Re: [Monotone-devel] Two new lists to get commit logs

[Richard Levitte - VMS Whacker]

In message <address@hidden> on Thu, 14 Apr 2005 21:04:47 -0700, Nathaniel Smith 
<address@hidden> said:

njs> On Fri, Apr 15, 2005 at 03:43:57AM +0200, Richard Levitte - VMS Whacker 
wrote:
njs> > In message <address@hidden> on Thu, 14 Apr 2005 19:36:54 -0600, Derek 
Scherger <address@hidden> said:
njs> > 
njs> > derek> Hmm... one thing that comes to mind is that monotone
njs> > derek> diff currently lists out "diffs" for binary files, which
njs> > derek> are sometimes rather large and might not really be what
njs> > derek> you want in your in-box.
njs> > 
njs> > Hmm, true.  What would be the best way to detect those?  If I
njs> > can do that easily, it wouldn't be too hard for me to figure is
njs> > should print those diffs...
njs> 
njs> Maybe fix the monotone bug that makes it happen? ;-)

I was a bit clumsy in the way I wrote that.  My question is, how do I
correctly detect a binary blob from text?  What can I assume?  Can I
assume all the text is mainly ASCII, so anything with more than, say,
20% are bytes with the highest bit set, it's a binary?  I'm definitely
not an expert at character encoding in all sorts of languages, so I
feel a bit unsafe in this area.

Either way, regardless of the place I would have to fix this, it comes
down to defining "what's a binary" in a computationally sane way.  I
need help with idea here, please.

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