line endings redux (was: 0.15 is)

[Daniel Parks]

On Mar 5, 2007, at 8:18 PM, John Whitley wrote:
> * Binary file detection/handling.  EOL conversion can corrupt  
> files, so care
>   is needed here.

I want to point out an alternate way to implement this: files are  
committed as they are now, but are transformed when they're accessed  
in the repository. In other words, files are normalized on read,  
instead of write. This could be used for keywords (e.g. $Id$), too.

I imagine this would make annotation harder, but it is very safe.  
This doesn't seem like it would require a working tree format bump,  
but I don't really understand bzr internals.

Another way to get the same effect is to normalize line endings on  
commit and then record which line endings were changed.

Just throwing it out there -- I'm not advocating one way or another.  
Hope this hasn't already been suggested...

Oh, and Wikipedia has a list of line endings:
http://en.wikipedia.org/wiki/Newline#Unicode

Daniel