3 ==============================
4 CommandLine 2.0 Library Manual
5 ==============================
10 This document describes the CommandLine argument processing library. It will
11 show you how to use it, and what it can do. The CommandLine library uses a
12 declarative approach to specifying the command line options that your program
13 takes. By default, these options declarations implicitly hold the value parsed
14 for the option declared (of course this `can be changed`_).
16 Although there are a **lot** of command line argument parsing libraries out
17 there in many different languages, none of them fit well with what I needed. By
18 looking at the features and problems of other libraries, I designed the
19 CommandLine library to have the following features:
21 #. Speed: The CommandLine library is very quick and uses little resources. The
22 parsing time of the library is directly proportional to the number of
23 arguments parsed, not the number of options recognized. Additionally,
24 command line argument values are captured transparently into user defined
25 global variables, which can be accessed like any other variable (and with the
28 #. Type Safe: As a user of CommandLine, you don't have to worry about
29 remembering the type of arguments that you want (is it an int? a string? a
30 bool? an enum?) and keep casting it around. Not only does this help prevent
31 error prone constructs, it also leads to dramatically cleaner source code.
33 #. No subclasses required: To use CommandLine, you instantiate variables that
34 correspond to the arguments that you would like to capture, you don't
35 subclass a parser. This means that you don't have to write **any**
38 #. Globally accessible: Libraries can specify command line arguments that are
39 automatically enabled in any tool that links to the library. This is
40 possible because the application doesn't have to keep a list of arguments to
41 pass to the parser. This also makes supporting `dynamically loaded options`_
44 #. Cleaner: CommandLine supports enum and other types directly, meaning that
45 there is less error and more security built into the library. You don't have
46 to worry about whether your integral command line argument accidentally got
47 assigned a value that is not valid for your enum type.
49 #. Powerful: The CommandLine library supports many different types of arguments,
50 from simple `boolean flags`_ to `scalars arguments`_ (`strings`_,
51 `integers`_, `enums`_, `doubles`_), to `lists of arguments`_. This is
52 possible because CommandLine is...
54 #. Extensible: It is very simple to add a new argument type to CommandLine.
55 Simply specify the parser that you want to use with the command line option
56 when you declare it. `Custom parsers`_ are no problem.
58 #. Labor Saving: The CommandLine library cuts down on the amount of grunt work
59 that you, the user, have to do. For example, it automatically provides a
60 ``-help`` option that shows the available command line options for your tool.
61 Additionally, it does most of the basic correctness checking for you.
63 #. Capable: The CommandLine library can handle lots of different forms of
64 options often found in real programs. For example, `positional`_ arguments,
65 ``ls`` style `grouping`_ options (to allow processing '``ls -lad``'
66 naturally), ``ld`` style `prefix`_ options (to parse '``-lmalloc
67 -L/usr/lib``'), and interpreter style options.
69 This document will hopefully let you jump in and start using CommandLine in your
70 utility quickly and painlessly. Additionally it should be a simple reference
71 manual to figure out how stuff works.
76 This section of the manual runs through a simple CommandLine'ification of a
77 basic compiler tool. This is intended to show you how to jump into using the
78 CommandLine library in your own program, and show you some of the cool things it
81 To start out, you need to include the CommandLine header file into your program:
85 #include "llvm/Support/CommandLine.h"
87 Additionally, you need to add this as the first line of your main program:
91 int main(int argc, char **argv) {
92 cl::ParseCommandLineOptions(argc, argv);
96 ... which actually parses the arguments and fills in the variable declarations.
98 Now that you are ready to support command line arguments, we need to tell the
99 system which ones we want, and what type of arguments they are. The CommandLine
100 library uses a declarative syntax to model command line arguments with the
101 global variable declarations that capture the parsed values. This means that
102 for every command line option that you would like to support, there should be a
103 global variable declaration to capture the result. For example, in a compiler,
104 we would like to support the Unix-standard '``-o <filename>``' option to specify
105 where to put the output. With the CommandLine library, this is represented like
108 .. _scalars arguments:
113 cl::opt<string> OutputFilename("o", cl::desc("Specify output filename"), cl::value_desc("filename"));
115 This declares a global variable "``OutputFilename``" that is used to capture the
116 result of the "``o``" argument (first parameter). We specify that this is a
117 simple scalar option by using the "``cl::opt``" template (as opposed to the
118 "``cl::list``" template), and tell the CommandLine library that the data
119 type that we are parsing is a string.
