Vm-dev December 2010

vm-dev@lists.squeakfoundation.org

42 participants
109 discussions

Couple of POSSIBLE Cog fixes [WAS] Re: [Vm-dev] [COG] primitiveClosureValue assert is failing with StackVM (possible BUG in bytecodes!!)
by Mariano Martinez Peck 23 Dec '10

23 Dec '10

On Wed, Dec 22, 2010 at 10:34 PM, Eliot Miranda <eliot.miranda(a)gmail.com>wrote: > > Hi Mariano, > > indeed it's a bug. Your fix (objectMemory is: rcvr instanceOf: (self > splObj: ClassBlockClosure) compactClassIndex: > ClassBlockClosureCompactIndex) is the correct one. There are a few other > places where this needs to be done also. I'll roll out new code soon. > > Hi Eliot. Before you roll out new code, I would like to share with you some "fixes" I have to do to Cog, mostly to improve the support of SmallInteger as methods. I attach a .cs but still, some need some explanations. Don't expect them to be the better solution/fix, but hopefully their are useful to show you a bug/problem and you may come with a better solution. So, some notes about the changes: - #primitiveIndexOf: now answers 0 if it is not a ooCompiledMethod. Otherwise, it crash. This method is called by, for example, #primitiveFlushCacheByMethod. If I want to use SmallInteger as methods, I need to implement #flushCache and call the primitive. But this crashes if we don't apply this change. - Igor have changed CoInterpreter >> readImageFromFile: f HeapSize: desiredHeapSize StartingAt: imageOffset and added a <var: #desiredHeapSize type: 'usqInt'>. With this we can compile the MacOS platform code. However, the same change need to be applied to StackInterpreter. - In #activateInterpreterMethodFromMachineCode the assert "self assert: (self primitiveIndexOf: newMethod) ~= 0" fails when using Objects as methods, since "primitiveFunctionPointer" is different than zero, but #primitiveIndexOf: is zero. Check #addNewMethodToCache: to see where the #primitiveInvokeObjectAsMethod is set FOr this problem, there is probably a better way to solve my ugly hack. But at least, the assert doesn't fail ;) - #methodHasCogMethod: I am not sure about this one. What do you think? Because I have some crashes but I cannot garuantee 100% it was because of that. - #primitiveFlushCacheByMethod needs to check if the method is really a method, because it can be a SmallINteger or any other object That's all. Cheers Mariano > best > Eliot > > On Wed, Dec 22, 2010 at 12:03 PM, Mariano Martinez Peck < > marianopeck(a)gmail.com> wrote: > >> >> Hi Eliot. If I compile a StackVM, and I just press the left arrow in a >> code pane of a OB windows, in a pharo image, I get an asser that fails: >> (oop & 1) 19202 >> >> Sorry I bother with these asserts, but I am trying to solve the crashes I >> have and I think starting for the asserts is a good idea. >> >> The assert that fails is in primitiveClosureValue() the part: >> >> blockClosure = longAt(GIV(stackPointer) + (GIV(argumentCount) * >> BytesPerWord)); >> /* begin quickFetchInteger:ofObject: */ >> oop = longAt((blockClosure + BaseHeaderSize) + (ClosureNumArgsIndex << >> ShiftForWord)); >> assert((oop & 1)); >> numArgs = (oop >> 1); >> if (!(GIV(argumentCount) == numArgs)) { >> /* begin primitiveFail */ >> if (GIV(primFailCode) == 0) { >> GIV(primFailCode) = 1; >> } >> return; >> } >> >> >> numArgs results to be a large number (using #printNum: was 718613205), >> and if I try to do a #printOop: I got a " 0x55aa55aa0x55aa55aa is not on >> the heap" >> >> Now, of course, the "if (!(GIV(argumentCount) == numArgs)) " fails. >> >> What it is funny is that I did a #printOop: of blockClosure instVar, and >> this is what I get: 0x1f59b578: a(n) Association >> >> Association? WTF ? so...I guess something weird is happening. I continue >> debugging it..and I noticed that: >> >> - Several classes were called in addition to Association,like Character. >> These seems to be compact classes and implement #value and #value: ;) >> - Both, #bytecodePrimValue and #bytecodePrimValueWithArg were not calling >> #primitiveClosureValue. So, "isBlock" was false... >> >> I tried modifying #bytecodePrimValueWithArg and #bytecodePrimValue this >> line: >> >> isBlock := objectMemory is: rcvr instanceOf: ClassBlockClosurecompactClassIndex: ClassBlockClosureCompactIndex. >> >> to this one: >> >> isBlock := objectMemory is: rcvr instanceOf: (self splObj: >> ClassBlockClosure) compactClassIndex: ClassBlockClosureCompactIndex. >> >> WIth this change, at least, the assert doesn't fail anymore. I made >> progress (maybe). But still, for normal closures, "isBlock" seems to be >> false. So.....all block values are being managed as a normal send ???? >> >> Now, I don't understand why ClassBlockClosureCompactIndex is 0. I think >> this may be the problem. Why it is not 12? (in Pharo Smalltalk >> compactClassesArray at: 12 ->>> BlockClosure.) >> I notice that ObjectMemory >> initializeCompactClassIndices does: >> >> ClassBlockClosureCompactIndex := 0 "12". "Prospective. Currently >> TranslatedMethod class" >> >> >> Notice that this ZERO changes the sematic of #is: oop instanceOf: classOop >> compactClassIndex: compactClassIndex >> called from the #bytecodePrimValue and bytecodePrimValueWithArg >> >> Finally, I changed the mentioned line to this: >> >> isBlock := objectMemory is: rcvr instanceOf: (self splObj: >> ClassBlockClosure) compactClassIndex: 12. >> >> and finally, with block closures, it is calling the primitive, and for >> Association it doesn't fail the assert. ahh and the image doesn't crash ;) >> >> I don't know if this is a bug, if my solution was ok, nor it impact. >> Please let me know. >> >> Cheers >> >> Mariano >> >> >> >> >> >> > >

4 4

AA StrikeFonts and paragraph width
by Matthew Fulmer 23 Dec '10

23 Dec '10

In testing Tweak on Squeak 4.2, I discovered an inconsistency in the text layout engine: TextComposer and MultiCharacterScanner disagree with each other on the width of paragraphs made of AA StrikeFonts. They agree for other types of fonts. This was causing non-word-wrapped paragraphs in Tweak to lose their last character. Issue and test case are in mantis: http://bugs.squeak.org/view.php?id=7589

1 1

VM Maker: CMakeVMMaker-Igor.Stasenko.1.mcz
by squeak-dev-noreply＠lists.squeakfoundation.org 23 Dec '10

23 Dec '10

Igor Stasenko uploaded a new version of CMakeVMMaker to project VM Maker: http://www.squeaksource.com/VMMaker/CMakeVMMaker-Igor.Stasenko.1.mcz ==================== Summary ==================== Name: CMakeVMMaker-Igor.Stasenko.1 Author: Igor.Stasenko Time: 23 December 2010, 2:59:53 am UUID: 22d679bb-063a-4114-bf33-4599d79f0d71 Ancestors: - initial commit. try CMakeCogVMGenerator new generate: CogMacOSConfig

1 0

[4.2] - VM <-> image release coordination
by Chris Muller 23 Dec '10

23 Dec '10

Do we have any items needed the 4.2 image that would be essential for the upcoming VM release?

2 1

Return key in iOS?
by Casey Ransberger 22 Dec '10

22 Dec '10

So I have been hacking a bit in iOS land. I looked at Bert's Etoys example and ultimately settled on a similar approach to making the keyboard visible. Everything is more or less golden, except that the return key doesn't do anything. Smalltalk methods look funny all on one line:/ It's quite likely something I'm doing wrong in Cocoa land, but it occurs to me that the return key might be putting linefeeds into whatever text input buffer is there, and that might explain why Squeak isn't advancing to the next line. Is this a known issue? -- Casey Ransberger

2 1

SoundPlugin in Cog
by Matthew Fulmer 22 Dec '10

22 Dec '10

I'm testing Cobalt with the Cog VM. I'm using Eliot's precompiled VMs at http://www.mirandabanda.org/files/Cog/VM/VM.r2316/ They don't include SoundPlugin, and this is rather needed by Cobalt. Is SoundPlugin supported on Cog? If so, I'd really appreciate it if someone could compile a VM with it; none of us really know how to

2 3

[Cog] Trying to understand #assert:
by Mariano Martinez Peck 22 Dec '10

22 Dec '10

Hi Eliot. I noticed that in the Cog VM there are a lot of asserts in the code. I am trying to understand them but I have a couple of questions: - What happens with the code execution when an assert fails? it stops the execution? it continues? it raises an error? I tried to follow to the definition of assert() but I only came to this: # define assert(expr) ((expr)||(warning(#expr " " __stringifyNum(__LINE__)),0)) So....after time, I understood that when I compile in "development" and I run with GDB, and I have something printed in the console like: (primitiveIndexOf(GIV(newMethod))) != 0 8913 this means that in the line 8913 the assert has failed? so, do I understand correct? when an assert fails, the code continues to execute, but it just prints the expresion + number of line in the console? if it doesn't fail, it doesn't print. thanks in advance, Mariano

2 3

VM Maker: VMMaker-dtl.212.mcz
by squeak-dev-noreply＠lists.squeakfoundation.org 21 Dec '10

21 Dec '10

Dave Lewis uploaded a new version of VMMaker to project VM Maker: http://www.squeaksource.com/VMMaker/VMMaker-dtl.212.mcz ==================== Summary ==================== Name: VMMaker-dtl.212 Author: dtl Time: 21 December 2010, 11:14:23 am UUID: 51f37d0c-92f0-42af-9ade-fd6d48cf50ef Ancestors: VMMaker-dtl.211 VMMaker 4.4.3 Fix assignment to block variable in #saveConfig. Use StandardFileMenu if UIManager not present, for support of older images.

1 0

VM Maker: VMMaker-dtl.212.mcz
by squeak-dev-noreply＠lists.squeakfoundation.org 21 Dec '10

21 Dec '10

Dave Lewis uploaded a new version of VMMaker to project VM Maker: http://www.squeaksource.com/VMMaker/VMMaker-dtl.212.mcz ==================== Summary ==================== Name: VMMaker-dtl.212 Author: dtl Time: 21 December 2010, 10:46:08 am UUID: 8403f4f1-e690-4806-a014-767721afdec2 Ancestors: VMMaker-dtl.211 VMMaker 4.4.3 Fix assignment to block variable in #saveConfig. Use StandardFileMenu if UIManager not present, for support of older images.

2 2

Re: [Pharo-users] Unicode in Pharo1.1
by Mariano Martinez Peck 20 Dec '10

20 Dec '10

On Fri, Dec 17, 2010 at 4:50 PM, Stanislav Paskalev <kshorg(a)gmail.com>wrote: > Hi, > I also have to write in cyrilic in pharo 1.1 / 1.1.1 and it works. > > You have to first update the fonts from system and then set new, > unicode enabled fonts on all the elements. Updating the fonts just > adds them to a list of known fonts, it doesn't make the image to use > them in other words. > Also, with the latest build (4 oct) of CogVM, updating fonts from > system does not work on any system (win,mac,linux) - so use the > standart VM. > > Maybe we should report this to Eliot? cheers mariano > Best regards, > Stanislav Paskalev > > > > On Fri, Dec 17, 2010 at 5:45 PM, qpogo <alex.albitov(a)gmail.com> wrote: > > Hi all, I'm a real newbie in Pharo and Smalltalk. I want to type in > Cyrillic > > and Hebrew. However I get question marks. I tried to change fonts, but I > > have only to available fonts: * Accuny * Bitmap DejaVu Sans Which don't > have > > Unicode glyphs. When I try to update fonts from the system, I don't see > any > > result. Thank you. Alex > > ________________________________ > > View this message in context: Unicode in Pharo1.1 > > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. > > > >

4 7

← Newer
1
2
3
4
5
6
7
8
9
10
11
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev December 2010