On Mar 12, 2004, at 12:57 AM, Doug Way wrote:

>> Why don't we just convert line endings on file-in?
>
> Probably a good idea.  (I'm trying to think if there are any rare 
> cases in which we'd want to allow a linefeed character to appear in a 
> method.  But making someone jump through hoops in order to file in a 
> method like that might be reasonable anyway...)

Linefeed characters should be allowed inside string literals, which is 
something that one might want to do in test cases. MonticelloCVS does 
this, as will any application that exports text to the local platform. 
(Assuming it has a good test suite, that is.)

Ok, it's not a terribly common case, but I think it *is* something that 
should work without hoop-jumping.

It'd also support Tim's assertion that having fileOuts be "good 
citizens" of a platform's text conventions is not a requirement. Anyone 
manually editing fileOuts with a text editor is definitely messing 
around under the hood, and should know about the format of a fileOut. 
It's not like there are no tools that will handle CRs properly. Heck, 
even unix command line tools can usually be made to work right.

tail -f Squeak.changes | tr '\r' '\n'