121 The second and third parameters (which are optional) are used to specify what to
122 output for the "``-help``" option. In this case, we get a line that looks like
127 USAGE: compiler [options]
130 -help - display available options (-help-hidden for more)
131 -o <filename> - Specify output filename
133 Because we specified that the command line option should parse using the
134 ``string`` data type, the variable declared is automatically usable as a real
135 string in all contexts that a normal C++ string object may be used. For
141 std::ofstream Output(OutputFilename.c_str());
142 if (Output.good()) ...
145 There are many different options that you can use to customize the command line
146 option handling library, but the above example shows the general interface to
147 these options. The options can be specified in any order, and are specified
148 with helper functions like `cl::desc(...)`_, so there are no positional
149 dependencies to remember. The available options are discussed in detail in the
152 Continuing the example, we would like to have our compiler take an input
153 filename as well as an output filename, but we do not want the input filename to
154 be specified with a hyphen (ie, not ``-filename.c``). To support this style of
155 argument, the CommandLine library allows for `positional`_ arguments to be
156 specified for the program. These positional arguments are filled with command
157 line parameters that are not in option form. We use this feature like this:
162 cl::opt<string> InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
164 This declaration indicates that the first positional argument should be treated
165 as the input filename. Here we use the `cl::init`_ option to specify an initial
166 value for the command line option, which is used if the option is not specified
167 (if you do not specify a `cl::init`_ modifier for an option, then the default
168 constructor for the data type is used to initialize the value). Command line
169 options default to being optional, so if we would like to require that the user
170 always specify an input filename, we would add the `cl::Required`_ flag, and we
171 could eliminate the `cl::init`_ modifier, like this:
175 cl::opt<string> InputFilename(cl::Positional, cl::desc("<input file>"), cl::Required);
177 Again, the CommandLine library does not require the options to be specified in
178 any particular order, so the above declaration is equivalent to:
182 cl::opt<string> InputFilename(cl::Positional, cl::Required, cl::desc("<input file>"));
184 By simply adding the `cl::Required`_ flag, the CommandLine library will
185 automatically issue an error if the argument is not specified, which shifts all
186 of the command line option verification code out of your application into the
187 library. This is just one example of how using flags can alter the default
188 behaviour of the library, on a per-option basis. By adding one of the
189 declarations above, the ``-help`` option synopsis is now extended to:
193 USAGE: compiler [options] <input file>
196 -help - display available options (-help-hidden for more)
197 -o <filename> - Specify output filename
199 ... indicating that an input filename is expected.
204 In addition to input and output filenames, we would like the compiler example to
205 support three boolean flags: "``-f``" to force writing binary output to a
206 terminal, "``--quiet``" to enable quiet mode, and "``-q``" for backwards
207 compatibility with some of our users. We can support these by declaring options
208 of boolean type like this:
212 cl::opt<bool> Force ("f", cl::desc("Enable binary output on terminals"));
213 cl::opt<bool> Quiet ("quiet", cl::desc("Don't print informational messages"));
214 cl::opt<bool> Quiet2("q", cl::desc("Don't print informational messages"), cl::Hidden);
216 This does what you would expect: it declares three boolean variables
217 ("``Force``", "``Quiet``", and "``Quiet2``") to recognize these options. Note
218 that the "``-q``" option is specified with the "`cl::Hidden`_" flag. This
219 modifier prevents it from being shown by the standard "``-help``" output (note
220 that it is still shown in the "``-help-hidden``" output).
