| blob | 74cf389894a8e00262dcf970c59cd1eac04ce3e5 |
3 <html>
4 <head>
8 </head>
9 <body>
12 CommandLine 2.0 Library Manual
13 </div>
15 <ol>
19 <ol>
23 set of possibilities</a></li>
28 </ol></li>
31 <ol>
33 <ul>
36 getPosition</a></li>
38 modifier</a></li>
39 </ul></li>
46 <ul>
48 output</a></li>
50 required and allowed</a></li>
52 specified</a></li>
56 </ul></li>
59 <ul>
65 function</a></li>
71 </ul></li>
74 <ul>
76 parser</a></li>
78 specialization</a></li>
80 specialization</a></li>
82 specialization</a></li>
84 specialization</a></li>
87 </ul></li>
88 </ol></li>
90 <ol>
94 options</a></li>
95 </ol></li>
96 </ol>
100 </div>
102 <!-- *********************************************************************** -->
105 </div>
106 <!-- *********************************************************************** -->
110 <p>This document describes the CommandLine argument processing library. It will
111 show you how to use it, and what it can do. The CommandLine library uses a
112 declarative approach to specifying the command line options that your program
113 takes. By default, these options declarations implicitly hold the value parsed
118 out there in many different languages, none of them fit well with what I needed.
119 By looking at the features and problems of other libraries, I designed the
120 CommandLine library to have the following features:</p>
122 <ol>
123 <li>Speed: The CommandLine library is very quick and uses little resources. The
124 parsing time of the library is directly proportional to the number of arguments
125 parsed, not the the number of options recognized. Additionally, command line
126 argument values are captured transparently into user defined global variables,
127 which can be accessed like any other variable (and with the same
128 performance).</li>
130 <li>Type Safe: As a user of CommandLine, you don't have to worry about
131 remembering the type of arguments that you want (is it an int? a string? a
132 bool? an enum?) and keep casting it around. Not only does this help prevent
133 error prone constructs, it also leads to dramatically cleaner source code.</li>
135 <li>No subclasses required: To use CommandLine, you instantiate variables that
136 correspond to the arguments that you would like to capture, you don't subclass a
138 code.</li>
140 <li>Globally accessible: Libraries can specify command line arguments that are
141 automatically enabled in any tool that links to the library. This is possible
142 because the application doesn't have to keep a list of arguments to pass to
146 <li>Cleaner: CommandLine supports enum and other types directly, meaning that
147 there is less error and more security built into the library. You don't have to
148 worry about whether your integral command line argument accidentally got
149 assigned a value that is not valid for your enum type.</li>
151 <li>Powerful: The CommandLine library supports many different types of
158 <li>Extensible: It is very simple to add a new argument type to CommandLine.
159 Simply specify the parser that you want to use with the command line option when
162 <li>Labor Saving: The CommandLine library cuts down on the amount of grunt work
163 that you, the user, have to do. For example, it automatically provides a
165 tool. Additionally, it does most of the basic correctness checking for
166 you.</li>
168 <li>Capable: The CommandLine library can handle lots of different forms of
169 options often found in real programs. For example, <a
176 </ol>
178 <p>This document will hopefully let you jump in and start using CommandLine in
179 your utility quickly and painlessly. Additionally it should be a simple
180 reference manual to figure out how stuff works. If it is failing in some area
181 (or you want an extension to the library), nag the author, <a
184 </div>
186 <!-- *********************************************************************** -->
189 </div>
190 <!-- *********************************************************************** -->
194 <p>This section of the manual runs through a simple CommandLine'ification of a
195 basic compiler tool. This is intended to show you how to jump into using the
196 CommandLine library in your own program, and show you some of the cool things it
197 can do.</p>
199 <p>To start out, you need to include the CommandLine header file into your
200 program:</p>
203 #include "llvm/Support/CommandLine.h"
204 </pre></div>
206 <p>Additionally, you need to add this as the first line of your main
207 program:</p>
210 int main(int argc, char **argv) {
212 ...
213 }
214 </pre></div>
216 <p>... which actually parses the arguments and fills in the variable
217 declarations.</p>
219 <p>Now that you are ready to support command line arguments, we need to tell the
220 system which ones we want, and what type of arguments they are. The CommandLine
221 library uses a declarative syntax to model command line arguments with the
222 global variable declarations that capture the parsed values. This means that
223 for every command line option that you would like to support, there should be a
224 global variable declaration to capture the result. For example, in a compiler,
226 to specify where to put the output. With the CommandLine library, this is
227 represented like this:</p>
231 <a href="#cl::opt">cl::opt</a><string> OutputFilename("<i>o</i>", <a href="#cl::desc">cl::desc</a>("<i>Specify output filename</i>"), <a href="#cl::value_desc">cl::value_desc</a>("<i>filename</i>"));
232 </pre></div>
235 capture the result of the "<tt>o</tt>" argument (first parameter). We specify
236 that this is a simple scalar option by using the "<tt><a
239 that the data type that we are parsing is a string.</p>
241 <p>The second and third parameters (which are optional) are used to specify what
243 looks like this:</p>
246 USAGE: compiler [options]
248 OPTIONS:
249 -help - display available options (--help-hidden for more)
250 <b>-o <filename> - Specify output filename</b>
251 </pre></div>
253 <p>Because we specified that the command line option should parse using the
254 <tt>string</tt> data type, the variable declared is automatically usable as a
255 real string in all contexts that a normal C++ string object may be used. For
256 example:</p>
259 ...
260 std::ofstream Output(OutputFilename.c_str());
261 if (Output.good()) ...
262 ...
263 </pre></div>
265 <p>There are many different options that you can use to customize the command
266 line option handling library, but the above example shows the general interface
267 to these options. The options can be specified in any order, and are specified
269 there are no positional dependencies to remember. The available options are
272 <p>Continuing the example, we would like to have our compiler take an input
273 filename as well as an output filename, but we do not want the input filename to
274 be specified with a hyphen (ie, not <tt>-filename.c</tt>). To support this
275 style of argument, the CommandLine library allows for <a
277 These positional arguments are filled with command line parameters that are not
278 in option form. We use this feature like this:</p>
281 <a href="#cl::opt">cl::opt</a><string> InputFilename(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<i><input file></i>"), <a href="#cl::init">cl::init</a>("<i>-</i>"));
282 </pre></div>
284 <p>This declaration indicates that the first positional argument should be
285 treated as the input filename. Here we use the <tt><a
287 command line option, which is used if the option is not specified (if you do not
289 the default constructor for the data type is used to initialize the value).
