source: trunk/third/sendmail/libsm/debug.html @ 19204

Revision 19204, 7.7 KB checked in by zacheiss, 21 years ago (diff)
This commit was generated by cvs2svn to compensate for changes in r19203, which included commits to RCS files with non-trunk default branches.
Line 
1<html>
2<head>
3    <title>libsm : Debugging and Tracing</title>
4</head>
5<body>
6
7<a href="index.html">Back to libsm overview</a>
8
9<center>
10    <h1> libsm : Debugging and Tracing </h1>
11    <br> $Id: debug.html,v 1.1.1.1 2003-04-08 15:06:57 zacheiss Exp $
12</center>
13
14<h2> Introduction </h2>
15
16The debug and trace package provides abstractions for writing trace
17messages, and abstractions for enabling and disabling debug and
18trace code at run time.
19
20<p>
21Sendmail 8.11 and earlier has a <tt>-d</tt> option which
22lets you turn on debug and trace code.
23Debug categories are integers from 0 to 99, with the sole exception
24of "ANSI", which is a named debug category.
25
26<p>
27The libsm debug package supports named debug categories.
28Debug category names have the form of C identifiers.
29For example, <tt>sm_trace_heap</tt> controls the output of trace
30messages from the sm heap package, while <tt>sm_check_heap</tt>
31controls the argument validity checking and memory leak detection
32features of the sm heap package.
33
34<p>
35In sendmail 8.12, the <tt>-d</tt> flag is generalized
36to support both the original style numeric categories, for backwards
37compatibility, and the new style named categories implemented by libsm.
38With this change,
39"-dANSI" is implemented using a libsm named debug category.
40You will be able to set a collection of named debug categories to
41the same activation level by specifying a glob pattern.
42For example,
43<dl>
44<dt>
45    <tt> -dANSI </tt>
46<dd>
47    sets the named category "ANSI" to level 1,
48<dt>
49    <tt> -dfoo_*.3 </tt>
50<dd>
51    sets all named categories matching the glob pattern "foo_*" to level 3,
52<dt>
53    <tt> -d0-99.1 </tt>
54<dd>
55    sets the numeric categories 0 through 99 to level 1, and
56<dt>
57    <tt> -dANSI,foo_*.3,0-99.1 </tt>
58<dd>
59    does all of the above.
60</dl>
61
62<h2> Synopsis </h2>
63
64<pre>
65#include &lt;sm/debug.h&gt;
66
67/*
68**  abstractions for printing trace messages
69*/
70void sm_dprintf(char *fmt, ...)
71void sm_dflush()
72void sm_debug_setfile(SM_FILE_T *)
73
74/*
75**  abstractions for setting and testing debug activation levels
76*/
77void sm_debug_addsettings(char *settings)
78void sm_debug_addsetting(char *pattern, int level)
79
80typedef struct sm_debug SM_DEBUG_T;
81SM_DEBUG_T dbg = SM_DEBUG_INITIALIZER("name", "@(#)$Debug: name - description $");
82
83bool sm_debug_active(SM_DEBUG_T *debug, int level)
84int  sm_debug_level(SM_DEBUG_T *debug)
85bool sm_debug_unknown(SM_DEBUG_T *debug)
86</pre>
87
88<h2> Naming Conventions </h2>
89
90All debug categories defined by libsm have names of the form <tt>sm_*</tt>.
91Debug categories that turn on trace output have names of the form
92<tt>*_trace_*</tt>.
93Debug categories that turn on run time checks have names of the form
94<tt>*_check_*</tt>.
95Here are all of the libsm debug categories as of March 2000:
96
97<table>
98  <tr>
99    <td>Variable name</td>
100    <td>Category name</td>
101    <td>Meaning</td>
102  </tr>
103  <tr>
104    <td>SmExpensiveAssert</td>
105    <td>sm_check_assert</td>
106    <td>enable expensive SM_ASSERT checking</td>
107  </tr>
108  <tr>
109    <td>SmExpensiveRequire</td>
110    <td>sm_check_require</td>
111    <td>enable expensive SM_REQUIRE checking</td>
112  </tr>
113  <tr>
114    <td>SmExpensiveEnsure</td>
115    <td>sm_check_ensure</td>
116    <td>enable expensive SM_ENSURE checking</td>
117  </tr>
118  <tr>
119    <td>SmHeapTrace</td>
120    <td>sm_trace_heap</td>
121    <td>trace sm_{malloc,realloc,free} calls</td>
122  </tr>
123  <tr>
124    <td>SmHeapCheck</td>
125    <td>sm_check_heap</td>
126    <td>enable checking and memory leak detection in sm_{malloc,realloc,free}</td>
127  </tr>
128</table>
129
130<h2> Function Reference </h2>
131
132<dl>
133<dt>
134<tt> SM_DEBUG_INITIALIZER </tt>
135<dd>
136    To create a new debug category, use the SM_DEBUG_INITIALIZER macro
137    to initialize a static variable of type SM_DEBUG_T.  For example,
138<blockquote><pre>
139SM_DEBUG_T ANSI_debug = SM_DEBUG_INITIALIZER("ANSI",
140            "@(#)$Debug: ANSI - enable reverse video in debug output $");
141</pre></blockquote>
142    There is no centralized table of category names that needs to
143    be edited in order to add a new debug category.
