Magma September 2006

magma@lists.squeakfoundation.org

9 participants
42 discussions

Re: [Smallwiki] Re: MADynamicObject
by Keith Hodges 14 Sep '06

14 Sep '06

about MADynamicObject and serialization. > It is used to put a calculated value into a variable. Whenever you > send the first message to it, the block gets evaluated and the > This is the bit that I dont understand, I cant see where the 'evaluation' actually happens. Who does it? Where is the code that says 'aBlock value'? > What I find strange is that a MADynamicObject is ending up in the > code to serialize. Before persisting a PRContext, I was zapping its properties to nil. However I had omitted to do the same to the context's command's answer context, as in the case of a PRAddCommand. So it was in the dictionary of properties, containing references to components that was causing the problem. Exactly where WAProxyObject is in there I am not quite sure since I never got a proper debugger up on it. The PRCommandAdd-structureClasses appeared to be the one causing trouble, i.e. a reference to this dynamic object is contained in one of the components. ---- So now Pier-Magma is looking good to go. My main objective was to be able to have the Pier data persisted outside of the image so that an 'upgrade' of the image would be straightforward. There are (at least) two more issues to consider if Magma-Pier is to be a tour-de-force. 1) At the moment pier-magma marks objects for rematerialisation as needed, one-by-one. Magma can be configured for different read-strategies, so performance could be greatly increased (so I am told) if the read strategy re-materialised on a page-by-page or kernel-by-kernel basis (for a multi kernel site). 2) Using a non-shared MagmaSession takes time to get the initial session on the database (it could hold a cached session ready in order to improve this), but when the session is expired the objects may be free to be purged from the image. (Thinking about it I suspect that this will not happen automatically with my current implementation, but it should) My current preferred approach is to use a shared MagmaSession where there is only one MagmaSession for all the seaside sessions. Effectively this caches every used object in memory, and the opportunity for allowing objects to be purged/gc'd is less clear. I am assuming that Magma doesnt do an explicit purge, but relies upon the application to 'let go' of unneeded objects for gc to tidy up. Keith ___________________________________________________________ Try the all-new Yahoo! Mail. "The New Version is radically easier to use" The Wall Street Journal http://uk.docs.yahoo.com/nowyoucan.html

1 0

munging objects before persisting
by Keith Hodges 14 Sep '06

14 Sep '06

In the Pier-Magma integration there are instances od PRContext whose #properties instance var could really use zeroing to nil, before it is ever persisted. I had a look at asStorageObject, but this seems to need an object that has its own graph. I assume that this means that it is representable as a discrete entity, rather than as part of the overall objects graph. My current solution is to ensure that my 'dangerous' object is always sent #magmaCopy before it is stored in the persistent model, and also ensure that it is always sent #magmaCopy before it is likely to be used. This doesnt seem like the most efficient way of doing things. I thought of using preSerialisation, but looking at the code it appears that the result of preserializing has to be the same object, it cant be a copy, otherwise it gets treated as a StorageObject. At present it looks as though I can combine, copying my dangerous object, so as not to disturb the users of the dangerous properties, and using preserialisation to zap the properties before storing. any advice would be appreciated. Keith ___________________________________________________________ To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com

2 1

MagmaServerLoader-cmm.22 - indexing bug fixed
by Chris Muller 12 Sep '06

12 Sep '06

Ok, it looks like MaHashIndexTester is going to finish (been running for 24 hours). Here's what happened. In fixing the freaky index bug Göran identified, I improved the design of MaHashIndex to get rid of the special value (3) and use a new separate control field to properly indicate. But I did it in the airport and there wasn't time to run the (thorough) MaHashIndexTester to ensure no regressions were found. Having finally found time to do that last weekend, it revealed a slip in the new code when removing entries from an index. That's now fixed and MaHashIndexTester is finally approaching the finish line, so I've gone ahead and replaced the bad version on MagmaTester. MagmaTester is now the SqueakSource project where test/beta code goes. Unless I hear from anyone within a week or so of any problems, I'll post it to the "Magma" project. - Chris