222 The CommandLine library uses a `different parser`_ for different data types.
223 For example, in the string case, the argument passed to the option is copied
224 literally into the content of the string variable... we obviously cannot do that
225 in the boolean case, however, so we must use a smarter parser. In the case of
226 the boolean parser, it allows no options (in which case it assigns the value of
227 true to the variable), or it allows the values "``true``" or "``false``" to be
228 specified, allowing any of the following inputs:
232 compiler -f # No value, 'Force' == true
233 compiler -f=true # Value specified, 'Force' == true
234 compiler -f=TRUE # Value specified, 'Force' == true
235 compiler -f=FALSE # Value specified, 'Force' == false
237 ... you get the idea. The `bool parser`_ just turns the string values into
238 boolean values, and rejects things like '``compiler -f=foo``'. Similarly, the
239 `float`_, `double`_, and `int`_ parsers work like you would expect, using the
240 '``strtol``' and '``strtod``' C library calls to parse the string value into the
243 With the declarations above, "``compiler -help``" emits this:
247 USAGE: compiler [options] <input file>
250 -f - Enable binary output on terminals
251 -o - Override output filename
252 -quiet - Don't print informational messages
253 -help - display available options (-help-hidden for more)
255 and "``compiler -help-hidden``" prints this:
259 USAGE: compiler [options] <input file>
262 -f - Enable binary output on terminals
263 -o - Override output filename
264 -q - Don't print informational messages
265 -quiet - Don't print informational messages
266 -help - display available options (-help-hidden for more)
268 This brief example has shown you how to use the '`cl::opt`_' class to parse
269 simple scalar command line arguments. In addition to simple scalar arguments,
270 the CommandLine library also provides primitives to support CommandLine option
271 `aliases`_, and `lists`_ of options.
278 So far, the example works well, except for the fact that we need to check the
279 quiet condition like this now:
284 if (!Quiet && !Quiet2) printInformationalMessage(...);
287 ... which is a real pain! Instead of defining two values for the same
288 condition, we can use the "`cl::alias`_" class to make the "``-q``" option an
289 **alias** for the "``-quiet``" option, instead of providing a value itself:
293 cl::opt<bool> Force ("f", cl::desc("Overwrite output files"));
294 cl::opt<bool> Quiet ("quiet", cl::desc("Don't print informational messages"));
295 cl::alias QuietA("q", cl::desc("Alias for -quiet"), cl::aliasopt(Quiet));
297 The third line (which is the only one we modified from above) defines a "``-q``"
298 alias that updates the "``Quiet``" variable (as specified by the `cl::aliasopt`_
299 modifier) whenever it is specified. Because aliases do not hold state, the only
300 thing the program has to query is the ``Quiet`` variable now. Another nice
301 feature of aliases is that they automatically hide themselves from the ``-help``
302 output (although, again, they are still visible in the ``-help-hidden output``).
304 Now the application code can simply use:
309 if (!Quiet) printInformationalMessage(...);
312 ... which is much nicer! The "`cl::alias`_" can be used to specify an
313 alternative name for any variable type, and has many uses.
315 .. _unnamed alternatives using the generic parser:
317 Selecting an alternative from a set of possibilities
318 ----------------------------------------------------
320 So far we have seen how the CommandLine library handles builtin types like
321 ``std::string``, ``bool`` and ``int``, but how does it handle things it doesn't
322 know about, like enums or '``int*``'s?
324 The answer is that it uses a table-driven generic parser (unless you specify
325 your own parser, as described in the `Extension Guide`_). This parser maps
326 literal strings to whatever type is required, and requires you to tell it what
327 this mapping should be.
329 Let's say that we would like to add four optimization levels to our optimizer,
330 using the standard flags "``-g``", "``-O0``", "``-O1``", and "``-O2``". We
331 could easily implement this with boolean options like above, but there are
332 several problems with this strategy:
334 #. A user could specify more than one of the options at a time, for example,
335 "``compiler -O3 -O2``". The CommandLine library would not be able to catch
336 this erroneous input for us.
338 #. We would have to test 4 different variables to see which ones are set.
340 #. This doesn't map to the numeric levels that we want... so we cannot easily
341 see if some level >= "``-O1``" is enabled.