144    The sole purpose of the second argument to SM_DEBUG_INITIALIZER
145    is to provide an easy way to find out what named debug categories
146    are present in a sendmail binary.  You can use:
147<blockquote><pre>
148ident /usr/sbin/sendmail | grep Debug
149</pre></blockquote>
150    or:
151<blockquote><pre>
152what /usr/sbin/sendmail | grep Debug
153</pre></blockquote>
154
155
156<dt>
157<tt> void sm_debug_addsetting(char *pattern, int level) </tt>
158<dd>
159    All debug categories default to activation level 0, which means
160    no activity.
161    This function updates an internal database of debug settings,
162    setting all categories whose name matches the specified glob
163    pattern to the specified activation level.  The level argument
164    must be &gt;= 0.
165    <p>
166
167
168<dt>
169<tt> void sm_debug_addsettings(char *settings) </tt>
170<dd>
171    This function is used to process the <tt>-d</tt> command line
172    option of Sendmail 9.x, and of other programs that support the
173    setting of named debug categories.  The settings argument is a
174    comma-separated list of settings; each setting is a glob pattern,
175    optionally followed by a '.' and a decimal numeral.
176    <p>
177
178
179<dt>
180<tt> bool sm_debug_active(SM_DEBUG_T *debug, int level) </tt>
181<dd>
182    This macro returns <tt>true</tt> if the activation level of
183    the statically initialized debug structure <tt>debug</tt>
184    is &gt;= the specified level.
185    The test is performed very efficiently: in the most common case,
186    when the result is <tt>false</tt>, only a single comparison
187    operation is performed.
188    <p>
189    This macro performs a function call only if the debug structure has
190    an unknown activation level.  All debug structures are in this state
191    at the beginning of program execution, and after a call to
192    <tt>sm_debug_addsetting</tt>.
193    <p>
194
195
196<dt>
197<tt> int sm_debug_level(SM_DEBUG_T *debug) </tt>
198<dd>
199    This macro returns the activation level of the specified debug structure.
200    The comparison
201<blockquote><pre>
202sm_debug_level(debug) &gt;= level
203</pre></blockquote>
204    is slightly less efficient than, but otherwise semantically
205    equivalent to
206<blockquote><pre>
207sm_debug_active(debug, level)
208</pre></blockquote>
209    <p>
210
211
212<dt>
213<tt> bool sm_debug_unknown(SM_DEBUG_T *debug) </tt>
214<dd>
215    This macro returns true if the activation level of the specified
216    debug structure is unknown.
217    Here is an example of how the macro might be used:
218<blockquote><pre>
219if (sm_debug_unknown(&FooDebug))
220{
221        if (sm_debug_active(&FooDebug, 1))
222        {
223                ... perform some expensive data structure initializations
224                ... in order to enable the "foo" debugging mechanism
225        }
226        else
227        {
228                ... disable the "foo" debugging mechanism
229        }
230}
231</pre></blockquote>
232    The purpose of using <tt>sm_debug_unknown</tt> in the above example
233    is to avoid performing the expensive initializations each time through
234    the code.  So it's a performance hack.
235    A debug structure is in the "unknown" state at the beginning of
236    program execution, and after a call to <tt>sm_debug_addsetting</tt>.
237    A side effect of calling <tt>sm_debug_active</tt> is that the
238    activation level becomes known.
239    <p>
240
241
242<dt>
243<tt> void sm_dprintf(char *fmt, ...) </tt>
244<dd>
245    This function is used to print a debug message.
246    The standard idiom is
247<blockquote><pre>
248if (sm_debug_active(&BarDebug, 1))
249        sm_dprintf("bar: about to test tensile strength of bar %d\n", i);
250</pre></blockquote>
251    <p>
252
253<dt>
254<tt> void sm_dflush() </tt>
255<dd>
256    Flush the debug output stream.
257    <p>
258
259<dt>
260<tt> void sm_debug_setfile(SM_FILE_T *file) </tt>
261<dd>
262    This function lets you specify where debug output is printed.
263    By default, debug output is written to standard output.
264    <p>
265    We want to allow you to direct debug output to syslog.
266    The current plan is to provide a standard interface for
267    creating an SM_FILE_T object that writes to syslog.
268
269</dl>
270
271</body>
272</html>
Note: See TracBrowser for help on using the repository browser.