290 Command line options default to being optional, so if we would like to require
291 that the user always specify an input filename, we would add the <tt><a
296 <a href="#cl::opt">cl::opt</a><string> InputFilename(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<i><input file></i>"), <b><a href="#cl::Required">cl::Required</a></b>);
297 </pre></div>
299 <p>Again, the CommandLine library does not require the options to be specified
300 in any particular order, so the above declaration is equivalent to:</p>
303 <a href="#cl::opt">cl::opt</a><string> InputFilename(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::Required">cl::Required</a>, <a href="#cl::desc">cl::desc</a>("<i><input file></i>"));
304 </pre></div>
307 the CommandLine library will automatically issue an error if the argument is not
308 specified, which shifts all of the command line option verification code out of
309 your application into the library. This is just one example of how using flags
310 can alter the default behaviour of the library, on a per-option basis. By
311 adding one of the declarations above, the <tt>--help</tt> option synopsis is now
312 extended to:</p>
315 USAGE: compiler [options] <b><input file></b>
317 OPTIONS:
318 -help - display available options (--help-hidden for more)
319 -o <filename> - Specify output filename
320 </pre></div>
322 <p>... indicating that an input filename is expected.</p>
324 </div>
326 <!-- ======================================================================= -->
329 </div>
333 <p>In addition to input and output filenames, we would like the compiler example
336 compatibility with some of our users. We can support these by declaring options
337 of boolean type like this:</p>
340 <a href="#cl::opt">cl::opt</a><bool> Force ("<i>f</i>", <a href="#cl::desc">cl::desc</a>("<i>Overwrite output files</i>"));
341 <a href="#cl::opt">cl::opt</a><bool> Quiet ("<i>quiet</i>", <a href="#cl::desc">cl::desc</a>("<i>Don't print informational messages</i>"));
342 <a href="#cl::opt">cl::opt</a><bool> Quiet2("<i>q</i>", <a href="#cl::desc">cl::desc</a>("<i>Don't print informational messages</i>"), <a href="#cl::Hidden">cl::Hidden</a>);
343 </pre></div>
345 <p>This does what you would expect: it declares three boolean variables
353 for different data types. For example, in the string case, the argument passed
354 to the option is copied literally into the content of the string variable... we
355 obviously cannot do that in the boolean case, however, so we must use a smarter
356 parser. In the case of the boolean parser, it allows no options (in which case
357 it assigns the value of true to the variable), or it allows the values
359 following inputs:</p>
362 compiler -f # No value, 'Force' == true
363 compiler -f=true # Value specified, 'Force' == true
364 compiler -f=TRUE # Value specified, 'Force' == true
365 compiler -f=FALSE # Value specified, 'Force' == false
366 </pre></div>
369 the string values into boolean values, and rejects things like '<tt>compiler
372 like you would expect, using the '<tt>strtol</tt>' and '<tt>strtod</tt>' C
373 library calls to parse the string value into the specified data type.</p>
378 USAGE: compiler [options] <input file>
380 OPTIONS:
381 <b>-f - Overwrite output files</b>
382 -o - Override output filename
383 <b>-quiet - Don't print informational messages</b>
384 -help - display available options (--help-hidden for more)
385 </pre></div>
390 USAGE: compiler [options] <input file>
392 OPTIONS:
393 -f - Overwrite output files
394 -o - Override output filename
395 <b>-q - Don't print informational messages</b>
396 -quiet - Don't print informational messages
397 -help - display available options (--help-hidden for more)
398 </pre></div>
400 <p>This brief example has shown you how to use the '<tt><a
402 arguments. In addition to simple scalar arguments, the CommandLine library also
406 </div>
408 <!-- ======================================================================= -->
411 </div>
415 <p>So far, the example works well, except for the fact that we need to check the
416 quiet condition like this now:</p>
419 ...
420 if (!Quiet && !Quiet2) printInformationalMessage(...);
421 ...
422 </pre></div>
424 <p>... which is a real pain! Instead of defining two values for the same
425 condition, we can use the "<tt><a href="#cl::alias">cl::alias</a></tt>" class to make the "<tt>-q</tt>"
427 a value itself:</p>
430 <a href="#cl::opt">cl::opt</a><bool> Force ("<i>f</i>", <a href="#cl::desc">cl::desc</a>("<i>Overwrite output files</i>"));
431 <a href="#cl::opt">cl::opt</a><bool> Quiet ("<i>quiet</i>", <a href="#cl::desc">cl::desc</a>("<i>Don't print informational messages</i>"));
432 <a href="#cl::alias">cl::alias</a> QuietA("<i>q</i>", <a href="#cl::desc">cl::desc</a>("<i>Alias for -quiet</i>"), <a href="#cl::aliasopt">cl::aliasopt</a>(Quiet));
433 </pre></div>
435 <p>The third line (which is the only one we modified from above) defines a
438 specified. Because aliases do not hold state, the only thing the program has to
439 query is the <tt>Quiet</tt> variable now. Another nice feature of aliases is
440 that they automatically hide themselves from the <tt>-help</tt> output
441 (although, again, they are still visible in the <tt>--help-hidden
442 output</tt>).</p>
444 <p>Now the application code can simply use:</p>
447 ...
448 if (!Quiet) printInformationalMessage(...);
449 ...
450 </pre></div>
453 can be used to specify an alternative name for any variable type, and has many
454 uses.</p>
456 </div>
458 <!-- ======================================================================= -->
461 possibilities</a>
462 </div>
466 <p>So far we have seen how the CommandLine library handles builtin types like
467 <tt>std::string</tt>, <tt>bool</tt> and <tt>int</tt>, but how does it handle
468 things it doesn't know about, like enums or '<tt>int*</tt>'s?</p>
470 <p>The answer is that it uses a table-driven generic parser (unless you specify
472 Guide</a>). This parser maps literal strings to whatever type is required, and
473 requires you to tell it what this mapping should be.</p>
475 <p>Let's say that we would like to add four optimization levels to our
478 options like above, but there are several problems with this strategy:</p>
480 <ol>
481 <li>A user could specify more than one of the options at a time, for example,
483 catch this erroneous input for us.</li>
485 <li>We would have to test 4 different variables to see which ones are set.</li>
487 <li>This doesn't map to the numeric levels that we want... so we cannot easily
490 </ol>
492 <p>To cope with these problems, we can use an enum value, and have the
493 CommandLine library fill it in with the appropriate level directly, which is
494 used like this:</p>
497 enum OptLevel {
498 g, O1, O2, O3
499 };
501 <a href="#cl::opt">cl::opt</a><OptLevel> OptimizationLevel(<a href="#cl::desc">cl::desc</a>("<i>Choose optimization level:</i>"),
507 clEnumValEnd));
509 ...
510 if (OptimizationLevel >= O2) doPartialRedundancyElimination(...);
511 ...
512 </pre></div>
516 that are listed in the declaration (Note that the declaration list must be
518 library enforces
519 that the user can only specify one of the options, and it ensure that only valid
521 command line arguments matched the enum values. With this option added, our
522 help output now is:</p>
525 USAGE: compiler [options] <input file>
527 OPTIONS:
528 <b>Choose optimization level:
529 -g - No optimizations, enable debugging
530 -O1 - Enable trivial optimizations
531 -O2 - Enable default optimizations
532 -O3 - Enable expensive optimizations</b>
533 -f - Overwrite output files
534 -help - display available options (--help-hidden for more)
535 -o <filename> - Specify output filename
536 -quiet - Don't print informational messages
537 </pre></div>
539 <p>In this case, it is sort of awkward that flag names correspond directly to
541 in our program. Because of this, we can alternatively write this example like
542 this:</p>
545 enum OptLevel {
546 Debug, O1, O2, O3
547 };
549 <a href="#cl::opt">cl::opt</a><OptLevel> OptimizationLevel(<a href="#cl::desc">cl::desc</a>("<i>Choose optimization level:</i>"),
555 clEnumValEnd));
557 ...
