Setools August 2007

setools@lists.squeakfoundation.org

6 participants
12 discussions

Confirm an issue please.
by Igor Stasenko 31 Aug '07

31 Aug '07

Hi. I have an issue with viewing a case in remote magma setup. While in local magma setup all works just fine (in single image), but when i starting two images i got DNU's like: MessageNotUnderstood: UndefinedObject>>key it appears in Q2AssociationField>>renderViewControlOn: html component: fc because cache is nil. I investigated a problem and found, that field unable to read a value from case instance and returns nil values as default (so fieldcomponent case = nil). By placing a break in Q2AssociationField>>valueFromTarget: aTarget ifAbsent: aBlock and tracing it, i found that accessor is unable to send message to case instance because it returns false from #respondsTo: message. What is more strange is, that this happens following code to break: Q2SelectorAccessor>>valueFromTarget: aTarget ifAbsent: aBlock ^ (aTarget respondsTo: selector) ifTrue: [ aTarget perform: selector ] ifFalse: [ self break. aBlock value ] while aTarget is Q2Case instance and selector = #responsibleUserAndRole. And by entering in debugger and evaluating aTarget respondsTo: selector is returns true!!! I think something is wrong with maMutatingProxy when accessor object first time used to get value, its still is mutating proxy object and uses: #mutateAndSend: message. And something not working in this context. After mutating everything works ok, but what makes Q2Case instance respond false to first #respondsTo: message? I will continue debugging trying to figure out.. Maybe a 'selector' ivar in Q2SelectorAccessor is not reified correctly.. Btw, Levente i downloaded image from test2 (MagmaCounter.1.image), which you saying 'works fine'.. And guess what? It not working. So, please, confirm an issue or explain, what im doing wrong?? :) -- Best regards, Igor Stasenko AKA sig.

1 2

Q2v0.4-ul.338
by Levente Uzonyi 29 Aug '07

29 Aug '07

Hi! > doing cleanImage On client: > 2007-08-28T21:38:51.199+02:00 - warn(4): Released MagmaSession 2 from > service: nil but if was not in the pool before! (Q2MagmaSessionPool) This can happen if the pool is closed (all sessions are closed and removed from the pool), but the objects which are holding references to the sessions try to release them with the pool. Since the pool is empty, every released session will cause a warning. (Not that nice, i know :)) > doing cleanImage on server caused DNU too, in finalization of > MagmaCollectionManager with session = nil. This seems to be unclean shutdown of magma. > Also, i tried create new case using remote session and got DNU while > rendering Q2AttachmentField due to Q2Case>>attachments returnung nil. > Strange because case , when creating assigns empty OrderedCollection to > it. > On other hand, creating new case using single image works fine. Test2 currently has a double image setup (magma server + gjallar), and i can create cases on it. > Also, i remember that i removed 'selector' ivar from Q2Field, it > reappeared in Levente snapshot.. Strange. Just checked sig.337, which is the ancestor of ul.338 and it's Q2Field has the instance variable too. Cheers, Levente

1 0

Yet another diagram tool
by Igor Stasenko 29 Aug '07

29 Aug '07

http://smallwiki.unibe.ch/moose/tools/eyesee/ Comparing to gnuplot, it uses objects for creating diagrams, not literal numbers -- Best regards, Igor Stasenko AKA sig.

3 3

Re: Counter in Gjallar
by Keith Hodges 28 Aug '07

28 Aug '07

Göran Krampe wrote: > Hi Chris! > > >> Hi GÃ¶ran, no, I'm sorry to hear that since I spent a good 4 days on it >> and ended up quite pleased with the result. >> > > No need to be sorry! :) I haven't looked but IIRC from your post you > implemented an RC counter a la GemStone, right? Which is quite useful for > counting stuff of course. :) > > >> What's the problem with it? >> > > Well, we have multiple MagmaSessions creating Q2Txn objects, giving them a > strictly increasing unique number and adding them to a MagmaCollection. In > short we want the illusion of a very long OrderedCollection. :) > > In an OrderedCollection each object of course has its position and if you > succeed adding an object to it (without conflicts) then its number == its > position. > > With a MagmaCollection we need to give it a number ourselves (and use an > index for it) - and the problem is how we generate that number so that we > don't end up with multiple Q2Txns with the same number - and not having to > do multiple commits. > > In other words - an "integer series generator" is not exactly the same > problem as a "reduced conflict counter". Note btw that I am describing > this from MY understanding - Balazs and Levente are the ones who are > coding on this and perhaps I have misunderstood some parts :) > > The "simplistic" solution is of course to let all MagmaSessions make a > server call (to the Magma server) which has a counter wrapped up in > critical: blocks for increasing it. Thus making sure we serialize all > "increments". This is what Keith implemented which apparently did not work > I simply implemented a request-response which uses the existing protocol. It has a class which handles both ends of the communication. Of course this class (these classes) has to be present in the server for it to work. This effectively bypasses the database altogether, so there are no critical blocks involved. Since the server handles requests serially it does ensure that the counter does not conflict. The bit that I did not know how to do was how to persist the counter values. I just used a file on the server to serialize read/ unserialize write a dictionary with the counter values in it on every increment. I myself have not looked at MagmaCounter yet. > in the Magma version I tried a week ago (the server seems to not deal with > the MagmaServerRequest properly when running as a separate image - causing > the client image to "hang" waiting for a response. I logged all Magma > The server image has to have the responder classes loaded in order to respond. > requests going and all seems fine until this "MagmaServerRequest" coming > through the line, haven't debugged it further on the server side - failed > to see where to put the darn halt. :)). > best regards Keith

2 1

Q2v0.4-ul.338
by Igor Stasenko 28 Aug '07

28 Aug '07

Nice, seems we first time was able to do something with remote server connection. Congratulations! :) But there still some errors: doing cleanImage On client: 2007-08-28T21:38:51.199+02:00 - warn(4): Released MagmaSession 2 from service: nil but if was not in the pool before! (Q2MagmaSessionPool) doing cleanImage on server caused DNU too, in finalization of MagmaCollectionManager with session = nil. Also, i tried create new case using remote session and got DNU while rendering Q2AttachmentField due to Q2Case>>attachments returnung nil. Strange because case , when creating assigns empty OrderedCollection to it. On other hand, creating new case using single image works fine. Also, i remember that i removed 'selector' ivar from Q2Field, it reappeared in Levente snapshot.. Strange. -- Best regards, Igor Stasenko AKA sig.

