Squeak-dev March 2003

squeak-dev@lists.squeakfoundation.org

202 participants
818 discussions

Re: SPrevayler
by Hannes Hirzel 02 Mar '03

02 Mar '03

Marco Paga <mail(a)marco-paga.de> wrote: > In my mind that is something that is against the idea of prevalence. The > programmer would have to think on more details. SPrevayler would become > feature laden and that is something I don't want prevayler to be. In my > mind it should be easy to use. You should not see a difference between > code that uses prevayler and code that doesn't use prevayler. Seconded ! -- Hannes

1 0

Re: Possible goals for future releases (was: Re: Let's Release 3. 4)
by Hannes Hirzel 02 Mar '03

02 Mar '03

Marcus Marcus Denker <marcus(a)ira.uka.de> wrote: > On Sat, Mar 01, 2003 at 11:13:42AM -0800, Tim Rowledge wrote: > > Hannes Hirzel <hannes.hirzel.squeaklist(a)bluewin.ch> wrote: > > > > > release cycle - 10 days alpha - 10 days beta 10 days gamma) and then > > > switch to your proposition of quaterly releases? > > Not Enough Time To Do Anything. > > > > I have this impression, too. > > Hannes, remember: We are here to have fun... deadlines are something > you'd need to pay me for. > > Marcus > Well the proposal I came up with is to have a one month cycle only for 3.5 and then switch to a quaterly release scheme. Daniel Vainsencher gives a reason for this http://lists.squeakfoundation.org/pipermail/squeakfoundation/2003-March/ 000986.html >Suggestion - make 3.5 a very short release, consisting mainly of getting >as many fixes in as we can, and giving them a proper testing. It's an >opportunity for us Guides **to warm up at doing the whole process** >ourselves. I'm talking about a time scale of 1 month total - 2 weeks >alpha, 1 week beta, 1 week gamma, release. We start with specific, very >doable goals (maybe even decide in advance a prioritized list of fixes >we want to get in, not trying to catch up with everything that's up for >harvesting), and basically spill them all into the update stream on day >one, with the rest of the time for integration. Fixes that aren't ready >for integration (meaning, posted tested, and reviewed by someone) by the >end of week 1 will wait for 3.6. (emphasis mine) Cees de Groot favours a well having a quaterly release schedule. This means that the amount of new things you have to put in a release may vary. If something doesn't make it for the current release then it will be put in the next one. No big deal. I do not see that this diminishes the fun. But other people may have other preferences. This brings us to an important question is of course - what does it "cost" (in terms of time , "non fun work", frustration) to do a release? I am still under the impression that finalizing a release is doing some checking procedures, running some scripts and uploading the image to the library. But of course there might be more things involved here than I am aware of. I think it would be good to have a rough idea people have to put in such a thing. As you know in XP cycles they have releases for example every week. >From this perspective having a one month release once doesn't seem such a big deal. The whole issue is open to discussion - for exactly this reason I started this thread. I would like us to list the pros and cons of various release schemes - probably on a squeakfoundation swiki page which is accessible from Life cycle of a release http://swiki.squeakfoundation.org/squeakfoundation/89 or Release Plan http://swiki.squeakfoundation.org/squeakfoundation/79 The point is: We have to decide between goal oriented release schemes vs. schedule based release schemes. At this moment I favour having more releases than we had now. I think it would motivate people considerable because they see that people are active and their contributions will show up in due time. What do you think for example about MCP. Is it a motivation for them not knowing if their work will perhaps be included in June, maybe October maybe next March. I really think we need some synchronisation. But for sure - keep up the good work! Hannes P.S. An somewhat modified release scheme could be 3.5 final beginning of May 3.6 final beginning of August 3.7 final beginning of November 3.8 final beginning of February 2004 to avoid somehow the "deadlyness" of the "deadlines". "I do not get paid for this!"

1 0

RE: Possible goals for future releases (was: Re: Let's Release 3. 4)
by Hannes Hirzel 02 Mar '03

02 Mar '03

Cees de Groot <cg(a)cdegroot.com> wrote: > Furthermore, I'm not saying that we should stick with this regime > forever - see it as a 'therapy' against: > - slipping release dates and thereby uncertainty with outsiders what is > happening with Squeak; > - a rusty release process - with a strict quarterly schedule, you get > into a rythm and the whole process can be smoothed out quickly; > - to a lesser extent, protection against 'too big' releases that have > the associated risk of breakage.=20 > As I said, my experience is that it helps tremendously; my experience is > usually with weekly or fortnightly release cycles, but that's clearly to > short for Squeak. What about having end of March for the release date for 3.5 (only on month release cycle - 10 days alpha - 10 days beta 10 days gamma) and then switch to your proposition of quaterly releases? End of June, end of September, end of December? -- Hannes