558 if (OptimizationLevel == Debug) outputDebugInfo(...);
559 ...
560 </pre></div>
563 can directly specify the name that the flag should get. In general a direct
564 mapping is nice, but sometimes you can't or don't want to preserve the mapping,
565 which is when you would use it.</p>
567 </div>
569 <!-- ======================================================================= -->
572 </div>
576 <p>Another useful argument form is a named alternative style. We shall use this
577 style in our compiler to specify different debug levels that can be used.
578 Instead of each debug level being its own switch, we want to support the
579 following options, of which only one can be specified at a time:
582 our optimization level flags, but we also specify an option name. For this
583 case, the code looks like this:</p>
586 enum DebugLev {
587 nodebuginfo, quick, detailed
588 };
590 // Enable Debug Options to be specified on the command line
591 <a href="#cl::opt">cl::opt</a><DebugLev> DebugLevel("<i>debug_level</i>", <a href="#cl::desc">cl::desc</a>("<i>Set the debugging level:</i>"),
596 clEnumValEnd));
597 </pre></div>
601 is just the interface exposed to the user of your program and the help output by
605 USAGE: compiler [options] <input file>
607 OPTIONS:
608 Choose optimization level:
609 -g - No optimizations, enable debugging
610 -O1 - Enable trivial optimizations
611 -O2 - Enable default optimizations
612 -O3 - Enable expensive optimizations
613 <b>-debug_level - Set the debugging level:
614 =none - disable debug information
615 =quick - enable quick debug information
616 =detailed - enable detailed debug information</b>
617 -f - Overwrite output files
618 -help - display available options (--help-hidden for more)
619 -o <filename> - Specify output filename
620 -quiet - Don't print informational messages
621 </pre></div>
623 <p>Again, the only structural difference between the debug level declaration and
624 the optimization level declaration is that the debug level declaration includes
626 library processes the argument. The CommandLine library supports both forms so
627 that you can choose the form most appropriate for your application.</p>
629 </div>
631 <!-- ======================================================================= -->
634 </div>
638 <p>Now that we have the standard run-of-the-mill argument types out of the way,
639 lets get a little wild and crazy. Lets say that we want our optimizer to accept
640 a <b>list</b> of optimizations to perform, allowing duplicates. For example, we
642 this case, the order of the arguments and the number of appearances is very
644 template is for. First, start by defining an enum of the optimizations that you
645 would like to perform:</p>
648 enum Opts {
649 // 'inline' is a C++ keyword, so name it 'inlining'
650 dce, constprop, inlining, strip
651 };
652 </pre></div>
657 <a href="#cl::list">cl::list</a><Opts> OptimizationList(<a href="#cl::desc">cl::desc</a>("<i>Available Optimizations:</i>"),
663 clEnumValEnd));
664 </pre></div>
666 <p>This defines a variable that is conceptually of the type
668 vector methods:</p>
671 for (unsigned i = 0; i != OptimizationList.size(); ++i)
672 switch (OptimizationList[i])
673 ...
674 </pre></div>
676 <p>... to iterate through the list of options specified.</p>
679 completely general and may be used with any data types or other arguments that
681 especially useful way to use a list is to capture all of the positional
682 arguments together if there may be more than one specified. In the case of a
683 linker, for example, the linker takes several '<tt>.o</tt>' files, and needs to
684 capture them into a list. This is naturally specified as:</p>
687 ...
688 <a href="#cl::list">cl::list</a><std::string> InputFilenames(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<Input files>"), <a href="#cl::OneOrMore">cl::OneOrMore</a>);
689 ...
690 </pre></div>
693 such, accessing the list is simple, just like above. In this example, we used
695 CommandLine library that it is an error if the user does not specify any
696 <tt>.o</tt> files on our command line. Again, this just reduces the amount of
697 checking we have to do.</p>
699 </div>
701 <!-- ======================================================================= -->
704 </div>
708 <p>Instead of collecting sets of options in a list, it is also possible to
709 gather information for enum values in a <b>bit vector</b>. The represention used by
711 integer. An enum value is represented by a 0/1 in the enum's ordinal value bit
712 position. 1 indicating that the enum was specified, 0 otherwise. As each
713 specified value is parsed, the resulting enum's bit is set in the option's bit
714 vector:</p>
717 <i>bits</i> |= 1 << (unsigned)<i>enum</i>;
718 </pre></div>
720 <p>Options that are specified multiple times are redundant. Any instances after
721 the first are discarded.</p>
727 <a href="#cl::bits">cl::bits</a><Opts> OptimizationBits(<a href="#cl::desc">cl::desc</a>("<i>Available Optimizations:</i>"),
733 clEnumValEnd));
734 </pre></div>
736 <p>To test to see if <tt>constprop</tt> was specified, we can use the
737 <tt>cl:bits::isSet</tt> function:</p>
740 if (OptimizationBits.isSet(constprop)) {
741 ...
742 }
743 </pre></div>
745 <p>It's also possible to get the raw bit vector using the
746 <tt>cl::bits::getBits</tt> function:</p>
749 unsigned bits = OptimizationBits.getBits();
750 </pre></div>
752 <p>Finally, if external storage is used, then the location specified must be of
753 <b>type</b> <tt>unsigned</tt>. In all other ways a <a
757 </div>
760 <!-- ======================================================================= -->
763 </div>
767 <p>As our program grows and becomes more mature, we may decide to put summary
768 information about what it does into the help output. The help output is styled
769 to look similar to a Unix <tt>man</tt> page, providing concise information about
770 a program. Unix <tt>man</tt> pages, however often have a description about what
771 the program does. To add this to your CommandLine program, simply pass a third
772 argument to the <a
774 call in main. This additional argument is then printed as the overview
775 information for your program, allowing you to include any additional information
776 that you want. For example:</p>
779 int main(int argc, char **argv) {
780 <a href="#cl::ParseCommandLineOptions">cl::ParseCommandLineOptions</a>(argc, argv, " CommandLine compiler example\n\n"
782 ...
783 }
784 </pre></div>
786 <p>would yield the help output:</p>
789 <b>OVERVIEW: CommandLine compiler example
791 This program blah blah blah...</b>
793 USAGE: compiler [options] <input file>
795 OPTIONS:
796 ...
797 -help - display available options (--help-hidden for more)
798 -o <filename> - Specify output filename
799 </pre></div>
801 </div>
804 <!-- *********************************************************************** -->
807 </div>
808 <!-- *********************************************************************** -->
812 <p>Now that you know the basics of how to use the CommandLine library, this
813 section will give you the detailed information you need to tune how command line
815 processing capabilities.</p>
817 </div>
819 <!-- ======================================================================= -->
822 </div>
826 <p>Positional arguments are those arguments that are not named, and are not
827 specified with a hyphen. Positional arguments should be used when an option is
828 specified by its position alone. For example, the standard Unix <tt>grep</tt>
829 tool takes a regular expression argument, and an optional filename to search
830 through (which defaults to standard input if a filename is not specified).
