Magma test failure

[Chris Muller]

Hmm, that this default seems to have disrupted the "one-click" nature
of the Magma install bugs me.

My choices are:

  1) revert the default WriteBarrier setting back to "off"
  2) include NewCompiler in the Magma install.
  3) extract the ByteCodeGenerator from NewCompiler package, the only
class WriteBarrier needs to make use of..

I'm loathe to do #2, Pharo folks may not want the NewCompiler in their
image (even though it should be harmless).

Any opinions?


On Wed, Sep 9, 2009 at 9:54 AM, Chris Muller <asqueaker at gmail.com> wrote:
> Hi, sorry about that.  This *should* be mentioned in the documentation
> (I will update it).
>
> To move forward, the choices are either:
>
>  1)  install NewCompiler from SqueakMap
>  2)  Evaluate:  MagmaTestCase allowWriteBarrier: false
>
> And restart the test suite.
>
> It was a tough choice to turn on WriteBarrier by default, but the the
> commit-rate can be significantly improved when it is on, and it seems
> to work, so why not..
>
> Regards,
>  Chris
>
> On Wed, Sep 9, 2009 at 1:03 AM, Ross Boylan <ross at biostat.ucsf.edu> wrote:
>> Shortly after starting the test _magmaTestclient1.image show
>> "MagmaUserError: WriteBarrier requires NewCompiler."
>>
>> Hmm, I was going to copy the tail end of the Transcript of
>> _magmaTestConductor into this message, but that image is unresponsive.
>>
>> I'm running on Debian GNU/Linux with squeak-vm 3.10.3+svn1902.dfsg-1 and
>> the 3.10 image downloaded from
>> http://ftp.squeak.org/3.10/Squeak3.10.2-7179-basic.zip.
>>
>> Magma was tested with 3.10.4 (VM).  Is the new compiler something added
>> with the 3.10.4 release?  I see a package by that name on SM (marked as
>> very alpha).
>>
>>
>> _______________________________________________
>> Magma mailing list
>> Magma at lists.squeakfoundation.org
>> http://lists.squeakfoundation.org/mailman/listinfo/magma
>>
>