[oprof-cvs] CVS: oprofile ChangeLog,1.964.2.76,1.964.2.77 TODO,1.203.2.30,1.203.2.31

[Philippe E. <ph...@us...>]

Update of /cvsroot/oprofile/oprofile
In directory sc8-pr-cvs1:/tmp/cvs-serv14972

Modified Files:
      Tag: pp-interface-branch
	ChangeLog TODO 
Log Message:

don't use profile_t to read sample file header

regards,
Phil


Index: ChangeLog
===================================================================
RCS file: /cvsroot/oprofile/oprofile/ChangeLog,v
retrieving revision 1.964.2.76
retrieving revision 1.964.2.77
diff -u -d -r1.964.2.76 -r1.964.2.77
--- ChangeLog	8 May 2003 23:53:15 -0000	1.964.2.76
+++ ChangeLog	9 May 2003 03:41:56 -0000	1.964.2.77
@@ -1,5 +1,14 @@
 2003-05-09  Philippe Elie  <ph...@wa...>
 
+	* libop++/op_header.h:
+	* libop++/op_header.cpp: read_header() new
+	* libop++/Makefile.in: we depend on libdb
+
+	* pp/opannotate.cpp:
+	* pp/opreport.cpp: better way to get sample file header
+
+2003-05-09  Philippe Elie  <ph...@wa...>
+
 	* libpp/format_output.cpp: remove redudant map formater
 	* pp/opannotate.cpp: spurious cout
 

Index: TODO
===================================================================
RCS file: /cvsroot/oprofile/oprofile/TODO,v
retrieving revision 1.203.2.30
retrieving revision 1.203.2.31
diff -u -d -r1.203.2.30 -r1.203.2.31
--- TODO	8 May 2003 19:54:54 -0000	1.203.2.30
+++ TODO	9 May 2003 03:41:56 -0000	1.203.2.31
@@ -93,6 +93,10 @@
   of add_sample_filename() (we don't want to do that for all the profiles at once)
   we have a smaller constructor, and we still delay the actual processing of the
   data until later. Comments phil ?
+  We can do that but I think it's better to sperate the error checking.
+  It will allow more freedom on how to refactorize partition_files and
+  image_set
+  
 
  o #pp/opreport -l -p /lib/modules/2.5.69/kernel/ oprofile.ko
    No samples files found: profile specification too strict ?
@@ -103,6 +107,11 @@
    ...
 
    It would be nice to allow at least the first alternate form.
+
+ o test with 2.95 if cout.rdbuf(os.rdbuf()); work even when there is non
+  flushed data in cout buffer. It's also not clear if at exit() we have not
+  a problem (who is the owner of the buffer and in what order of and cout will
+  be destroyed ?)
 
  o fix John movement on 2.5: odb_insert failure (and message error corrupted ?)
    - haven't been able to reproduce yet, I just get too many files open instead ;)