831 Using the CommandLine library, this would be specified as:</p>
834 <a href="#cl::opt">cl::opt</a><string> Regex (<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<i><regular expression></i>"), <a href="#cl::Required">cl::Required</a>);
835 <a href="#cl::opt">cl::opt</a><string> Filename(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<i><input file></i>"), <a href="#cl::init">cl::init</a>("<i>-</i>"));
836 </pre></div>
838 <p>Given these two option declarations, the <tt>--help</tt> output for our grep
839 replacement would look like this:</p>
842 USAGE: spiffygrep [options] <b><regular expression> <input file></b>
844 OPTIONS:
845 -help - display available options (--help-hidden for more)
846 </pre></div>
848 <p>... and the resultant program could be used just like the standard
849 <tt>grep</tt> tool.</p>
851 <p>Positional arguments are sorted by their order of construction. This means
852 that command line options will be ordered according to how they are listed in a
853 .cpp file, but will not have an ordering defined if the positional arguments
854 are defined in multiple .cpp files. The fix for this problem is simply to
855 define all of your positional arguments in one .cpp file.</p>
857 </div>
860 <!-- _______________________________________________________________________ -->
863 </div>
867 <p>Sometimes you may want to specify a value to your positional argument that
868 starts with a hyphen (for example, searching for '<tt>-foo</tt>' in a file). At
869 first, you will have trouble doing this, because it will try to find an argument
870 named '<tt>-foo</tt>', and will fail (and single quotes will not save you).
871 Note that the system <tt>grep</tt> has the same problem:</p>
874 $ spiffygrep '-foo' test.txt
875 Unknown command line argument '-foo'. Try: spiffygrep --help'
877 $ grep '-foo' test.txt
878 grep: illegal option -- f
879 grep: illegal option -- o
880 grep: illegal option -- o
881 Usage: grep -hblcnsviw pattern file . . .
882 </pre></div>
884 <p>The solution for this problem is the same for both your tool and the system
885 version: use the '<tt>--</tt>' marker. When the user specifies '<tt>--</tt>' on
886 the command line, it is telling the program that all options after the
887 '<tt>--</tt>' should be treated as positional arguments, not options. Thus, we
888 can use it like this:</p>
891 $ spiffygrep -- -foo test.txt
892 ...output...
893 </pre></div>
895 </div>
897 <!-- _______________________________________________________________________ -->
900 </div>
902 <p>Sometimes an option can affect or modify the meaning of another option. For
903 example, consider <tt>gcc</tt>'s <tt>-x LANG</tt> option. This tells
904 <tt>gcc</tt> to ignore the suffix of subsequent positional arguments and force
905 the file to be interpreted as if it contained source code in language
906 <tt>LANG</tt>. In order to handle this properly, you need to know the
907 absolute position of each argument, especially those in lists, so their
908 interaction(s) can be applied correctly. This is also useful for options like
909 <tt>-llibname</tt> which is actually a positional argument that starts with
910 a dash.</p>
911 <p>So, generally, the problem is that you have two <tt>cl::list</tt> variables
912 that interact in some way. To ensure the correct interaction, you can use the
913 <tt>cl::list::getPosition(optnum)</tt> method. This method returns the
914 absolute position (as found on the command line) of the <tt>optnum</tt>
915 item in the <tt>cl::list</tt>.</p>
916 <p>The idiom for usage is like this:</p>
919 static cl::list<std::string> Files(cl::Positional, cl::OneOrMore);
922 int main(int argc, char**argv) {
923 // ...
924 std::vector<std::string>::iterator fileIt = Files.begin();
925 std::vector<std::string>::iterator libIt = Libraries.begin();
926 unsigned libPos = 0, filePos = 0;
927 while ( 1 ) {
928 if ( libIt != Libraries.end() )
929 libPos = Libraries.getPosition( libIt - Libraries.begin() );
930 else
931 libPos = 0;
932 if ( fileIt != Files.end() )
933 filePos = Files.getPosition( fileIt - Files.begin() );
934 else
935 filePos = 0;
937 if ( filePos != 0 && (libPos == 0 || filePos < libPos) ) {
938 // Source File Is next
939 ++fileIt;
940 }
941 else if ( libPos != 0 && (filePos == 0 || libPos < filePos) ) {
942 // Library is next
943 ++libIt;
944 }
945 else
946 break; // we're done with the list
947 }
948 }</pre></div>
950 <p>Note that, for compatibility reasons, the <tt>cl::opt</tt> also supports an
951 <tt>unsigned getPosition()</tt> option that will provide the absolute position
952 of that option. You can apply the same approach as above with a
953 <tt>cl::opt</tt> and a <tt>cl::list</tt> option as you can with two lists.</p>
954 </div>
956 <!-- _______________________________________________________________________ -->
959 </div>
965 this style of option processing, all arguments specified after the last
966 positional argument are treated as special interpreter arguments that are not
967 interpreted by the command line argument.</p>
969 <p>As a concrete example, lets say we are developing a replacement for the
970 standard Unix Bourne shell (<tt>/bin/sh</tt>). To run <tt>/bin/sh</tt>, first
971 you specify options to the shell itself (like <tt>-x</tt> which turns on trace
972 output), then you specify the name of the script to run, then you specify
973 arguments to the script. These arguments to the script are parsed by the Bourne
974 shell command line option processor, but are not interpreted as options to the
975 shell itself. Using the CommandLine library, we would specify this as:</p>
978 <a href="#cl::opt">cl::opt</a><string> Script(<a href="#cl::Positional">cl::Positional</a>, <a href="#cl::desc">cl::desc</a>("<i><input script></i>"), <a href="#cl::init">cl::init</a>("-"));
979 <a href="#cl::list">cl::list</a><string> Argv(<a href="#cl::ConsumeAfter">cl::ConsumeAfter</a>, <a href="#cl::desc">cl::desc</a>("<i><program arguments>...</i>"));
980 <a href="#cl::opt">cl::opt</a><bool> Trace("<i>x</i>", <a href="#cl::desc">cl::desc</a>("<i>Enable trace output</i>"));
981 </pre></div>
983 <p>which automatically provides the help output:</p>
986 USAGE: spiffysh [options] <b><input script> <program arguments>...</b>
988 OPTIONS:
989 -help - display available options (--help-hidden for more)
990 <b>-x - Enable trace output</b>
991 </pre></div>
993 <p>At runtime, if we run our new shell replacement as `<tt>spiffysh -x test.sh
994 -a -x -y bar</tt>', the <tt>Trace</tt> variable will be set to true, the
997 were specified after the last positional argument (which is the script
998 name).</p>
1000 <p>There are several limitations to when <tt>cl::ConsumeAfter</tt> options can
1001 be specified. For example, only one <tt>cl::ConsumeAfter</tt> can be specified
1004 positional arguments, and the <tt>cl::ConsumeAfter</tt> option should be a <a
1007 </div>
1009 <!-- ======================================================================= -->
1012 </div>
1016 <p>By default, all command line options automatically hold the value that they
1017 parse from the command line. This is very convenient in the common case,
1018 especially when combined with the ability to define command line options in the
1019 files that use them. This is called the internal storage model.</p>
1021 <p>Sometimes, however, it is nice to separate the command line option processing
1022 code from the storage of the value parsed. For example, lets say that we have a
1023 '<tt>-debug</tt>' option that we would like to use to enable debug information
1024 across the entire body of our program. In this case, the boolean value
1025 controlling the debug code should be globally accessable (in a header file, for
1026 example) yet the command line option processing code should not be exposed to
1027 all of these clients (requiring lots of .cpp files to #include
1028 <tt>CommandLine.h</tt>).</p>
1030 <p>To do this, set up your .h file with your option, like this for example:</p>
1033 <pre>
1034 <i>// DebugFlag.h - Get access to the '-debug' command line option
1035 //
1037 // DebugFlag - This boolean is set to true if the '-debug' command line option
1038 // is specified. This should probably not be referenced directly, instead, use
1039 // the DEBUG macro below.
