Re: [Monotone-devel] Re: Line ending conversion

[Richard Levitte - VMS Whacker]

In message <address@hidden> on Mon, 13 Jun 2005 19:59:40 +0200, Zbynek Winkler 
<address@hidden> said:

zwin> Me too. I like the idea of "normalized storage form" that has
zwin> only one type of EOL but that means that when adding the file we
zwin> should somehow "guess" what normalization to perform.

I think normalization when storing the files is a really bad idea.
The way I see it, normalization should be done at the last possible
moment, if at all.  For monotone, that would basically be when doing a
diff and similar operations (not on checkout!).

The benefits with not normalizing is manyfold, especially for binary
files:

 - if there has been a mistake in guessing the file format, it can be
   corrected afterwards with no loss of data (it's not possible to
   guarantee that a revertion of a CRLF->LF conversion will give you
   the original data back).
 - there will never be a difference between the work file that the
   user just committed and the file stored in the database (!!!!!).
 - there will really be no need for monotone to know the local EOL.
   When doing a diff, it should just recognise all EOL variants (CRLF,
   CR and LF) then figure out the differences and output lines using
   standard functions (it's important that it shows when someone has
   changed the line endings).

I have more concerns, but can't remember them right now.

Cheers,
Richard

-----
Please consider sponsoring my work on free software.
See http://www.free.lp.se/sponsoring.html for details.

-- 
Richard Levitte                         address@hidden
                                        http://richard.levitte.org/

"When I became a man I put away childish things, including
 the fear of childishness and the desire to be very grown up."
                                                -- C.S. Lewis