October 2009 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] FilePliugin???

by John M McIntosh

Well the FilePlugin is built into the VM versus being an external VM and it supplies the basic file I/O operations. So you won't find a plugin called that unless some-one decided to compile up one. So either the plugin method call fails because it cannot be found, or it fails because it explicitly failed on bad data, or other reason (an error?) The failure cases are: (1) the id object is not a byte array type of object. Given MultiByteFileStream(StandardFileStream)>>primGetPosition: Receiver: MultiByteFileStream: '/Data/Downloads/pharo1.0-10470- BETAdev09.10.2/installScript.st' Arguments and temporary variables: id: #[153 12 133 142 168 105 19 160 0 0 0 0 18 18 0 0 0 0 0 0 0 0 0 0] Receiver's instance variables: Yes that looks like a byte array object. (b) the size of id is wrong This is unlikely since the id bytearray hasn't changed size in decades. Let me assume the number of bytes is correct. (c) the bytes 5-8 have to be non-zero, otherwise we fail Nope, it's 168 105 19 160 (d) bytes 1-3 have to be non-zero AND match some magic value in the VM Well 153 12 133 142 is non zero, but what's the magic value? MMm I don't think you can get the magic value directly, but at the point of failure you could create a new file and examine the 'id' value for the StandardFileStream and see if the first 4 bytes match the first 4 bytes of this value (153 12 133 142) Failure to match would cause a prim failure. Oh wait I see you can invoke self primSizeNoError: fileID and if that returns NIL the handle is stale. See ensureOpen (e) the ftello() fails because the file doesn't exist, or an I/O error happened, or the path is bad. Or some icky file system error. Alas we don't feed back or record what the error is, but I think you fail on (d)? I'd bet on (d) the MultiByteFileStream instance is some cached value from a previous VM usage, try the primSizeNoError: and see what it returns when you're inspecting the MultiByteFileStream. On 2009-10-12, at 9:42 PM, StÃ©phane Ducasse wrote: > Hi > > when trying to build the dev image I systematically get a DNU > > primGetPosition: id > "Get this files current position." > > <primitive: 'primitiveFileGetPosition' module: 'FilePlugin'> > self primitiveFailed > > and I do not have the FilePlugin in the resources VM folder. > > I checked the web site of john but could not find an explanation. > > Stef > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

Oct. 13, 2009

Re: [Pharo-project] Preparing the sprint

by StÃ©phane Ducasse

Hernan > > How long the community will support the deliberate decision of not > answering to a major contributor? > How do you expect contribution of large open source companies to this > project avoiding political or technical conflicts (or both)? this is not that we do not reply. This is that keith is always saying the same. Now I do not have to power to say to people not to develop their own stuff and also people have the freedom not to understand - sake - now you can ask jannik for example: we started to write something on installer and sake but we got confused on sake (may be with a physical meeting at esug this would make a lot of point clearer) - not to like some part of the installer - not to contribute - not have the time to understand something - prefer to build thier own even if this is limited - For example I tried to check the SUnitextensions and I could not find them. Did you see me not integrating something in Pharo Did you see me not mentioning rio? Even if they are some points I do not like. > > > Focusing only on bug-driven development and ingoring valid questions > with a kind of systematic corporative silence is not the way to go. > And if the managerial staff decide to continue ignoring, sooner or > later you would have two problems to resolve: The unsolved pending > questions plus the management of the emergence of social subsystems > (if the number of contributors increases) without enough coordinators > for pivots, because of the limited size of contributors for a niche > technology like Smalltalk. > > Censorship of uninformed newcomers is common and an implicit norm, > censorship of a known member is disgusting. sorry hernan but I cannot let you say that in this mailing-list. Did we ever not reply to a newbie email? > Such attitude is limiting involvement and participation in the > community, because implicitly > define clear roles which affects the discussion space and group > dynamics. > > Finally, I do not want a holy war here, I just want responses without > personal attacks and without quoting out of context fallacies. Are my answer satisfactory? Now I have a question: - do you use sake personnally? - do you have positive experience? - did you check the code? what is your feedback? - is there tests? - what do you think about task definition pros and cons? You can replace Sake but RIO, Sunit extensions (if you find it) Please do it. So far in our group people succeeded to make TestServer running for Moose and if they come and tell me that sake is better/easier/ then we will probably switch Now I asked the summertalk student to also consider evaluating Sake and we will see. Stef

Oct. 13, 2009

[Pharo-project] FilePliugin???

by StÃ©phane Ducasse

Hi when trying to build the dev image I systematically get a DNU primGetPosition: id "Get this files current position." <primitive: 'primitiveFileGetPosition' module: 'FilePlugin'> self primitiveFailed and I do not have the FilePlugin in the resources VM folder. I checked the web site of john but could not find an explanation. Stef

Oct. 13, 2009

Re: [Pharo-project] Preparing the sprint

by keith

On 12 Oct 2009, at 22:04, Dale Henrichs wrote: > RT @tinybuddha: "Do not give your attention to what others do or > fail to do; give it to what you do or fail to do." ~Dhammapada Not much of a team player this dude then. How about this one... And so, my fellow Smalltalkers: ask not what your community can do for you - ask what you can do for your community. My fellow coders of the world: ask not what Pharo will do for you, but what together we can do for the freedom of man etc etc etc Keith