1040 //</i>
1041 extern bool DebugFlag;
1043 <i>// DEBUG macro - This macro should be used by code to emit debug information.
1044 // In the '-debug' option is specified on the command line, and if this is a
1045 // debug build, then the code specified as the option to the macro will be
1046 // executed. Otherwise it will not be.</i>
1048 #define DEBUG(X)
1049 #else
1050 #define DEBUG(X)</span> do { if (DebugFlag) { X; } } while (0)
1052 </pre>
1053 </div>
1055 <p>This allows clients to blissfully use the <tt>DEBUG()</tt> macro, or the
1056 <tt>DebugFlag</tt> explicitly if they want to. Now we just need to be able to
1057 set the <tt>DebugFlag</tt> boolean when the option is set. To do this, we pass
1058 an additional argument to our command line argument processor, and we specify
1060 attribute:</p>
1063 <pre>
1064 bool DebugFlag; <i>// the actual value</i>
1066 Debug("<i>debug</i>", <a href="#cl::desc">cl::desc</a>("<i>Enable debug output</i>"), <a href="#cl::Hidden">cl::Hidden</a>, <a href="#cl::location">cl::location</a>(DebugFlag));
1067 </pre>
1068 </div>
1072 template should not maintain a copy of the value itself. In addition to this,
1074 that <tt>DebugFlag</tt> is automatically set.</p>
1076 </div>
1078 <!-- ======================================================================= -->
1081 </div>
1085 <p>This section describes the basic attributes that you can specify on
1086 options.</p>
1088 <ul>
1090 <li>The option name attribute (which is required for all options, except <a
1092 This option is specified in simple double quotes:
1094 <pre>
1096 </pre>
1098 </li>
1101 description for the option to be shown in the <tt>--help</tt> output for the
1102 program.</li>
1105 specifies a string that can be used to fine tune the <tt>--help</tt> output for
1107 example.</li>
1111 not specified then the command line option value defaults to the value created
1112 by the default constructor for the type. <b>Warning</b>: If you specify both
1113 <b><tt>cl::init</tt></b> and <b><tt>cl::location</tt></b> for an option,
1114 you must specify <b><tt>cl::location</tt></b> first, so that when the
1115 command-line parser sees <b><tt>cl::init</tt></b>, it knows where to put the
1116 initial value. (You will get an error at runtime if you don't put them in
1117 the right order.)</li>
1120 store the value for a parsed command line option if using external storage. See
1122 information.</li>
1126 an alias for.</li>
1129 the string-to-value mapping to be used by the generic parser. It takes a
1130 <b>clEnumValEnd terminated</b> list of (option, value, description) triplets
1131 that
1132 specify the option name, the value mapped to, and the description shown in the
1133 <tt>--help</tt> for the tool. Because the generic parser is used most
1134 frequently with enum values, two macros are often useful:
1136 <ol>
1139 nice simple way to specify a triplet for an enum. This macro automatically
1140 makes the option name be the same as the enum name. The first option to the
1141 macro is the enum, the second is the description for the command line
1142 option.</li>
1145 specify macro options where the option name doesn't equal the enum name. For
1146 this macro, the first argument is the enum value, the second is the flag name,
1147 and the second is the description.</li>
1150 attribute specifies that this option takes has multiple values
1151 (example: <tt>-sectalign segname sectname sectvalue</tt>). This
1152 attribute takes one unsigned argument - the number of values for the
1153 option. This attribute is valid only on <tt>cl::list</tt> options (and
1154 will fail with compile error if you try to use it with other option
1155 types). It is allowed to use all of the usual modifiers on
1156 multi-valued options (besides <tt>cl::ValueDisallowed</tt>,
1157 obviously).</li>
1160 </ol>
1162 You will get a compile time error if you try to use cl::values with a parser
1163 that does not support it.</li>
1165 </ul>
1167 </div>
1169 <!-- ======================================================================= -->
1172 </div>
1176 <p>Option modifiers are the flags and expressions that you pass into the
1179 tweak how options are parsed and how <tt>--help</tt> output is generated to fit
1180 your application well.</p>
1182 <p>These options fall into five main catagories:</p>
1184 <ol>
1187 required and allowed</a></li>
1189 specified</a></li>
1192 </ol>
1194 <p>It is not possible to specify two options from the same catagory (you'll get
1195 a runtime error) to a single option, except for options in the miscellaneous
1196 catagory. The CommandLine library specifies defaults for all of these settings
1197 that are the most useful in practice and the most common, which mean that you
1198 usually shouldn't have to worry about these.</p>
1200 </div>
1202 <!-- _______________________________________________________________________ -->
1205 </div>
1209 <p>The <tt>cl::NotHidden</tt>, <tt>cl::Hidden</tt>, and
1210 <tt>cl::ReallyHidden</tt> modifiers are used to control whether or not an option
1211 appears in the <tt>--help</tt> and <tt>--help-hidden</tt> output for the
1212 compiled program:</p>
1214 <ul>
1219 in both help listings.</li>
1223 the option should not appear in the <tt>--help</tt> output, but should appear in
1224 the <tt>--help-hidden</tt> output.</li>
1227 indicates that the option should not appear in any help output.</li>
1229 </ul>
1231 </div>
1233 <!-- _______________________________________________________________________ -->
1236 allowed</a>
1237 </div>
1241 <p>This group of options is used to control how many time an option is allowed
1242 (or required) to be specified on the command line of your program. Specifying a
1243 value for this setting allows the CommandLine library to do error checking for
1244 you.</p>
1246 <p>The allowed values for this option group are:</p>
1248 <ul>
1253 allow either zero or one occurrence of the option to be specified.</li>
1257 indicates that your program will allow the option to be specified zero or more
1258 times.</li>
1261 indicates that the specified option must be specified exactly one time.</li>
1264 indicates that the option must be specified at least one time.</li>
1266 <li>The <b><tt>cl::ConsumeAfter</tt></b> modifier is described in the <a
1269 </ul>
1271 <p>If an option is not specified, then the value of the option is equal to the
1274 option value is initialized with the default constructor for the data type.</p>
1276 <p>If an option is specified multiple times for an option of the <tt><a
1278 retained.</p>
1280 </div>
1282 <!-- _______________________________________________________________________ -->
1285 </div>
1289 <p>This group of options is used to control whether or not the option allows a
1290 value to be present. In the case of the CommandLine library, a value is either
1291 specified with an equal sign (e.g. '<tt>-index-depth=17</tt>') or as a trailing
1292 string (e.g. '<tt>-o a.out</tt>').</p>
1294 <p>The allowed values for this option group are:</p>
1296 <ul>
1299 (which is the default for <tt>bool</tt> typed options) specifies that it is
1300 acceptable to have a value, or not. A boolean argument can be enabled just by
1301 appearing on the command line, or it can have an explicit '<tt>-foo=true</tt>'.
