hdf5/doc/html/Debugging.html
Frank Baker ad81c8163e [svn-r3823]
Purpose:
    Update "Describes HDF5 Release..." footer from R1.4 to
    "Release 1.5, Unreleased Development Branch"
Platforms tested:
    IE 5
2001-04-18 17:03:09 -05:00

504 lines
16 KiB
HTML

<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML//EN">
<html>
<head>
<title>Debugging HDF5 Applications</title>
</head>
<body bgcolor="#FFFFFF">
<hr>
<center>
<table border=0 width=98%>
<tr><td valign=top align=left>
<a href="H5.intro.html">Introduction to HDF5</a>&nbsp;<br>
<a href="RM_H5Front.html">HDF5 Reference Manual</a>&nbsp;<br>
<a href="index.html">Other HDF5 documents and links</a>&nbsp;<br>
<!--
<a href="Glossary.html">Glossary</a><br>
-->
</td>
<td valign=top align=right>
And in this document, the
<a href="H5.user.html"><strong>HDF5 User's Guide:</strong></a>&nbsp;&nbsp;&nbsp;&nbsp;
<br>
<a href="Files.html">Files</a>&nbsp;&nbsp;
<a href="Datasets.html">Datasets</a>&nbsp;&nbsp;
<a href="Datatypes.html">Datatypes</a>&nbsp;&nbsp;
<a href="Dataspaces.html">Dataspaces</a>&nbsp;&nbsp;
<a href="Groups.html">Groups</a>&nbsp;&nbsp;
<br>
<a href="References.html">References</a>&nbsp;&nbsp;
<a href="Attributes.html">Attributes</a>&nbsp;&nbsp;
<a href="Properties.html">Property Lists</a>&nbsp;&nbsp;
<a href="Errors.html">Error Handling</a>&nbsp;&nbsp;
<br>
<a href="Filters.html">Filters</a>&nbsp;&nbsp;
<a href="Caching.html">Caching</a>&nbsp;&nbsp;
<a href="Chunking.html">Chunking</a>&nbsp;&nbsp;
<a href="MountingFiles.html">Mounting Files</a>&nbsp;&nbsp;
<br>
<a href="Performance.html">Performance</a>&nbsp;&nbsp;
Debugging&nbsp;&nbsp;
<a href="Environment.html">Environment</a>&nbsp;&nbsp;
<a href="ddl.html">DDL</a>&nbsp;&nbsp;
</td></tr>
</table>
</center>
<hr>
<h1>Debugging HDF5 Applications</h1>
<h2>Introduction</h2>
<p>The HDF5 library contains a number of debugging features to
make programmers' lives easier including the ability to print
detailed error messages, check invariant conditions, display
timings and other statistics, and trace API function calls and
return values.
<dl>
<dt><b>Error Messages</b>
<dd>Error messages are normally displayed automatically on the
standard error stream and include a stack trace of the library
including file names, line numbers, and function names. The
application has complete control over how error messages are
displayed and can disable the display on a permanent or
temporary basis. Refer to the documentation for the H5E error
handling package.
<br><br>
<dt><b>Invariant Conditions</b>
<dd>Unless <code>NDEBUG</code> is defined during compiling, the
library will include code to verify that invariant conditions
have the expected values. When a problem is detected the
library will display the file and line number within the
library and the invariant condition that failed. A core dump
may be generated for post mortem debugging. The code to
perform these checks can be included on a per-package bases.
<br><br>
<dt><b>Timings and Statistics</b>
<dd>The library can be configured to accumulate certain
statistics about things like cache performance, datatype
conversion, data space conversion, and data filters. The code
is included on a per-package basis and enabled at runtime by
an environment variable.
<br><br>
<dt><b>API Tracing</b>
<dd>All API calls made by an application can be displayed and
include formal argument names and actual values and the
function return value. This code is also conditionally
included at compile time and enabled at runtime.
