care longly and strongly enough about this to watch over it and guide it to see that it happens and keeps happening? >> >> >We are putting the infrastructure in place to do exactly this kind of >testing in an automated fashion. I have a machine spare with a new 160Gb >of HD spare for it. > If you are putting yourself or even Matthew in the critical path for this happening, are you also creating a bottleneck? We also need others with similar facilities to go beyond the limit of time and service you and Matthew can put in. I'm talking about fresh blood here. I know it is easier to choose to do what is within your personal control. All I am saying is that might not be the only way or the best way to go. Leave time for doing what you alone can do best. Offload the tasks that another member of the community can do easily. (Once trained). If we are to get to where we wish to go we will have to get a lot of folks to pull in the right direction. And give them the encouragement they need. >"All" that is needed is to... > >A) have all packages loading tested relative to an Image with a certain >base functionality (i.e. LevelPlayingField.) Okay. That's 3.10.3 onward and the other branches that wish to join the LPF movement. Seems to me we are there with the ftp release you just made. > >B) Fix dependencies in Sake/Packages as a community effort. Huh? Actually this sounds reasonable. What do you mean by it? How is it to be accomplished? When do you teach the community the techniques involved? > >and away we shall go. > >Keith > > >

1 0

Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2....
by Jerome Peace 17 Dec '08

17 Dec '08

[squeak-dev] Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2.... Keith Hodges keith_hodges at yahoo.co.uk >Wed Dec 17 02:17:56 UTC 2008 > >Dear Greg, > >some of us have spent a lot of time trying to improve things in work >that has been made available SINCE 3.10's release. > >However you are basically ignoring what people are trying to tell you, >and insist on being >negative and non-constructive. There is something that Shaw said about progress and the unreasonable man: http://www.elise.com/quotes/quotes/shawquotes.htm >So I wont be replying any further. Puck points out that Greg is no more unreasonable than you or I. Just a little more ornery. Don't take this as a call to action for you personally. I figure you are doing quite a lot and maybe are at even your energetic limit of service. What's being called for here is a recruitment of resources (i.e. others to do the testing). And the call to conscience to listen to the result and live up to high standards. All of these will improve the community. And Greg is starting to come up with realistic suggestions to get to his unrealistic ideals. Can we find a box big enough and strong enough to hold an entire universe or two? Can you or someone else write a script to load the guru's hot-dog?* Can we get someone different, new, fresh, to do the periodic and episodic testing on a regular enough basis? Can we get someone in authority to care longly and strongly enough about this to watch over it and guide it to see that it happens and keeps happening? <...> Yours in curiosity and service, --Jerome Peace *the guru's hot-dog Puck reminds me to tell the story: A guru goes up to the hot-dog vendor and asks Guru:"Can you give me a hot dog?" vendor: "Sure, how would you like it." Guru: "Make me one with everything. :)" vendor: "OKay, that'll be a dollar fifty." Guru: "Here's two dollars, can I have change?" vendor: "No. Change comes from within. :)"

2 1

Agreement among packages better than name-spaces from( Re: worst crash yet ... thread)
by Jerome Peace 17 Dec '08

17 Dec '08

Agreement among packages better than name-spaces from( Re: worst crash yet ... thread) [squeak-dev] Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2.... Andreas Raab andreas.raab at gmx.de >Tue Dec 16 05:00:30 UTC 2008 >Greg A. Woods; Planix, Inc. wrote: >>> Puck would like to know exactly how to recreate this failure. >>> (So he can try it out on some of his "friends" :). >> >> Looks like all I had to do was try to load SmallDEVS from the default >> package universe. > >No surprise. Polymorph as a class called LabelMorph. SmallDEVS has a >class called LabelMorph. They don't agree. So much for not having name >spaces. Um. I don't see this as a name spaces problem so much as an agreement problem. If both packages live in the same universe then they have a need and responsiblity to agree on terms. Namespaces would avoid this but create the dual of this problem. You then would have the clutter of needing two definitions in different namespaces to mean the exact same thing. My take is that disagreements are inevitable. The repair is conversation, negotiation and agreement. That way you have one language not two or many. The other opportunity here is this defines another Simple test for universes. Invariant: No two packages living in the same universe shall define the same class. One will have to differ to the other. Or they shall ever be galaxies apart. The same is probably true for method overides. The indicated repair is to repackage the conflicting code into its own package and have both other packages be dependent on the common package. I say repackage because packages should have clearly defined responsibilities just like objects. The hierarchy of packages should make sense in that way also. These things sort themselves out well if you look at objects (here packages) responsibilities and collaborations. Gary is reinventing things that with more time and understanding would be created with reused classes and objects. SmallDEVS is probably doing the same. Its not like the code doesn't need another iteration of thought. There are magnitudes of differences in speed between 3.10.2 Morphic UI and the Squeak-dev UI. Same with the MVC UI and Morphic UI in 3.10.2. Yours in curiosity and service, --Jerome Peace > ><...>

