[Vm-dev] Re: [squeak-dev] Squeak4.3alpha-11481 build
eliot.miranda at gmail.com
Wed Jun 15 17:45:17 UTC 2011
On Mon, Jun 13, 2011 at 12:39 PM, Edgar J. De Cleene <edgardec2005 at gmail.com
> On 6/13/11 2:55 PM, "Levente Uzonyi" <leves at elte.hu> wrote:
> > This one looks better. Did you run the tests before saving the image?
> > I'm asking because tests may have unwanted side effects.
> I know. Run after the .zip go to ftp and not save
> > I guess your VM is pretty old and probably that's the cause of most of
> > those failures and errors. Running the tests in this image gives the
> > following list on my pc with Cog r2363:
> Yes, the VM is old.
> Still puzzled why different VM (Mac and Windows) give us different results.
> Test don't should be VM independent ?
Not if the VM has bugs and/or limitations, and not if the tests in question
specifically test VM functionality. If the VM is buggy then potentially any
and all tests can fail. But more generally useful is several sets of tests
that test VM functionality such as CroquetVMTests CogVMTests
IslandVMTweaksTestCase STackVMTest Win32VMTest FFIPluginTests
FloatMathPluginTests JPEGReadWriter2PluginTest MiscPrimitivePluginTest
CompilerTests (mirror primitive tests). (Yes, these need rationalizing, but
this takes time). Then as far as platform independence goes, things like the
FFI and OSProcess are windows into platform specific behaviour, and so one
can't expect platform independence.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Vm-dev