4 3

Re: [FIX] #union: broken for sets
by Richard A. O'Keefe 02 Mar '03

02 Mar '03

Daniel Vainsencher <danielv(a)netvision.net.il> wrote: I wonder if #union: of Bags should use Set semantics or Bag semantics. This has been discussed before. In fact, I recall raising the issue myself quite some time ago. Currently, and with your suggestion - #(1 2 1) asBag union: #(1 2) asBag -> a Set(1 2) which is precisely what it is documented as doing. I would personally _prefer_ it to return a Bag, but I can't call it a bug when a method delivers the result its documentation says it delivers, without any nasty side effects. It would be more meaningful, and one can always convert to Sets first to get the usual functionality. Well, you can always write (aCollection copy asBag) addAll: anotherCollection; yourself if you want bag union. Or even (aCollection asArray , anotherCollection asArray) asBag More importantly, if #union: were changed in this way, #intersection: and #difference: should also be changed. To be really honest, I think #union:, #intersection:, and #difference: belong in Set (and its subclasses), not Collection. How often do you want to compute the union of a Dictionary and a SortedCollection? But what Bag>>union: does and whether #union: should be in Collection at all are design issues that deserve wider debate and probably can't be changed without breaking someone's code. We don't have to settle such matters before fixing an obvious bug.

2 1

Re: Storage
by None 01 Mar '03

01 Mar '03

Chad Nantais <chad(a)whistler.com> wrote: > Hi Hannes, > > Yes. I am considering using seaside for my web application framework. > > Chad Hi Chad this is very interesting. An innovative approach actually. Of course you are on the mailing list http://lists.squeakfoundation.org/pipermail/seaside/ then. But it would be great to hear from you how things are working from time to time. I would love to add a story about what you are doing to the documentation swiki. Squeak related problems are probably answered here quicker because of the larger audience. Feel free to post anytime. Probably you are going to do a quick prototype next to show feasability. Cheers Hannes

2 1

Unable to install SqueakMap in Squeak3.4a-5125 image
by Joseph Frippiat 01 Mar '03

01 Mar '03

Hi everyone, I just tried to install SqueakMap in a fresh Squeak 3.4alpha 5125 image but I didn't succeed. I got a debugger window with the following message: MessageNotUnderstood: created:updated:name:currentVersion:summary:description:url:downloadUrl: author:maintainer:registrator:password:categories: and arguments: #(3222872289 3222872289 'Flow' '2 alpha' 'a framework for streaming, including networking, filesystem, and device driver support' 'Please see http://netjam.org/flow .' 'http://netjam.org/flow' 'http://netjam.org/smalltalk/sources/flow/flow2a.zip' 'Craig Latta <craig(a)netjam.org>' 'Craig Latta <craig(a)netjam.org>' 'Craig Latta <craig(a)netjam.org>' 1259017709235410723060456039043336829716473140864 #('25243867-b8c9-4ce8-bbdd-0c83bac07b02' 'd233e4ae-4420-41d4-ad97-4b8c52139812' '0990c683-1a6e-4452-98b4-f17ded06a3c9' 'fab30799-b80d-4598-b740-b6d1e30254b7' '82f328ab-175d-40b4-a363-bc0287dabd77' 'b4fc1809-3d56-4145-94b7-036531b9dc21' '7295c005-2305-4565-b718-fb97d0f82843')) Could someone please help me ? Thanks Joseph Frippiat