1 0

Re: worst crash yet trying to load stuff from Packages Universe in 3.10.2....
by Jerome Peace 17 Dec '08

17 Dec '08

[squeak-dev] worst crash yet trying to load stuff from Packages Universe in 3.10.2.... Hi Greg, I like your spirit of adventure. My bug finding/creating assistant Puck thinks you do good work. :) Sure you wouldn't like to become a mantis reporter? *** >Greg A. Woods; Planix, Inc. woods at planix.ca >Mon Dec 15 00:13:37 UTC 2008 > > >OK, one more try I thought..... > >I took a fresh 3.10.2-7179 image, set my preferences, opened up the >Packages Universe, carefully selected the packages I wanted, being >extra careful to avoid the ones that I've had problems with already, >and started the installs...... > >Things looked like they had finished fine, and so I tried clicking on >the Packages Universe window again to see what it might say and >KABOOM! (this all retyped by hand) > > ***System error handling failed*** > Original error: MessageNotUnderstood: LabelMorph class>>contents:. > Debugger error: MessageNotUnderstood: LabelMorph class>>contensts:: > [] in Debugger class>>openOn:context:label:contents:fullView: {[:ex | >self primitiveError: 'Original error: ', title asString, ', Debuge...]} > BlockContext>>valueWithPossibleArgs: > [] in MethodContext(ContextPart)>>handleSignal: {[(self tempAt: 2) >valueWithPossibleArgs: {exception}]} > blockContext>>ensure: > MethodContext(ContextPart)>>handleSignal: > MessageNotUnderstood(Exception)>>signal > LabelMorph class(Object)>>doesNotUnderstand: #contents: > UIThemeSoftSqueak(UITheme)>>buttonLabelForText: > UIThemeSoftSqueak>>buttonLabelForText: > UIThemeSoftSqueak(UITheme)>>buttonLabelFor: Ok. From the evidence so far. You have gotten to the emergency evaluator because the debugger blew up on whatever was the original error. The debugger blew up probably because of a misdirected message. People tend to put in: answer when they meant: ^ answer. In this case something in the Pinesoft ui enhancements looks like it left one out in a LabelMorph class message. Instead of answering an instance it returned the class itself which had no idea what to do with an message to its instance. Congratulations you found a bug. Puck would like to know exactly how to recreate this failure. (So he can try it out on some of his "friends" :). My curiosity would like to know why you are so ambitious as to load so much. And I sympathize with you wanting it all to work right away. Out of the box. What you may be slowly realizing is that all this is Beta software at best. Test coverage in squeak is a recent thought. Most of the software is released with prayers for its survival. First time heavy users find the integration bugs. That would be you. Oldtimers, by instinct born of experience, tend to treat squeak tenderly. As if they are walking on eggs that might break at any instant and give off a rotten stentch. A sad state of affairs to be sure, but the alternative would be a lot of unpaid work. And the person who would take that on would have to have a powerful personal reason for doing so. So far in this small community no one has shown up with that need. Still you never know. Leaving aside the who. Do you or anyone else out there have an idea of how that testing could be done? Yours in curiosity and service, --Jerome Peace

4 14

← Newer
1
2
3
4
5
6
7
8
9
...
12
Older →