Oct. 13, 2009

Re: [Pharo-project] Preparing the sprint

by HernÃ¡n Morales Durand

How long the community will support the deliberate decision of not answering to a major contributor? How do you expect contribution of large open source companies to this project avoiding political or technical conflicts (or both)? Focusing only on bug-driven development and ingoring valid questions with a kind of systematic corporative silence is not the way to go. And if the managerial staff decide to continue ignoring, sooner or later you would have two problems to resolve: The unsolved pending questions plus the management of the emergence of social subsystems (if the number of contributors increases) without enough coordinators for pivots, because of the limited size of contributors for a niche technology like Smalltalk. Censorship of uninformed newcomers is common and an implicit norm, censorship of a known member is disgusting. Such attitude is limiting involvement and participation in the community, because implicitly define clear roles which affects the discussion space and group dynamics. Finally, I do not want a holy war here, I just want responses without personal attacks and without quoting out of context fallacies. Cheers, HernÃ¡n 2009/10/12 keith <keith_hodges(a)yahoo.co.uk>: > > On 12 Oct 2009, at 17:45, StÃ©phane Ducasse wrote: > >> Mike is it related to the testServer publihsed on squeaksource? >> Because ESUG will also sponsor a summer talk project on the topics. >> So if we could join forces it would be great. >> >> For the server this is not clear yet. Internally we have our server >> that we could use but >> we have to check the security. >> >> I'm waiting that marcus arrives to check that with him and damien. >> >> Stef > > Why why why why why is compete more attractive than contributing. > > Let me explain.... > > Step 1. We would like a test server. > Step 2. Is anyone working on this already? > Step 3. Would anyone like to collaborate to specify and work on a test > server? > > Bob is already written it just needs configuring, TestReporter is > already written, > and has been available for 3 years or more. > > The same thing is happening again and again and again. > > Why did anyone even start working on Metacello when Sake/Packages was > already there a year before hand. > > Why are the steps above not the default way of working for the > community? It is beyond incredulity. > > Keith > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 13, 2009

[Pharo-project] Polymorph observations

by Schwab,Wilhelm K

Gary, I have done a little with Polymorph and am quite impressed. A small usability suggestion for the examples is to include comments with the text on a line by itself. It is a lot easier to execute something formatted as " ThisOrThat example ThisOrThat someOtherExample " than it is the (very common in Squeak) "GoAhead iChallengeYouToSelectMe andMissTheQuotes". On to more substantive things. TEasilyThemed is a big help, but some model and/or presenter classes would help. Consider a list. One has to define an aspect for the selection index and initialize it to zero. Similar comments apply to trees, with the added wrinkle of wrapping the nodes with appropriate objects. Dolphin's MVP framework creates default models that can be replaced. One weakness of the defaults is that they can take significant time to create for substantial objects. I hacked around that, and we could easily design it out of anything that we build. Dolphin lists and trees treat their selection in terms of the object selected rather than the index. One can get and set the index if needed, but it is not generally as interesting as the object behind the selection. One good thing about specifying the aspect is that one can can connect a view to an aspect of any model. Dolphin allows for the same flexibility via value adapters and type converters. Does TEasilyThemed include helpers for creating splitters? Bill

Oct. 13, 2009

Re: [Pharo-project] Preparing the sprint

by Michael Roberts

http://code.google.com/p/pharo/issues/detail?id=1299 Mike

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by keith

On 12 Oct 2009, at 22:28, keith wrote: > > On 12 Oct 2009, at 22:08, Lukas Renggli wrote: > >>> If you can post the code and maybe some comments on how to use it, >>> that would be great. Maybe we can do some testing and investigation >>> into the problem during the sprint. >>> >>> Do we have an issue for the OSProcess problem or has it been >>> discussed >>> before? >>> >>> We'll discuss where to host this server. We have several options >>> there. >> >> We are using Mason from <http://source.wiresong.ca/mc> to build >> Seaside images. We don't have a server setup yet, but it also works >> locally. >> >> Mason is a very flexible system, it has infrastructure to build >> images, build SAR archives, build documentation, load Monticello 1 >> and >> 2 packages, load Metacello configurations, run tests, run code >> critics >> and execute arbitrary scripts, etc. The OB and MC2 distributions are >> built using Mason for example: >> <http://www.wiresong.ca/static/releases/>. >> >> Lukas >> > > Absolutely incredible. > > Keith translated = "oh look its Sake" Keith

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Adrian Lienhard

I think it makes sense to open a ticket in our tracker, yes. Especially since OSProcess is an important package for us. Adrian On Oct 12, 2009, at 23:19 , Michael Roberts wrote: > Adrian, sorry i did not read your mail carefully ;-) > > no we don't have an issue on the tracker yet. whilst it is an external > package I think it is important enough we maintain compatibility. Is > it reasonable for us to track on our tracker? I would say so, but in > the past sometimes i think we have closed issues for not being pharo > directly. > > cheers Mike > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by keith

On 12 Oct 2009, at 22:08, Lukas Renggli wrote: >> If you can post the code and maybe some comments on how to use it, >> that would be great. Maybe we can do some testing and investigation >> into the problem during the sprint. >> >> Do we have an issue for the OSProcess problem or has it been >> discussed >> before? >> >> We'll discuss where to host this server. We have several options >> there. > > We are using Mason from <http://source.wiresong.ca/mc> to build > Seaside images. We don't have a server setup yet, but it also works > locally. > > Mason is a very flexible system, it has infrastructure to build > images, build SAR archives, build documentation, load Monticello 1 and > 2 packages, load Metacello configurations, run tests, run code critics > and execute arbitrary scripts, etc. The OB and MC2 distributions are > built using Mason for example: > <http://www.wiresong.ca/static/releases/>. > > Lukas > Absolutely incredible. Keith

Oct. 12, 2009