1302 If an option is specified with this mode, it is illegal for the value to be
1303 provided without the equal sign. Therefore '<tt>-foo true</tt>' is illegal. To
1304 get this behavior, you must use the <a
1308 (which is the default for all other types except for <a
1310 specifies that a value must be provided. This mode informs the command line
1311 library that if an option is not provides with an equal sign, that the next
1312 argument provided must be the value. This allows things like '<tt>-o
1313 a.out</tt>' to work.</li>
1317 alternatives using the generic parser</a>) indicates that it is a runtime error
1318 for the user to specify a value. This can be provided to disallow users from
1319 providing options to boolean options (like '<tt>-foo=true</tt>').</li>
1321 </ul>
1323 <p>In general, the default values for this option group work just like you would
1324 want them to. As mentioned above, you can specify the <a
1326 argument to restrict your command line parser. These options are mostly useful
1329 </div>
1331 <!-- _______________________________________________________________________ -->
1334 </div>
1338 <p>The formatting option group is used to specify that the command line option
1339 has special abilities and is otherwise different from other command line
1340 arguments. As usual, you can only specify one of these arguments at most.</p>
1342 <ul>
1345 modifier (which is the default all options) specifies that this option is
1349 specifies that this is a positional argument that does not have a command line
1351 Arguments</a> section for more information.</li>
1354 specifies that this option is used to capture "interpreter style" arguments. See <a href="#cl::ConsumeAfter">this section for more information</a>.</li>
1357 that this option prefixes its value. With 'Prefix' options, the equal sign does
1358 not separate the value from the option name specified. Instead, the value is
1359 everything after the prefix, including any equal sign if present. This is useful
1360 for processing odd arguments like <tt>-lmalloc</tt> and <tt>-L/usr/lib</tt> in a
1361 linker tool or <tt>-DNAME=value</tt> in a compiler tool. Here, the
1362 '<tt>l</tt>', '<tt>D</tt>' and '<tt>L</tt>' options are normal string (or list)
1364 modifier added to allow the CommandLine library to recognize them. Note that
1367 specified.</li>
1370 to implement Unix-style tools (like <tt>ls</tt>) that have lots of single letter
1371 arguments, but only require a single dash. For example, the '<tt>ls -labF</tt>'
1372 command actually enables four different options, all of which are single
1374 options cannot have values.</li>
1376 </ul>
1378 <p>The CommandLine library does not restrict how you use the <b><tt><a
1381 specify ambiguous argument settings. Thus, it is possible to have multiple
1382 letter options that are prefix or grouping options, and they will still work as
1383 designed.</p>
1385 <p>To do this, the CommandLine library uses a greedy algorithm to parse the
1386 input option into (potentially multiple) prefix and grouping options. The
1387 strategy basically looks like this:</p>
1391 <ol>
1392 <li><tt>string input = OrigInput;</tt>
1393 <li><tt>if (isOption(input)) return getOption(input).parse();</tt> <i>// Normal option</i>
1394 <li><tt>while (!isOption(input) && !input.empty()) input.pop_back();</tt> <i>// Remove the last letter</i>
1395 <li><tt>if (input.empty()) return error();</tt> <i>// No matching option</i>
1396 <li><tt>if (getOption(input).isPrefix())<br>
1397 return getOption(input).parse(input);</tt>
1398 <li><tt>while (!input.empty()) { <i>// Must be grouping options</i><br>
1399 getOption(input).parse();<br>
1400 OrigInput.erase(OrigInput.begin(), OrigInput.begin()+input.length());<br>
1401 input = OrigInput;<br>
1402 while (!isOption(input) && !input.empty()) input.pop_back();<br>
1403 }</tt>
1404 <li><tt>if (!OrigInput.empty()) error();</tt></li>
1405 </ol>
1407 <p><tt>}</tt></p>
1408 </div>
1410 </div>
1412 <!-- _______________________________________________________________________ -->
1415 </div>
1419 <p>The miscellaneous option modifiers are the only flags where you can specify
1420 more than one flag from the set: they are not mutually exclusive. These flags
1421 specify boolean properties that modify the option.</p>
1423 <ul>
1426 indicates that any commas specified for an option's value should be used to
1427 split the value up into multiple values for the option. For example, these two
1428 options are equivalent when <tt>cl::CommaSeparated</tt> is specified:
1430 makes sense to be used in a case where the option is allowed to accept one or
1434 <b><tt>cl::PositionalEatsArgs</tt></b></a> modifier (which only applies to
1435 positional arguments, and only makes sense for lists) indicates that positional
1436 argument should consume any strings after it (including strings that start with
1444 used to handle unknown options. If there is at least one option with
1445 <b><tt>cl::Sink</tt></b></a> modifier specified, the parser passes
1446 unrecognized option strings to it as values instead of signaling an
1447 error. As with <b><tt>cl::CommaSeparated</tt></b></a>, this modifier
1450 </ul>
1452 <p>So far, these are the only three miscellaneous option modifiers.</p>
1454 </div>
1456 <!-- _______________________________________________________________________ -->
1459 </div>
1463 <p>Some systems, such as certain variants of Microsoft Windows and
1464 some older Unices have a relatively low limit on command-line
1465 length. It is therefore customary to use the so-called 'response
1466 files' to circumvent this restriction. These files are mentioned on
1468 files and inserts the contents into argv, thereby working around the
1469 command-line length limits. Response files are enabled by an optional
1470 fourth argument to
1472 and
1474 </p>
1476 </div>
1479 <!-- ======================================================================= -->
1482 </div>
1486 <p>Despite all of the built-in flexibility, the CommandLine option library
1487 really only consists of one function (<a
1492 classes in detail.</p>
1494 </div>
1496 <!-- _______________________________________________________________________ -->
1499 function</a>
1500 </div>
1504 <p>The <tt>cl::ParseCommandLineOptions</tt> function is designed to be called
1505 directly from <tt>main</tt>, and is used to fill in the values of all of the
1506 command line option variables once <tt>argc</tt> and <tt>argv</tt> are
1507 available.</p>
1509 <p>The <tt>cl::ParseCommandLineOptions</tt> function requires two parameters
1510 (<tt>argc</tt> and <tt>argv</tt>), but may also take an optional third parameter
1512 <tt>--help</tt> option is invoked, and a fourth boolean parameter that enables
1515 </div>
1517 <!-- _______________________________________________________________________ -->
1520 function</a>
1521 </div>
1525 <p>The <tt>cl::ParseEnvironmentOptions</tt> function has mostly the same effects
1526 as <a
1528 except that it is designed to take values for options from an environment
1529 variable, for those cases in which reading the command line is not convenient or
1530 desired. It fills in the values of all the command line option variables just
1531 like <a
1533 does.</p>
1535 <p>It takes four parameters: the name of the program (since <tt>argv</tt> may
1536 not be available, it can't just look in <tt>argv[0]</tt>), the name of the
1537 environment variable to examine, the optional
1539 <tt>--help</tt> option is invoked, and the boolean
1541 should be read.</p>
1543 <p><tt>cl::ParseEnvironmentOptions</tt> will break the environment
1544 variable's value up into words and then process them using