343 To cope with these problems, we can use an enum value, and have the CommandLine
344 library fill it in with the appropriate level directly, which is used like this:
352 cl::opt<OptLevel> OptimizationLevel(cl::desc("Choose optimization level:"),
354 clEnumVal(g , "No optimizations, enable debugging"),
355 clEnumVal(O1, "Enable trivial optimizations"),
356 clEnumVal(O2, "Enable default optimizations"),
357 clEnumVal(O3, "Enable expensive optimizations"),
361 if (OptimizationLevel >= O2) doPartialRedundancyElimination(...);
364 This declaration defines a variable "``OptimizationLevel``" of the
365 "``OptLevel``" enum type. This variable can be assigned any of the values that
366 are listed in the declaration (Note that the declaration list must be terminated
367 with the "``clEnumValEnd``" argument!). The CommandLine library enforces that
368 the user can only specify one of the options, and it ensure that only valid enum
369 values can be specified. The "``clEnumVal``" macros ensure that the command
370 line arguments matched the enum values. With this option added, our help output
375 USAGE: compiler [options] <input file>
378 Choose optimization level:
379 -g - No optimizations, enable debugging
380 -O1 - Enable trivial optimizations
381 -O2 - Enable default optimizations
382 -O3 - Enable expensive optimizations
383 -f - Enable binary output on terminals
384 -help - display available options (-help-hidden for more)
385 -o <filename> - Specify output filename
386 -quiet - Don't print informational messages
388 In this case, it is sort of awkward that flag names correspond directly to enum
389 names, because we probably don't want a enum definition named "``g``" in our
390 program. Because of this, we can alternatively write this example like this:
398 cl::opt<OptLevel> OptimizationLevel(cl::desc("Choose optimization level:"),
400 clEnumValN(Debug, "g", "No optimizations, enable debugging"),
401 clEnumVal(O1 , "Enable trivial optimizations"),
402 clEnumVal(O2 , "Enable default optimizations"),
403 clEnumVal(O3 , "Enable expensive optimizations"),
407 if (OptimizationLevel == Debug) outputDebugInfo(...);
410 By using the "``clEnumValN``" macro instead of "``clEnumVal``", we can directly
411 specify the name that the flag should get. In general a direct mapping is nice,
412 but sometimes you can't or don't want to preserve the mapping, which is when you
418 Another useful argument form is a named alternative style. We shall use this
419 style in our compiler to specify different debug levels that can be used.
420 Instead of each debug level being its own switch, we want to support the
421 following options, of which only one can be specified at a time:
422 "``--debug-level=none``", "``--debug-level=quick``",
423 "``--debug-level=detailed``". To do this, we use the exact same format as our
424 optimization level flags, but we also specify an option name. For this case,
425 the code looks like this:
430 nodebuginfo, quick, detailed
433 // Enable Debug Options to be specified on the command line
434 cl::opt<DebugLev> DebugLevel("debug_level", cl::desc("Set the debugging level:"),
436 clEnumValN(nodebuginfo, "none", "disable debug information"),
437 clEnumVal(quick, "enable quick debug information"),
438 clEnumVal(detailed, "enable detailed debug information"),
441 This definition defines an enumerated command line variable of type "``enum
442 DebugLev``", which works exactly the same way as before. The difference here is
443 just the interface exposed to the user of your program and the help output by
444 the "``-help``" option:
448 USAGE: compiler [options] <input file>
451 Choose optimization level:
452 -g - No optimizations, enable debugging
453 -O1 - Enable trivial optimizations
454 -O2 - Enable default optimizations
455 -O3 - Enable expensive optimizations
456 -debug_level - Set the debugging level:
457 =none - disable debug information
458 =quick - enable quick debug information
459 =detailed - enable detailed debug information
460 -f - Enable binary output on terminals
461 -help - display available options (-help-hidden for more)
462 -o <filename> - Specify output filename
463 -quiet - Don't print informational messages
465 Again, the only structural difference between the debug level declaration and
466 the optimization level declaration is that the debug level declaration includes
467 an option name (``"debug_level"``), which automatically changes how the library
468 processes the argument. The CommandLine library supports both forms so that you
469 can choose the form most appropriate for your application.