1 0

Re: [Smallwiki] Re: Context-#session
by Keith Hodges 11 Sep '06

11 Sep '06

> I see, this is not really useful if you use a persistency mechanism > that dumps out the whole objects. The idea was that Pier knows using > Indeed. > the description what parts of the objects to serialize, something > that is probably not easily controllable with a framework using the > Smalltalk reach-ability to serialize objects. > > So I would suggest that you remove this cache. > > It seems that most of the properties in PRContext are not very useful when persisted. My current approach explicitly copies the parts of the object that I wish to persist, and so now it is not absolutely necessary to remove this cache. I have removed Pier-Magma's use of this cache, and so it remains to see who else actually uses it. > What's the trouble with MAProxyObject instances? > > I am not really sure of the ins and outs of this. I think that Magma was attempting to serialise part of the PRContext structure at the time. The proxy was to an Array of Commands, containing 8 items. Magma was looping through them to serialise them and asked for maInstSize which returned 9 rather than 8, owing to the fact that ProtoObject (i.e MAProxyObject) returns maInstSize as (^self class instSize + self basicSize ) differently to Object. I.e. the call to maInstSize is not being proxied. A fix would be to add an explicit maInstSize to MAProxyObject. I suspect that it would be worth setting up a test case for this one and to work out just what the desired behaviour actually is. Keith ___________________________________________________________ The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html

1 0

Preloaded Seaside-Magma-Pier 0.9 image
by Keith Hodges 11 Sep '06

11 Sep '06

Latest release of image available at ftp.seasidehosting.st Username: preloaded password: squeak Sqk3.8f-smp0.9.image (and changes) This release supports the logging of changes. I do this by copying the useful parts of a PRContext before putting them into the repository and when taking them out of the repository copying them again before using them. For now this seems to work ok. Keith ---------------- Welcome to the Seaside Magma Pier preloaded image. Please send your comments to the magma or pier mailing lists. ---------------------------------------------------------- This release smp 0.9 11thSept2006 First release fully supporting Magma-Pier integration with history and changes logs. ___________________________________________________________ Try the all-new Yahoo! Mail. "The New Version is radically easier to use" The Wall Street Journal http://uk.docs.yahoo.com/nowyoucan.html

1 0

Context-#session
by Keith Hodges 11 Sep '06

11 Sep '06

In trying to implement MagmaPersistency it turns out that the fact that PRContext holds a local cache to the current session may be a bit of a pain. Not only may there be a lot more stuff accidentally persisted through this reference, there may actually be objects that magma gurus say 'don't persist these if you know what is good for you'. I cant see that this cache is widely used, so I wondered if anyone would object if it was removed i.e. make this method lookup the current session the hard but not all that slow way (about 100 microseconds, on my oldish 700Mhz machine) PRContext>>session ^ WACurrentSession value any thoughts/opinions? Keith ___________________________________________________________ The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html

2 2

Questions
by keith hodges 11 Sep '06

11 Sep '06

I am looking for a reliable way to clear up old MagmaSessions. My use of seaside is a bit hit and miss, and I seem to have a load of MagmaSessions left dangling. There seems to be a process that is merrily chugging through every class in the system what is going on. lots of allSubclassesDoGently: Would it be safe to persist a "Semaphore"? Keith ___________________________________________________________ All New Yahoo! Mail Tired of Vi@gr@! come-ons? Let our SpamGuard protect you. http://uk.docs.yahoo.com/nowyoucan.html

3 2

Note: Freaky index bug NOT fixed
by Chris Muller 10 Sep '06

10 Sep '06

Its a good thing for the extensive MaHashIndex test cases. Although they take a long time to run, their thoroughness has found my slip. I'll get it fixed, sorry for the inconvenience. - Chris

1 0

3.9
by Ramiro Diaz Trepat 09 Sep '06

09 Sep '06

I am trying Magma (MagmaServerLoader 20) on the latest 3.9 image (56) and I get the following error while trying to create a repository Outage occurred while writing system-definitions! Will now attempt to repair. Select Proceed to continue, or close this window to cancel the operation. Does Magma work on 3.9? Any hint on what this could be? thanks. r.

2 1

MakoServer missing
by Keith Hodges 09 Sep '06

09 Sep '06

Chris, I thought I would try the latest packages, but I got errors about a class MakoServer which I could not find, but it appears in 'Ma client server'. I went back to version 116. to gte things working but I thought I would let you know Keith ___________________________________________________________ To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Magma September 2006