I'm inquiring and it seems related to bad handling of author/timestamps. Here: (CompiledMethod allInstances select: [:e | e fileIndex = 1]) detectMax: [:e | e filePosition] (ZipWriteStream class>>#regressionTestFrom: "a CompiledMethod(353370112)") filePosition 16234210 (SourceFiles at: 1) size 16249263 So, the highest position still in the source file. Let's see the lowest: (CompiledMethod allInstances select: [:e | e fileIndex = 1]) detectMin: [:e | e filePosition] (AColorSelectorMorph>>#color: "a CompiledMethod(894697472)") filePosition 468 Now, let us see what is in the source file: (SourceFiles at: 1) position: 0; next: 468 '''From Pharo1.0beta of ''''18 October 2009'''' [Latest update: #10487] on 19 October 2009 at 6:18:53 pm''! BracketSliderMorph subclass: #AColorSelectorMorph instanceVariableNames: '''' classVariableNames: '''' poolDictionaries: '''' category: ''Polymorph-Widgets''! !AColorSelectorMorph commentStamp: ''gvc 5/18/2007 13:52'' prior: 0! ColorComponentSelector showing an alpha gradient over a hatched background.! !AColorSelectorMorph methodsFor: ''accessing'' stamp: ''gvc 9/3/2009 ' (SourceFiles at: 1) next: 100. '13:43''! color: aColor "Set the gradient colors." super color: aColor asNontranslucentColor. sel' Yeah, that's what I get in the browser 13:43, it's not the source, it's the time...` In 1.2: (AColorSelectorMorph>>#color: ) filePosition 476 (ZipWriteStream class>>#regressionTestFrom: ) filePosition 16248101 So, someone broke my pointers, or someone have a different sources. Nicolas 2011/3/21 Stéphane Ducasse <stephane.ducasse@inria.fr>:
Hi nicolas
this is strange because I always update all the updates from the latest unstable published by marcus (so yes it can be long) and I never get this error. May be you have a broken sources files or something like that. I update often multiple times especially when I integrate changes to make sure that I do not a strange changes around. and I never get such error. What is your OS?
Stef
On Mar 21, 2011, at 9:53 PM, Nicolas Cellier wrote:
Hi guys, Updating Pharo from the update stream failed many times and keep on failing for me... But maybe it's just me...
So I picked the unstable link from the web site (http://www.pharo-project.org/pharo-download/unstable-core) This drives me to a 13091 image. The first thing I do is select the 'update system' menu... Err, it fails... It seems related to broken sources... Well the image points to which source file ? can't remember the incantation, let just browse it... Err, no source, I can't browse. No matter, I set the browser to decompile mode... Err, this does not work, I still can't browse code. Hmm very bad starting experience, I know what unstable means, but I was dreaming of other words like continuous development.
So ket's open an older image (my own updated 1.2) to find the incantations. (Scanner>>#scanTokens:) fileIndex -> 1, so it seems that the source file is corrupted. (SourceFiles at: 1) -> MultiByteFileStream: '/Users/nicolas/Smalltalk/Squeak/pharo/PharoV10.sources'
Well, for my previous Pharo images, this very file works ! How damned can be the file be changed and the version not incremented ?
Is it me who did something wrong ? Or the update process somewhere did condense the sources ? Or did someone elese messed with the sources ? Is it a known problem ?
Any help appreciated.
Nicolas