[llvm-commits] CVS: llvm-www/pubs/2008-08-RTCodegen.html

[Chris Lattner]

Changes in directory llvm-www/pubs:

2008-08-RTCodegen.html updated: 1.1 -> 1.2
---
Log message:

remove ligatures and other non-ascii characters, slice lines.


---
Diffs of the changes:  (+7 -7)

 2008-08-RTCodegen.html |   14 +++++++-------
 1 files changed, 7 insertions(+), 7 deletions(-)


Index: llvm-www/pubs/2008-08-RTCodegen.html
diff -u llvm-www/pubs/2008-08-RTCodegen.html:1.1 llvm-www/pubs/2008-08-RTCodegen.html:1.2
--- llvm-www/pubs/2008-08-RTCodegen.html:1.1	Fri Sep  5 12:36:33 2008
+++ llvm-www/pubs/2008-08-RTCodegen.html	Sun Sep 21 17:14:08 2008
@@ -17,16 +17,16 @@
 <h2>Abstract:</h2>
 <blockquote>
   Typical material systems are based on a collection of properties 
-  that are set on a per–object or per–surface basis usually by evalua- 
-  tion of a description file at run–time. Shading of hit points is then 
+  that are set on a per-object or per-surface basis usually by
+evaluation of a description file at run-time. Shading of hit points is then 
   implemented by interpreting these properties. In order to be able 
-  to support a large feature set, different code paths need to be cre- 
-  ated, e.g. for reflection and refraction effects. Conditions for the 
+  to support a large feature set, different code paths need to be
+created, e.g. for reflection and refraction effects. Conditions for the 
   branches associated with these fragments of functionality need to 
   be evaluated for each shaded ray, which may degrade performance 
-  considerably. Material specific optimization opportunities are also 
-  missed out by having a generic function for all material configura- 
-  tions.  We propose the use of run–time code generation for materials.
+  considerably. Material specific optimization opportunities are also 
+  missed out by having a generic function for all material 
+configurations.  We propose the use of run-time code generation for materials.
 
 </blockquote>