1546 <b>Note:</b> Currently <tt>cl::ParseEnvironmentOptions</tt> does not support
1549 which is different from what you would get from the shell with the same
1550 input.</p>
1552 </div>
1554 <!-- _______________________________________________________________________ -->
1557 function</a>
1558 </div>
1562 <p>The <tt>cl::SetVersionPrinter</tt> function is designed to be called
1563 directly from <tt>main</tt> and <i>before</i>
1564 <tt>cl::ParseCommandLineOptions</tt>. Its use is optional. It simply arranges
1565 for a function to be called in response to the <tt>--version</tt> option instead
1566 of having the <tt>CommandLine</tt> library print out the usual version string
1567 for LLVM. This is useful for programs that are not part of LLVM but wish to use
1568 the <tt>CommandLine</tt> facilities. Such programs should just define a small
1569 function that takes no arguments and returns <tt>void</tt> and that prints out
1570 whatever version information is appropriate for the program. Pass the address
1571 of that function to <tt>cl::SetVersionPrinter</tt> to arrange for it to be
1572 called when the <tt>--version</tt> option is given by the user.</p>
1574 </div>
1575 <!-- _______________________________________________________________________ -->
1578 </div>
1582 <p>The <tt>cl::opt</tt> class is the class used to represent scalar command line
1583 options, and is the one used most of the time. It is a templated class which
1584 can take up to three arguments (all except for the first have default values
1585 though):</p>
1588 <b>namespace</b> cl {
1589 <b>template</b> <<b>class</b> DataType, <b>bool</b> ExternalStorage = <b>false</b>,
1590 <b>class</b> ParserClass = parser<DataType> >
1591 <b>class</b> opt;
1592 }
1593 </pre></div>
1595 <p>The first template argument specifies what underlying data type the command
1596 line argument is, and is used to select a default parser implementation. The
1597 second template argument is used to specify whether the option should contain
1598 the storage for the option (the default) or whether external storage should be
1600 vs External Storage</a> for more information).</p>
1602 <p>The third template argument specifies which parser to use. The default value
1603 selects an instantiation of the <tt>parser</tt> class based on the underlying
1604 data type of the option. In general, this default works well for most
1605 applications, so this option is only used when using a <a
1608 </div>
1610 <!-- _______________________________________________________________________ -->
1613 </div>
1617 <p>The <tt>cl::list</tt> class is the class used to represent a list of command
1618 line options. It too is a templated class which can take up to three
1619 arguments:</p>
1622 <b>namespace</b> cl {
1623 <b>template</b> <<b>class</b> DataType, <b>class</b> Storage = <b>bool</b>,
1624 <b>class</b> ParserClass = parser<DataType> >
1625 <b>class</b> list;
1626 }
1627 </pre></div>
1629 <p>This class works the exact same as the <a
1631 the <b>type</b> of the external storage, not a boolean value. For this class,
1632 the marker type '<tt>bool</tt>' is used to indicate that internal storage should
1633 be used.</p>
1635 </div>
1637 <!-- _______________________________________________________________________ -->
1640 </div>
1644 <p>The <tt>cl::bits</tt> class is the class used to represent a list of command
1645 line options in the form of a bit vector. It is also a templated class which
1646 can take up to three arguments:</p>
1649 <b>namespace</b> cl {
1650 <b>template</b> <<b>class</b> DataType, <b>class</b> Storage = <b>bool</b>,
1651 <b>class</b> ParserClass = parser<DataType> >
1652 <b>class</b> bits;
1653 }
1654 </pre></div>
1656 <p>This class works the exact same as the <a
1658 must be of <b>type</b> <tt>unsigned</tt> if external storage is used.</p>
1660 </div>
1662 <!-- _______________________________________________________________________ -->
1665 </div>
1669 <p>The <tt>cl::alias</tt> class is a nontemplated class that is used to form
1670 aliases for other arguments.</p>
1673 <b>namespace</b> cl {
1674 <b>class</b> alias;
1675 }
1676 </pre></div>
1679 used to specify which option this is an alias for. Alias arguments default to
1681 the conversion from string to data.</p>
1683 </div>
1685 <!-- _______________________________________________________________________ -->
1688 </div>
1692 <p>The <tt>cl::extrahelp</tt> class is a nontemplated class that allows extra
1693 help text to be printed out for the <tt>--help</tt> option.</p>
1696 <b>namespace</b> cl {
1697 <b>struct</b> extrahelp;
1698 }
1699 </pre></div>
1701 <p>To use the extrahelp, simply construct one with a <tt>const char*</tt>
1702 parameter to the constructor. The text passed to the constructor will be printed
1703 at the bottom of the help message, verbatim. Note that multiple
1704 <tt>cl::extrahelp</tt> <b>can</b> be used, but this practice is discouraged. If
1705 your tool needs to print additional help information, put all that help into a
1706 single <tt>cl::extrahelp</tt> instance.</p>
1707 <p>For example:</p>
1710 </pre></div>
1711 </div>
1713 <!-- ======================================================================= -->
1716 </div>
1720 <p>Parsers control how the string value taken from the command line is
1721 translated into a typed value, suitable for use in a C++ program. By default,
1722 the CommandLine library uses an instance of <tt>parser<type></tt> if the
1723 command line option specifies that it uses values of type '<tt>type</tt>'.
1724 Because of this, custom option processing is specified with specializations of
1725 the '<tt>parser</tt>' class.</p>
1727 <p>The CommandLine library provides the following builtin parser
1728 specializations, which are sufficient for most applications. It can, however,
1729 also be extended to work with new data types and new ways of interpreting the
1731 details on this type of library extension.</p>
1733 <ul>
1736 can be used to map strings values to any data type, through the use of the <a
1738 information. The most common use of this parser is for parsing enum values,
1739 which allows you to use the CommandLine library for all of the error checking to
1740 make sure that only valid enum values are specified (as opposed to accepting
1741 arbitrary strings). Despite this, however, the generic parser class can be used
1742 for any data type.</li>
1745 is used to convert boolean strings to a boolean value. Currently accepted
1750 specialization</b></a> is used for cases where the value is boolean,
1751 but we also need to know whether the option was specified at all. boolOrDefault
1752 is an enum with 3 values, BOU_UNSET, BOU_TRUE and BOU_FALSE. This parser accepts
1753 the same strings as <b><tt>parser<bool></tt></b>.</li>
1756 specialization</b></a> simply stores the parsed string into the string value
1757 specified. No conversion or modification of the data is performed.</li>
1760 uses the C <tt>strtol</tt> function to parse the string input. As such, it will
1761 accept a decimal number (with an optional '+' or '-' prefix) which must start
1762 with a non-zero digit. It accepts octal numbers, which are identified with a
1763 '<tt>0</tt>' prefix digit, and hexadecimal numbers with a prefix of
1764 '<tt>0x</tt>' or '<tt>0X</tt>'.</li>
1767 <b><tt>parser<float></tt> specializations</b> use the standard C
1768 <tt>strtod</tt> function to convert floating point strings into floating point
1769 values. As such, a broad range of string formats is supported, including
1770 exponential notation (ex: <tt>1.7e15</tt>) and properly supports locales.