473 Parsing a list of options
474 -------------------------
476 Now that we have the standard run-of-the-mill argument types out of the way,
477 lets get a little wild and crazy. Lets say that we want our optimizer to accept
478 a **list** of optimizations to perform, allowing duplicates. For example, we
479 might want to run: "``compiler -dce -constprop -inline -dce -strip``". In this
480 case, the order of the arguments and the number of appearances is very
481 important. This is what the "``cl::list``" template is for. First, start by
482 defining an enum of the optimizations that you would like to perform:
487 // 'inline' is a C++ keyword, so name it 'inlining'
488 dce, constprop, inlining, strip
491 Then define your "``cl::list``" variable:
495 cl::list<Opts> OptimizationList(cl::desc("Available Optimizations:"),
497 clEnumVal(dce , "Dead Code Elimination"),
498 clEnumVal(constprop , "Constant Propagation"),
499 clEnumValN(inlining, "inline", "Procedure Integration"),
500 clEnumVal(strip , "Strip Symbols"),
503 This defines a variable that is conceptually of the type
504 "``std::vector<enum Opts>``". Thus, you can access it with standard vector
509 for (unsigned i = 0; i != OptimizationList.size(); ++i)
510 switch (OptimizationList[i])
513 ... to iterate through the list of options specified.
515 Note that the "``cl::list``" template is completely general and may be used with
516 any data types or other arguments that you can use with the "``cl::opt``"
517 template. One especially useful way to use a list is to capture all of the
518 positional arguments together if there may be more than one specified. In the
519 case of a linker, for example, the linker takes several '``.o``' files, and
520 needs to capture them into a list. This is naturally specified as:
525 cl::list<std::string> InputFilenames(cl::Positional, cl::desc("<Input files>"), cl::OneOrMore);
528 This variable works just like a "``vector<string>``" object. As such, accessing
529 the list is simple, just like above. In this example, we used the
530 `cl::OneOrMore`_ modifier to inform the CommandLine library that it is an error
531 if the user does not specify any ``.o`` files on our command line. Again, this
532 just reduces the amount of checking we have to do.
534 Collecting options as a set of flags
535 ------------------------------------
537 Instead of collecting sets of options in a list, it is also possible to gather
538 information for enum values in a **bit vector**. The representation used by the
539 `cl::bits`_ class is an ``unsigned`` integer. An enum value is represented by a
540 0/1 in the enum's ordinal value bit position. 1 indicating that the enum was
541 specified, 0 otherwise. As each specified value is parsed, the resulting enum's
542 bit is set in the option's bit vector:
546 bits |= 1 << (unsigned)enum;
548 Options that are specified multiple times are redundant. Any instances after
549 the first are discarded.
551 Reworking the above list example, we could replace `cl::list`_ with `cl::bits`_:
555 cl::bits<Opts> OptimizationBits(cl::desc("Available Optimizations:"),
557 clEnumVal(dce , "Dead Code Elimination"),
558 clEnumVal(constprop , "Constant Propagation"),
559 clEnumValN(inlining, "inline", "Procedure Integration"),
560 clEnumVal(strip , "Strip Symbols"),
563 To test to see if ``constprop`` was specified, we can use the ``cl:bits::isSet``
568 if (OptimizationBits.isSet(constprop)) {
572 It's also possible to get the raw bit vector using the ``cl::bits::getBits``
577 unsigned bits = OptimizationBits.getBits();
579 Finally, if external storage is used, then the location specified must be of
580 **type** ``unsigned``. In all other ways a `cl::bits`_ option is equivalent to a
583 .. _additional extra text:
585 Adding freeform text to help output
586 -----------------------------------
588 As our program grows and becomes more mature, we may decide to put summary
589 information about what it does into the help output. The help output is styled
590 to look similar to a Unix ``man`` page, providing concise information about a
591 program. Unix ``man`` pages, however often have a description about what the
592 program does. To add this to your CommandLine program, simply pass a third
593 argument to the `cl::ParseCommandLineOptions`_ call in main. This additional
594 argument is then printed as the overview information for your program, allowing
595 you to include any additional information that you want. For example:
599 int main(int argc, char **argv) {
