Am 2012-03-11 um 15:10 schrieb Nicolas Cellier:

> Le 10 mars 2012 22:45, Tobias Pape <Das.Linux at gmx.de> a écrit :
> 
> Am 2012-03-10 um 17:27 schrieb commits at source.squeak.org:
> 
> > Nicolas Cellier uploaded a new version of Monticello to project The Inbox:
> > http://source.squeak.org/inbox/Monticello-nice.500.mcz
> >
> > ==================== Summary ====================
> >
> > Name: Monticello-nice.500
> > Author: nice
> > Time: 10 March 2012, 6:27:05.52 pm
> > UUID: 432c3785-149a-4822-ace4-e14b1b792dfa
> > Ancestors: Monticello-bf.496
> >
> > Use a ReferenceStream rather than a DataStream in #serializeInBinary:
> > because storing some object require a ReferenceStream
> > The bug did show up when I added tokens to MCMethodDefinition as shown below
> 
> Note that this change possibly can break compatibily with Gemstone/Pharo
> folks, although I hope, it does not :)
> 
> best
>        -tobias
> 
> I don't know...
> An alternative would be to define DataStream>>insideASegment ^false
> That's just showing how fragile this part of the code is...

I was just mentioning.
 I had weird results when dealing with source code containing
“extended” characters (as extended as an “ can be) and had them
travel between the worlds (Squeak/Gemstone/Pharo in this case).
 I just want to raise the point here, that the whole Monticello implementation
in all three platforms I know is based on the…buggy?… DataStream.
 In fact, that was why I started the Monticello/Fuel thing some month ago,
in the first place.

We just should be aware.

just my 2ct.

Best
	-Tobias