1771 </li>
1773 </ul>
1775 </div>
1777 <!-- *********************************************************************** -->
1780 </div>
1781 <!-- *********************************************************************** -->
1785 <p>Although the CommandLine library has a lot of functionality built into it
1786 already (as discussed previously), one of its true strengths lie in its
1787 extensibility. This section discusses how the CommandLine library works under
1788 the covers and illustrates how to do some simple, common, extensions.</p>
1790 </div>
1792 <!-- ======================================================================= -->
1795 </div>
1799 <p>One of the simplest and most common extensions is the use of a custom parser.
1801 of the CommandLine library that turns string input from the user into a
1802 particular parsed data type, validating the input in the process.</p>
1804 <p>There are two ways to use a new parser:</p>
1806 <ol>
1808 <li>
1811 your custom data type.<p>
1813 <p>This approach has the advantage that users of your custom data type will
1814 automatically use your custom parser whenever they define an option with a value
1815 type of your data type. The disadvantage of this approach is that it doesn't
1816 work if your fundamental data type is something that is already supported.</p>
1818 </li>
1820 <li>
1822 <p>Write an independent class, using it explicitly from options that need
1823 it.</p>
1825 <p>This approach works well in situations where you would line to parse an
1826 option using special syntax for a not-very-special data-type. The drawback of
1827 this approach is that users of your parser have to be aware that they are using
1828 your parser instead of the builtin ones.</p>
1830 </li>
1832 </ol>
1834 <p>To guide the discussion, we will discuss a custom parser that accepts file
1835 sizes, specified with an optional unit after the numeric size. For example, we
1837 this case, the underlying data type we want to parse into is
1838 '<tt>unsigned</tt>'. We choose approach #2 above because we don't want to make
1839 this the default for all <tt>unsigned</tt> options.</p>
1841 <p>To start out, we declare our new <tt>FileSizeParser</tt> class:</p>
1844 <b>struct</b> FileSizeParser : <b>public</b> cl::basic_parser<<b>unsigned</b>> {
1845 <i>// parse - Return true on error.</i>
1846 <b>bool</b> parse(cl::Option &O, <b>const char</b> *ArgName, <b>const</b> std::string &ArgValue,
1847 <b>unsigned</b> &Val);
1848 };
1849 </pre></div>
1851 <p>Our new class inherits from the <tt>cl::basic_parser</tt> template class to
1852 fill in the default, boiler plate code for us. We give it the data type that
1853 we parse into, the last argument to the <tt>parse</tt> method, so that clients of
1854 our custom parser know what object type to pass in to the parse method. (Here we
1855 declare that we parse into '<tt>unsigned</tt>' variables.)</p>
1857 <p>For most purposes, the only method that must be implemented in a custom
1858 parser is the <tt>parse</tt> method. The <tt>parse</tt> method is called
1859 whenever the option is invoked, passing in the option itself, the option name,
1860 the string to parse, and a reference to a return value. If the string to parse
1861 is not well-formed, the parser should output an error message and return true.
1862 Otherwise it should return false and set '<tt>Val</tt>' to the parsed value. In
1863 our example, we implement <tt>parse</tt> as:</p>
1866 <b>bool</b> FileSizeParser::parse(cl::Option &O, <b>const char</b> *ArgName,
1867 <b>const</b> std::string &Arg, <b>unsigned</b> &Val) {
1868 <b>const char</b> *ArgStart = Arg.c_str();
1869 <b>char</b> *End;
1871 <i>// Parse integer part, leaving 'End' pointing to the first non-integer char</i>
1872 Val = (unsigned)strtol(ArgStart, &End, 0);
1874 <b>while</b> (1) {
1875 <b>switch</b> (*End++) {
1876 <b>case</b> 0: <b>return</b> false; <i>// No error</i>
1877 <b>case</b> 'i': <i>// Ignore the 'i' in KiB if people use that</i>
1878 <b>case</b> 'b': <b>case</b> 'B': <i>// Ignore B suffix</i>
1879 <b>break</b>;
1881 <b>case</b> 'g': <b>case</b> 'G': Val *= 1024*1024*1024; <b>break</b>;
1882 <b>case</b> 'm': <b>case</b> 'M': Val *= 1024*1024; <b>break</b>;
1883 <b>case</b> 'k': <b>case</b> 'K': Val *= 1024; <b>break</b>;
1885 default:
1886 <i>// Print an error message if unrecognized character!</i>
1888 }
1889 }
1890 }
1891 </pre></div>
1893 <p>This function implements a very simple parser for the kinds of strings we are
1895 example), it is good enough for this example. Note that we use the option
1896 itself to print out the error message (the <tt>error</tt> method always returns
1897 true) in order to get a nice error message (shown below). Now that we have our
1898 parser class, we can use it like this:</p>
1901 <b>static</b> <a href="#cl::opt">cl::opt</a><<b>unsigned</b>, <b>false</b>, FileSizeParser>
1902 MFS(<i>"max-file-size"</i>, <a href="#cl::desc">cl::desc</a>(<i>"Maximum file size to accept"</i>),
1904 </pre></div>
1906 <p>Which adds this to the output of our program:</p>
1909 OPTIONS:
1910 -help - display available options (--help-hidden for more)
1911 ...
1912 <b>-max-file-size=<size> - Maximum file size to accept</b>
1913 </pre></div>
1915 <p>And we can test that our parse works correctly now (the test program just
1916 prints out the max-file-size argument value):</p>
1919 $ ./test
1920 MFS: 0
1921 $ ./test -max-file-size=123MB
1922 MFS: 128974848
1923 $ ./test -max-file-size=3G
1924 MFS: 3221225472
1925 $ ./test -max-file-size=dog
1926 -max-file-size option: 'dog' value invalid for file size argument!
1927 </pre></div>
1929 <p>It looks like it works. The error message that we get is nice and helpful,
1931 tutorial.</p>
1933 </div>
1935 <!-- ======================================================================= -->
1938 </div>
1941 <p>Several of the LLVM libraries define static <tt>cl::opt</tt> instances that
1942 will automatically be included in any program that links with that library.
1943 This is a feature. However, sometimes it is necessary to know the value of the
1944 command line option outside of the library. In these cases the library does or
1945 should provide an external storage location that is accessible to users of the
1946 library. Examples of this include the <tt>llvm::DebugFlag</tt> exported by the
1947 <tt>lib/Support/Debug.cpp</tt> file and the <tt>llvm::TimePassesIsEnabled</tt>
1948 flag exported by the <tt>lib/VMCore/Pass.cpp</tt> file.</p>
1950 <p>TODO: complete this section</p>
1952 </div>
1954 <!-- ======================================================================= -->
1957 </div>
1961 <p>TODO: fill in this section</p>
1963 </div>
1965 <!-- *********************************************************************** -->
1967 <hr>
1968 <address>
1976 Last modified: $Date$
1977 </address>
1979 </body>
1980 </html>