[Checkins] SVN: zc.zservertracelog/trunk/ Tweeked documentation.

[Jim Fulton]

Log message for revision 110651:
  Tweeked documentation.
  

Changed:
  U   zc.zservertracelog/trunk/README.txt
  U   zc.zservertracelog/trunk/setup.py
  U   zc.zservertracelog/trunk/src/zc/zservertracelog/README.txt

-=-
Modified: zc.zservertracelog/trunk/README.txt
===================================================================
--- zc.zservertracelog/trunk/README.txt	2010-04-08 15:30:52 UTC (rev 110650)
+++ zc.zservertracelog/trunk/README.txt	2010-04-08 15:33:19 UTC (rev 110651)
@@ -7,10 +7,33 @@
 request.  Each log entry starts with a record type, a request
 identifier and the time.  Some log records have additional data.
 
+To create a trace log, you need to:
 
-Trace Codes
-===========
+- Include the zc.zservertracelog configuration in your site zcml file::
 
+    <include package="zc.zservertracelog" />
+
+- Define where messages to the 'zc.tracelog' logger should go. In your
+  zope.cond file, use something like::
+
+    <logger>
+      name zc.tracelog
+      propagate false
+
+      <logfile>
+        format %(message)s
+        path /home/jim/p/zc.zservertracelog/dev/trace.log
+      </logfile>
+
+    </logger>
+
+
+The analysis script, tracereport, can be used to analyze the trace
+log. I recommend the html output option.
+
+Trace log records
+=================
+
 - Request begins:
 
   B -1214390740 2007-04-27T20:16:55.582940 GET /
@@ -27,6 +50,23 @@
 
   C -1214390740 2007-04-27T20:16:55.703829
 
+- Database activity
+
+
+  D -1223774356 2007-04-27T20:16:55.890371 42 0 x 2 1
+
+  The data includes objects loaded and saved for each database except
+  databases for which there was no activity.  Note that it's common
+  for the main database to be unnamed, and the data often starts with
+  objects loaded and saved for the main database.
+
+  In the example above, 42 objects were loaded from the unnamed
+  database. Two objects were loaded from and one saved to the database
+  named 'x'.
+
+  If requests are retried due to conflict errors, then there will be
+  multiple 'D' records.
+
 - Application done:
 
   A -1223774356 2007-04-27T20:16:55.890371 500 84
@@ -44,27 +84,3 @@
 Tracelog extension records are prefixed with a '-':
 
   - -1223774356 2008-09-12T15:51:05.559302 zc.example.extension message
-
-To create a trace log, include the zc.zservertracelog package in your
-site ZCML configuration.  Also, define a logger section in zope.conf:
-
-.. warning: The logger section is named "zc.tracelog" not "zc.zservertracelog";
-    that might change one day.
-
-::
-
-  <logger>
-    name zc.tracelog
-    propagate false
-
-    <logfile>
-      format %(message)s
-      path /home/jim/p/zc.zservertracelog/dev/trace.log
-    </logfile>
-
-  </logger>
-
-Where, of course, you'll need to specify a log file path.
-
-The analysis script, tracereport, can be used to analyze the trace
-log. I recommend the html output option.

Modified: zc.zservertracelog/trunk/setup.py
===================================================================
--- zc.zservertracelog/trunk/setup.py	2010-04-08 15:30:52 UTC (rev 110650)
+++ zc.zservertracelog/trunk/setup.py	2010-04-08 15:33:19 UTC (rev 110651)
@@ -35,11 +35,7 @@
     long_description=(
         read('README.txt')
         + '\n\n'
-        + 'Detailed Documentation\n'
-        + '**********************\n'
         + '\n\n'
-        + read('src', 'zc', 'zservertracelog', 'README.txt')
-        + '\n\n'
         + read('CHANGES.txt')
     ),
     license='ZPL 2.1',

Modified: zc.zservertracelog/trunk/src/zc/zservertracelog/README.txt
===================================================================
--- zc.zservertracelog/trunk/src/zc/zservertracelog/README.txt	2010-04-08 15:30:52 UTC (rev 110650)
+++ zc.zservertracelog/trunk/src/zc/zservertracelog/README.txt	2010-04-08 15:33:19 UTC (rev 110651)
@@ -283,8 +283,9 @@
     A 49234448 2010-04-07 17:03:41.230364 200 ?
     E 23418928 2008-08-26 10:55:00.000000
 
-Here we got multiple D records due to conflicts. We show database
-activity for those databases for which there was any. The databases
-are sorted by name, with the unnamed database coming first.  For each
-database, the number of object's read and written are provided.
+Here we got multiple D records due to (simulated) conflicts. We show
+database activity for those databases for which there was any. The
+databases are sorted by name, with the unnamed database coming first.
+For each database, the number of object's loaded and saved are
+provided.