1 0

Yet again, troubles with indexing under win32
by Igor Stasenko 21 Aug '07

21 Aug '07

Q2Win32SwishEWrapper>>index:into: have a bug ProcessWrapper global is undefined I don't know, maybe i lost something during merging. I just testing field accessors and going through most frequent uses cases like adding case , editing forms e.t.c -- Best regards, Igor Stasenko AKA sig.

1 0

Does gjallar have this fix?
by Igor Stasenko 15 Aug '07

15 Aug '07

http://bugs.squeak.org/view.php?id=6588 maybe client 'locking', which i experienced is because of it? -- Best regards, Igor Stasenko AKA sig.

1 0

[Fwd: Re: Hi Chris]
by Keith Hodges 13 Aug '07

13 Aug '07

3 3

Settings and forms
by Igor Stasenko 08 Aug '07

08 Aug '07

At current stage we have variable code which uses different approaches and different messages to access settings and show UI for editing them. I really think that we need to develop a common way how to access settings and how provide users a way to manage them. Currently there are 3 use cases, where we need to access settings: - .ini file settings - global settings (aka model) - process settings For accessing .ini file settings i think we should use 'self session settings' but Q2Object>>settings must not point to them by default, actually this must be an abstract method, which defines a protocol. For accessing global settings we must use 'self model settings' and for process - 'self process settings'. I don't see a need in defining other messages, like #processSetting: in Q2Object class, or #modelSettingAt: in Q2Model class. Maybe, for cleaning out the code we may add #processSettings and #modelSettings to Q2Object, which will return corresponding: 'self model settings' and 'self process settings' ? But such changes will require from us to return some object, known as 'settings' which will have some protocol (basically #at:, #at:put:). Should we enforce such behavior? or maybe better define a protocol in Q2Object, like: #settingAt:, #settingAt:put: , so we will not need an intermediate object(s) to access different settings? But returning intermediate object haves own benefits - when we need to show UI for editing settings, we can simply use something like: self model settings getEditorComponent , which should return a ready for use component for editing it. Now, about providing UI for editing settings. A Q2Object have a methods in category 'formvalue' - #editForm - #editFormBuffer I think this is very good scheme which gives us ability to edit contents of any Q2Object by sending #editForm message to it. But its have own limitations - in some cases we need to build a multi-tab forms containing multi-form value for showing settings UI. Maybe we should, instead of asking form from object, ask for component(s) which represent UI for editing its contents? A Q2Object having an abstract method #addComponentsIn: , so what if we do: self model settings addComponentsIn: ... for getting editor UI? Then particular Q2Object can determine what component it needs to add to given component to represent itself. It can check access and provide viewer or editor component depends on access rights. But in some cases, we need to tell that we need only view of object's contents, so what if we define method like following: Q2Object>>addComponentsIn: blabla ^ self addComponentsIn: blabla forEditing: true. And now, #addComponentsIn:forEditing: can be implemented in subclasses. Another problem, on which i want to pay your attention, is about changing forms dynamically. In current design, a form is build as something static and then based on it we build a components which then live long and don't change even if form changed. I know, that this was done for speed reasons , to not create components each time we need to render forms. But we need some mechanisms to synch forms and their corresponding components, when form was changed. Actually even forms can be changed dynamically (for example - i need a value for each stage name in color settings, and for this i need to rebuild form each time a stage with new name appeared / removed). Maybe we should add a 'dynamic field' class, a subclass of Q2Field, which can be used for editing/showing contents of current state of model? Lets say, model is array of strings - and we need to show a list of fields for editing each string, and then if something in model got changed (new string added or removed) we need to show different list of edit controls for it. The best way i think is to let fields render themselves without creating components. So, the only component will be a Q2FormComponent , but fields will be rendered without creating components. This will reduce the amount of complexity and will give us a way to make dynamic forms which always will represent editor/viewer for model in current state. Actually , there already some misconcept: fields used #addComponentsIn: to define components which represent for editing values of target, but not field itself! I added a #fillForm: method to make editor UI for field in forms constructor, but i suspect that this should be #addComponentsIn: , since Q2Object states that this method must add editor/viewer or receiver's contents. Currently, im refactoring fields by adding accessors to them (instead of selector ivar). I think we should change the protocols how fields should be rendered, but i don't have a clear concept of it. I added a field grouping and divider , which is a Q2Field subclasses , but they don't have any values to edit , they simply a UI decorations. But to add this i was need to hack and write ugly code and i not very satisfied with it :) -- Best regards, Igor Stasenko AKA sig.

1 0

For Gjallar devs: New build available
by goran＠krampe.se 08 Aug '07

08 Aug '07

Hi! See: http://mc.gjallar.se/download/images/ Not tested, it was in the middle of the night when it finally was ready - so I zipped it up and posted it. But my guess is that it is "fairly ok". :) regards, Göran

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

Setools August 2007