</dl>
<p>The statistics and tracing can be displayed on any output
stream (including streams opened by the shell) with output from
different packages even going to different streams.
<h2>Error Messages</h2>
<p>By default any API function that fails will print an error
stack to the standard error stream.
<p>
<center>
<table border align=center width="100%">
<tr>
<td>
<p><code><pre>
HDF5-DIAG: Error detected in thread 0. Back trace follows.
#000: H5F.c line 1245 in H5Fopen(): unable to open file
major(04): File interface
minor(10): Unable to open file
#001: H5F.c line 846 in H5F_open(): file does not exist
major(04): File interface
minor(10): Unable to open file
</code></pre>
</td>
</tr>
</table>
</center>
<p>The error handling package (H5E) is described
<a href="Errors.html">elsewhere</a>.
<h2>Invariant Conditions</h2>
<p>To include checks for invariant conditions the library should
be configured with <code>--disable-production</code>, the
default for versions before 1.2. The library designers have made
every attempt to handle error conditions gracefully but an
invariant condition assertion may fail in certain cases. The
output from a failure usually looks something like this:
<p>
<center>
<table border align=center width="100%">
<tr>
<td>
<p><code><pre>
Assertion failed: H5.c:123: i&lt;NELMTS(H5_debug_g)
IOT Trap, core dumped.
</code></pre>
</td>
</tr>
</table>
</center>
<h2>Timings and Statistics</h2>
<p>Code to accumulate statistics is included at compile time by
using the <code>--enable-debug</code> configure switch. The
switch can be followed by an equal sign and a comma-separated
list of package names or else a default list is used.
<p>
<center>
<table border align=center width="80%">
<tr>
<th>Name</th>
<th>Default</th>
<th>Description</th>
</tr>
<tr>
<td align=center>a</td>
<td align=center>No</td>
<td>Attributes</td>
</tr>
<tr>
<td align=center>ac</td>
<td align=center>Yes</td>
<td>Meta data cache</td>
</tr>
<tr>
<td align=center>b</td>
<td align=center>Yes</td>
<td>B-Trees</td>
</tr>
<tr>
<td align=center>d</td>
<td align=center>Yes</td>
<td>Datasets</td>
</tr>
<tr>
<td align=center>e</td>
<td align=center>Yes</td>
<td>Error handling</td>
</tr>
<tr>
<td align=center>f</td>
<td align=center>Yes</td>
<td>Files</td>
</tr>
<tr>
<td align=center>g</td>
<td align=center>Yes</td>
<td>Groups</td>
</tr>
<tr>
<td align=center>hg</td>
<td align=center>Yes</td>
<td>Global heap</td>
</tr>
<tr>
<td align=center>hl</td>
<td align=center>No</td>
<td>Local heaps</td>
</tr>
<tr>
<td align=center>i</td>
<td align=center>Yes</td>
<td>Interface abstraction</td>
</tr>
<tr>
<td align=center>mf</td>
<td align=center>No</td>
<td>File memory management</td>
</tr>
<tr>
<td align=center>mm</td>
<td align=center>Yes</td>
<td>Library memory managment</td>
</tr>
<tr>
<td align=center>o</td>
<td align=center>No</td>
<td>Object headers and messages</td>
</tr>
<tr>
<td align=center>p</td>
<td align=center>Yes</td>
<td>Property lists</td>
</tr>
<tr>
<td align=center>s</td>
<td align=center>Yes</td>
<td>Data spaces</td>
</tr>
<tr>
<td align=center>t</td>
<td align=center>Yes</td>
<td>Datatypes</td>
</tr>
<tr>
<td align=center>v</td>
<td align=center>Yes</td>
<td>Vectors</td>
</tr>
<tr>
<td align=center>z</td>
<td align=center>Yes</td>
<td>Raw data filters</td>
</tr>
</table>
</center>
<p>In addition to including the code at compile time the
application must enable each package at runtime. This is done
by listing the package names in the <code>HDF5_DEBUG</code>
environment variable. That variable may also contain file
descriptor numbers (the default is `2') which control the output
for all following packages up to the next file number. The
word <code>all</code> refers to all packages. Any word my be
preceded by a minus sign to turn debugging off for the package.