2 1

Re: SUnit Test Failure on fresh Image - Win32
by Hannes Hirzel 01 Mar '03

01 Mar '03

Andrew Andrew Berg <andrew_c_berg(a)pop.mail.yahoo.com> wrote: > > That's a bit funny. I consider myself a novice, and I also get a similar > failed test: > TestUUIDPrimatives>>testCreationRandom > > I fiddled about with the debugger for a while but didn't really get > anywhere toward figuring out what was wrong. it looked like UUID's were > being made, but then the test was looking for some specific bits to be in > some specific pattern, and they were not. I'm stuck there. > > -andrew Actually SUnit tests are there to show if bugs are present or not. If a test fails that means that a bug is present. It may be in the test or the target software. Software often has bugs. The bug you see is not the only one in the Squeak image but it is visible for anybody because of the SUnit tests. For this reason Tim Rowledge wants to remove the SUnit tests to lower the visibility in the final release (you will then have to load SUnit and the tests from SqueakMap to do the tests). As an illustration I add an email by Doug Way below. -- Hannes P.S. You know, even an unnamed large software company ships software with bugs. But they don't provide the tools to easily detect them .... ;-) That these kinds of questions are discussed shows that we have an emerging testing culture within the Squeak community. In the long run having these tests will speed up the develoment process enourmously. At the moment most tests are performed by harvesters inspecting at the code, putting it in the update stream and waiting if somebody complains. This waiting process is responsible for long release cycles. ------------------------------------------------------------------------ ----------------------------------------- > Date: Sat, 21 Dec 2002 00:06:03 -0500 > From: Doug Way <dway(a)riskmetrics.com> > Subject: Failing SUnit tests in 3.4b > > One thing that would be nice for any "final" Squeak release, such as > 3.4, is to have the SUnit tests all passing. > > There are currently a few tests which are failing. Oddly enough, they > are different tests on different platforms, which means they're either > VM-related or otherwise platform-specific (such as the platform-specific > stuff in *FileDirectory, etc.) > > Here is what I am seeing with the current 3.4beta-5156: > > On Mac OS X (Carbon VM 3.2.8Beta9, and CocoaSqueak-3.2.4): > ------------ > 3 failed: > FileDirectoryTests>>testDeleteDirectory > FileDirectoryTests>>testDirectoryExists > FileDirectoryTests>>testExists > 1 with errors: > FileList2ModalDialogsTest>>testModalFolderSelector (these > failures/errors were not in 3.2) > > On Windows 2000 (3.2.3 VM / Tea 1.8 VM): > ---------------- > 1 failed: > TestUUIDPrimitives>>testCreationRandom (this failure was also in 3.2) > > > (Are any tests failing under Linux?) > > Anyway, I seem to remember something on the list around the time of the > FileDirectory fixes, about a change still needing to be made in the VM, > or in the tests, or something. :-) > > There are a few other fixes afoot right now for 3.4beta (including one > related to UUID). It would also be nice to get these tests cleaned up > while we're at it. Any insight on these would be appeciated. > > - Doug Way

2 1

Re: Storage
by Hannes Hirzel 01 Mar '03

01 Mar '03

Chad Nantais <chad(a)whistler.com> wrote: > Hi Alan, > > I had developed a system for car dealership service departments that > allowed for profiling of customers based on their repair activity, > satisfaction level, etc. The system was developed using Zope + MySQL. > I want to create a sister application that is used by sales departments > for lead management, customer profiling, etc. This new system _should_ > allow for migration of data from the service tables. It will have to > allow for the ability to manage new types of customers in different > states. It also has to deal with the fact that a vehicle is both a > deliverable (you can sell it) and a repairable (you can fix it). > > I have the luxury of time and very few implementation constraints so I > would like to experiment with using Squeak to build this system which > will have to manage thousands of vehicles, customers, leads, service > invoices, and follow-up calls and the many relationships between them. > The keyword here is "profiling". That's where catalogs and queries > come in. I tried to do this with Zope, but I found unit testing > awkward, if not impossible. I also tried to develop a solution using > the ZODB without Zope, but felt like I was just reinventing the wheel. > So, I'm going to see if Squeak is a better fit for this problem. > > Chad Just wondering .... Are you considering going to use 'seaside'? http://map2.squeakfoundation.org/sm/package/b135c62f-80b5-4a2f-80d4-7be8 389ed54e and http://beta4.com/seaside2/ -- Hannes

2 1

RE: Possible goals for future releases (was: Re: Let's Release 3. 4)
by Hannes Hirzel 01 Mar '03

01 Mar '03

Tim Rowledge <tim(a)sumeru.stanford.edu> wrote: > One might possibly claim that > we have enough stuff ready to start trying to consider alpha for a very > minor release but that still means assembling a test release, getting a > good number of people to try it etc. > > tim > Yes I vote for a minor release 3.5 by the end of March. And be assured there are enough people to help with it .... Hannes

1 0

Re: Storage
by Hannes Hirzel 01 Mar '03

01 Mar '03

Hi Chad Chad Nantais <chad(a)whistler.com> wrote: > I am building a system from the ground-up and would like to stick to > Squeak as much as possible to implement the system. The system will > eventually generate hundreds of thousands of instances and I will need > quick, query-based access to these instances. How far can I go with > using a Squeak image for storage before I need to use some form of > external storage like Relational or Object DBs? > > Chad Congratulations to you for choosing this approach! I would recommend you to do an experiment. Create your net of objects, fill it with random data (strings and random numbers), save the image, check the amount of memory it needs, restart the image create additional data, measure againg until you get enough information to have an estimate. Nowadays it is easily to work with let's say 512MB of RAM. Diego Gomez Deck says that he has been working with hundreds of thousands of instances. At the moment I am using this approach but only for about 30000 objects. I have an image size of 52MB, but because I have about 70 projects with a lot of graphics in it as well. I have 192MB of RAM. It is really exciting to work with a net of living objects you can query. You even can change the features of your living objects as you work (adding instance vars and methods). This is the *huge* difference to Java as Alejandero Raimondo likes to put it. Cheers Hannes

1 0

Jump to page: