Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 119 by luc.fabr...(a)gmail.com: SerialPlugin
primitiveOpenPortByName: ...
http://code.google.com/p/cog/issues/detail?id=119
Since VMMaker-oscog-EstebanLorenzano.166, there is a bug in the
implementation of
SerialPlugin>>primitiveSerialPortOpenByName: portName baudRate: baudRate
stopBitsType: stopBitsType parityType: parityType dataBits: dataBits
inFlowControlType: inFlowControl outFlowControlType: outFlowControl
xOnByte: xOnChar xOffByte: xOffChar
Find a (small) patch attached.
It is simple and it works.
Just need integration.
Luc
Attachments:
VMMaker-oscog-LucFabresse.229.mcz 2.6 MB
Status: Accepted
Owner: nicolas....(a)gmail.com
Labels: Type-Defect Priority-High
New issue 117 by nicolas....(a)gmail.com: primitiveRemLargeIntegers is
incorrect for negative quotient
http://code.google.com/p/cog/issues/detail?id=117
Now that primitive 20 is connected (COG 2662), I had a chance to run some
tests and discover a bug (stupid copy/paste from quo)...
The result should be negated iff receiver negative.
Attachments:
InterpreterPrimitives-primitiveRemLargeIntegers.st 871 bytes
Status: Accepted
Owner: nicolas....(a)gmail.com
Labels: Type-Defect Priority-Medium
New issue 115 by nicolas....(a)gmail.com: FFIPlugin unsignedShortAt: answer a
signed short
http://code.google.com/p/cog/issues/detail?id=115
(#[255 255 255 255] unsignedShortAt: 1) -> -1
This is exactly http://bugs.squeak.org/view.php?id=1415
It was reported in 2005, integrated in 2006 in VMMaker-tpr.55.mcz...
...and is still missing in COG branch
Attachments:
CogIssue115_FFIPlugin_unsignedShortAt_isSigned.st 1.6 KB
Status: Accepted
Owner: nicolas....(a)gmail.com
Labels: Type-Defect Priority-Medium
New issue 114 by nicolas....(a)gmail.com: BMPReadWriterPlugin odes not check
negative width and is subject to buffer overflow
http://code.google.com/p/cog/issues/detail?id=114
This was reported and fixed by Andreas in 2006 at
http://bugs.squeak.org/view.php?id=4360
And integrated in 2007 in VMMaker-tpr.64...
I fail to see why it should not be in COG branch too.
Attachments:
CogIssue114_fix_a_BMPReadWriterPlugin_potentialBufferOverflow.st 3.4 KB
Status: Accepted
Owner: nicolas....(a)gmail.com
Labels: Type-Defect Priority-High
New issue 112 by nicolas....(a)gmail.com: Literal > 16rFFFFFFFF should better
be generated as unsigned long long
http://code.google.com/p/cog/issues/detail?id=112
This was one change of issue 92 which was forgotten.
Currently, gcc is kind enough to correct 0xffffffffffffffffUL by itself,
but it issues a warning and that stinks like Undefined Behaviour.
Find attached correction based on .oscog-emm.241
Status: Accepted
Owner: nicolas....(a)gmail.com
Labels: Type-Enhancement Priority-Medium Performance Maintainability
New issue 111 by nicolas....(a)gmail.com: Use direction aware << and >>
instead of bitShift: to avoid useless Runtime test
http://code.google.com/p/cog/issues/detail?id=111
As reported in vm-dev mailing list "bitShift: and runtime sign discussion"
http://comments.gmane.org/gmane.comp.lang.smalltalk.squeak.vm.devel/9261
in (expr bitShift: shift),
some shift have well known direction for the programmer,
but the CCodeGenerator is unable to guess if shift is not literal.
This results in useless runtime tests.
Worse, it can increase the number of C compiler warnings after inlining.
I attach some changes to VMMaker (based on .oscog-eem.241 branch).
For sound, there is a shift sign discussion already, so we can eventually
use directed shift too
(based on trunk -ul.32 branch).
Attachments:
VMMaker_replace_bitShift_with_directedShift.cs 12.1 KB
Sound_replace_bitShift_with_directedShift.cs 1.8 KB
Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 108 by stm...(a)gmail.com: Access to stackPages broken in
Stack/CoInterpreter>>#commonReturn
http://code.google.com/p/cog/issues/detail?id=108
One of the branches in Stack/CoInterpreter>>#commonReturn does a `self
freeStackPage: thePage` where all other places in the code do `stackPages
freeStackPage: thePage`.
#freeStackPage: is not defined on self, so I suppose the attached patch is
necessary.
The bug should be in the CogVM branch and the Pharo branches of the
codebase.
Attachments:
StackPages Access.1.cs 11.5 KB
Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 107 by stm...(a)gmail.com: Bit-rot in the Pharo CogVM code based
(Simulator)
http://code.google.com/p/cog/issues/detail?id=107
The attached change set fixes bit rot in the *InterpreterSimulator and
related classes.
Here a brief summary:
- VMClass>>#doOrDefer: now uses ProcessBrowser>>#isUIProcess: instead of
Project
- it replaces the use of #asDisplayText with StringMorph>>#contents:.
- use of ThreadStream>>#on: is replaced with ThreadSafeTranscript>>#new
- Utilities is replaced by UIManager>>#default
- and #primitiveGetAttribute now uses the non-depricated `Smalltalk vm
getSystemAttribute: attr`
Attachments:
Fix simulator.3.cs 30.0 KB
Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 106 by stm...(a)gmail.com: Bit-rot in NOMS>>#sqGrowMemory:By: and
IS>>#ioUTCMicrosecondsNow
http://code.google.com/p/cog/issues/detail?id=106
In InterpreterSimulator the method #ioUTCMicrosecondsNow is missing.
I just added this:
ioUTCMicrosecondsNow
"STEFAN: not entierly sure what to do with this... but the method is
missing."
^ Time millisecondClockValue
In the NewObjectMemorySimulator>>#sqGrowMemory:By:, there is a reference to
coInterpreter which receives are #memory: message. However in the new
hierarchy, where the interpreter is not an ObjectMemory anymore, that
message is not implemented.
So, I removed it...
Not sure whether either of these solutions is appropriate, but would be
good to get the problems fixed and have the simulators usable.
Attachments:
more-bit-rot.1.cs 857 bytes