<p>
<center>
<table border align=center width="100%">
<caption align=top><b>Sample debug specifications</b></caption>
<tr valign=top>
<td><code>all</code></td>
<td>This causes debugging output from all packages to be
sent to the standard error stream.</td>
</tr>
<tr valign=top>
<td><code>all -t -s</code></td>
<td>Debugging output for all packages except datatypes
and data spaces will appear on the standard error
stream.</td>
</tr>
<tr valign=top>
<td><code>-all ac 255 t,s</code></td>
<td>This disables all debugging even if the default was to
debug something, then output from the meta data cache is
send to the standard error stream and output from data
types and spaces is sent to file descriptor 255 which
should be redirected by the shell.</td>
</tr>
</table>
</center>
<p>The components of the <code>HDF5_DEBUG</code> value may be
separated by any non-lowercase letter.
<h2>API Tracing</h2>
<p>The HDF5 library can trace API calls by printing the
function name, the argument names and their values, and the
return value. Some people like to see lots of output during
program execution instead of using a good symbolic debugger, and
this feature is intended for their consumption. For example,
the output from <code>h5ls foo</code> after turning on tracing,
includes:
<p>
<center>
<table border align=center width="100%">
<tr>
<td>
<code><pre>
H5Tcopy(type=184549388) = 184549419 (type);
H5Tcopy(type=184549392) = 184549424 (type);
H5Tlock(type=184549424) = SUCCEED;
H5Tcopy(type=184549393) = 184549425 (type);
H5Tlock(type=184549425) = SUCCEED;
H5Fopen(filename="foo", flags=0, access=H5P_DEFAULT) = FAIL;
HDF5-DIAG: Error detected in thread 0. Back trace follows.
#000: H5F.c line 1245 in H5Fopen(): unable to open file
major(04): File interface
minor(10): Unable to open file
#001: H5F.c line 846 in H5F_open(): file does not exist
major(04): File interface
minor(10): Unable to open file
</pre></code>
</td>
</tr>
</table>
</center>
<p>The code that performs the tracing must be included in the
library by specifying the <code>--enable-trace</code>
configuration switch (the default for versions before 1.2). Then
the word <code>trace</code> must appear in the value of the
<code>HDF5_DEBUG</code> variable. The output will appear on the
last file descriptor before the word <code>trace</code> or two
(standard error) by default.
<p>
<center>
<table border align=center width="100%">
<tr>
<td>To display the trace on the standard error stream:
<code><pre>
$ env HDF5_DEBUG=trace a.out
</pre></code>
</td>
</tr>
<tr>
<td>To send the trace to a file:
<code><pre>
$ env HDF5_DEBUG="55 trace" a.out 55>trace-output
</pre></code>
</td>
</tr>
</table>
</center>
<h3>Performance</h3>
<p>If the library was not configured for tracing then there is no
unnecessary overhead since all tracing code is excluded.
However, if tracing is enabled but not used there is a small
penalty. First, code size is larger because of extra
statically-declared character strings used to store argument
types and names and extra auto variable pointer in each
function. Also, execution is slower because each function sets
and tests a local variable and each API function calls the
<code>H5_trace()</code> function.
<p>If tracing is enabled and turned on then the penalties from the
previous paragraph apply plus the time required to format each
line of tracing information. There is also an extra call to
H5_trace() for each API function to print the return value.
<h3>Safety</h3>
<p>The tracing mechanism is invoked for each API function before
arguments are checked for validity. If bad arguments are passed
to an API function it could result in a segmentation fault.
However, the tracing output is line-buffered so all previous
output will appear.
<h3>Completeness</h3>
<p>There are two API functions that don't participate in
tracing. They are <code>H5Eprint()</code> and
<code>H5Eprint_cb()</code> because their participation would
mess up output during automatic error reporting.
