> On 11.06.2019, at 20:07, tim Rowledge <tim at rowledge.org> wrote:
> 
> 
> 
>> On 2019-06-11, at 10:06 AM, Tobias Pape <Das.Linux at gmx.de> wrote:
>>> 
>> 
>> Powerwash all the things!
> 
> I like the general approach of cleaning things up but do remember to test that the older images are ok afterwards. Eliot has nicely explained the requirement many times in the past.
> 
That's true and I am in no way opposed to that.

However, with (a) the change from “traditional” Squeak encoding (ie, macroman) to latin1 (aka iso-8859-1, hence "squeakToIso") and (b) then the  “rebrand” of bytestring-is-latin1/widestring-is-utf32 to "just"  unicode[1], the (a) step became lost to many people.

_if_ old images would not work with the changes Nicolas proposes, they would have been hosed long ago anyway.

Best regards
	-Tobias



[1]: this is the most amazing encoding "hack" I have seen, and I think that's actually why Unicode for codepoints <256 is the way it is: that latin1 strings are just that: valid unicode. I know it was not done _because_ of Squeak, but Andreas obviously saw then chance and took it.