<p>On the other hand, a number of API functions are called during
library initialization and they print tracing information.
<h3>Implementation</h3>
<p>For those interested in the implementation here is a
description. Each API function should have a call to one of the
<code>H5TRACE()</code> macros immediately after the
<code>FUNC_ENTER()</code> macro. The first argument is the
return type encoded as a string. The second argument is the
types of all the function arguments encoded as a string. The
remaining arguments are the function arguments. This macro was
designed to be as terse and unobtrousive as possible.
<p>In order to keep the <code>H5TRACE()</code> calls synchronized
with the source code we've written a perl script which gets
called automatically just before Makefile dependencies are
calculated for the file. However, this only works when one is
using GNU make. To reinstrument the tracing explicitly, invoke
the <code>trace</code> program from the hdf5 bin directory with
the names of the source files that need to be updated. If any
file needs to be modified then a backup is created by appending
a tilde to the file name.
<p>
<center>
<table border align=center width="100%">
<caption align=top><b>Explicit Instrumentation</b></caption>
<tr>
<td>
<code><pre>
$ ../bin/trace *.c
H5E.c: in function `H5Ewalk_cb':
H5E.c:336: warning: trace info was not inserted
</pre></code>
</td>
</tr>
</table>
</center>
<p>Note: The warning message is the result of a comment of the
form <code>/*NO TRACE*/</code> somewhere in the function
body. Tracing information will not be updated or inserted if
such a comment exists.
<p>Error messages have the same format as a compiler so that they
can be parsed from program development environments like
Emacs. Any function which generates an error will not be
modified.
<hr>
<center>
<table border=0 width=98%>
<tr><td valign=top align=left>
<a href="H5.intro.html">Introduction to HDF5</a>&nbsp;<br>
<a href="RM_H5Front.html">HDF5 Reference Manual</a>&nbsp;<br>
<a href="index.html">Other HDF5 documents and links</a>&nbsp;<br>
<!--
<a href="Glossary.html">Glossary</a><br>
-->
</td>
<td valign=top align=right>
And in this document, the
<a href="H5.user.html"><strong>HDF5 User's Guide:</strong></a>&nbsp;&nbsp;&nbsp;&nbsp;
<br>
<a href="Files.html">Files</a>&nbsp;&nbsp;
<a href="Datasets.html">Datasets</a>&nbsp;&nbsp;
<a href="Datatypes.html">Datatypes</a>&nbsp;&nbsp;
<a href="Dataspaces.html">Dataspaces</a>&nbsp;&nbsp;
<a href="Groups.html">Groups</a>&nbsp;&nbsp;
<br>
<a href="References.html">References</a>&nbsp;&nbsp;
<a href="Attributes.html">Attributes</a>&nbsp;&nbsp;
<a href="Properties.html">Property Lists</a>&nbsp;&nbsp;
<a href="Errors.html">Error Handling</a>&nbsp;&nbsp;
<br>
<a href="Filters.html">Filters</a>&nbsp;&nbsp;
<a href="Caching.html">Caching</a>&nbsp;&nbsp;
<a href="Chunking.html">Chunking</a>&nbsp;&nbsp;
<a href="MountingFiles.html">Mounting Files</a>&nbsp;&nbsp;
<br>
<a href="Performance.html">Performance</a>&nbsp;&nbsp;
Debugging&nbsp;&nbsp;
<a href="Environment.html">Environment</a>&nbsp;&nbsp;
<a href="ddl.html">DDL</a>&nbsp;&nbsp;
</td></tr>
</table>
</center>
<hr>
<address>
<a href="mailto:hdfhelp@ncsa.uiuc.edu">HDF Help Desk</a>
</address>
<!-- Created: Wed Jun 17 12:29:12 EDT 1998 -->
<!-- hhmts start -->
Last modified: 13 December 1999
<!-- hhmts end -->
<br>
Describes HDF5 Release 1.5, Unreleased Development Branch
</body>
</html>