600 cl::ParseCommandLineOptions(argc, argv, " CommandLine compiler example\n\n"
601 " This program blah blah blah...\n");
605 would yield the help output:
609 **OVERVIEW: CommandLine compiler example
611 This program blah blah blah...**
613 USAGE: compiler [options] <input file>
617 -help - display available options (-help-hidden for more)
618 -o <filename> - Specify output filename
625 Now that you know the basics of how to use the CommandLine library, this section
626 will give you the detailed information you need to tune how command line options
627 work, as well as information on more "advanced" command line option processing
631 .. _positional argument:
632 .. _Positional Arguments:
633 .. _Positional arguments section:
634 .. _positional options:
639 Positional arguments are those arguments that are not named, and are not
640 specified with a hyphen. Positional arguments should be used when an option is
641 specified by its position alone. For example, the standard Unix ``grep`` tool
642 takes a regular expression argument, and an optional filename to search through
643 (which defaults to standard input if a filename is not specified). Using the
644 CommandLine library, this would be specified as:
648 cl::opt<string> Regex (cl::Positional, cl::desc("<regular expression>"), cl::Required);
649 cl::opt<string> Filename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
651 Given these two option declarations, the ``-help`` output for our grep
652 replacement would look like this:
656 USAGE: spiffygrep [options] <regular expression> <input file>
659 -help - display available options (-help-hidden for more)
661 ... and the resultant program could be used just like the standard ``grep``
664 Positional arguments are sorted by their order of construction. This means that
665 command line options will be ordered according to how they are listed in a .cpp
666 file, but will not have an ordering defined if the positional arguments are
667 defined in multiple .cpp files. The fix for this problem is simply to define
668 all of your positional arguments in one .cpp file.
670 Specifying positional options with hyphens
671 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
673 Sometimes you may want to specify a value to your positional argument that
674 starts with a hyphen (for example, searching for '``-foo``' in a file). At
675 first, you will have trouble doing this, because it will try to find an argument
676 named '``-foo``', and will fail (and single quotes will not save you). Note
677 that the system ``grep`` has the same problem:
681 $ spiffygrep '-foo' test.txt
682 Unknown command line argument '-foo'. Try: spiffygrep -help'
684 $ grep '-foo' test.txt
685 grep: illegal option -- f
686 grep: illegal option -- o
687 grep: illegal option -- o
688 Usage: grep -hblcnsviw pattern file . . .
690 The solution for this problem is the same for both your tool and the system
691 version: use the '``--``' marker. When the user specifies '``--``' on the
692 command line, it is telling the program that all options after the '``--``'
693 should be treated as positional arguments, not options. Thus, we can use it
698 $ spiffygrep -- -foo test.txt
701 Determining absolute position with getPosition()
702 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
704 Sometimes an option can affect or modify the meaning of another option. For
705 example, consider ``gcc``'s ``-x LANG`` option. This tells ``gcc`` to ignore the
706 suffix of subsequent positional arguments and force the file to be interpreted
707 as if it contained source code in language ``LANG``. In order to handle this
708 properly, you need to know the absolute position of each argument, especially
709 those in lists, so their interaction(s) can be applied correctly. This is also
710 useful for options like ``-llibname`` which is actually a positional argument
711 that starts with a dash.
713 So, generally, the problem is that you have two ``cl::list`` variables that
714 interact in some way. To ensure the correct interaction, you can use the
715 ``cl::list::getPosition(optnum)`` method. This method returns the absolute
716 position (as found on the command line) of the ``optnum`` item in the
719 The idiom for usage is like this:
723 static cl::list<std::string> Files(cl::Positional, cl::OneOrMore);
724 static cl::list<std::string> Libraries("l", cl::ZeroOrMore);
726 int main(int argc, char**argv) {
728 std::vector<std::string>::iterator fileIt = Files.begin();
729 std::vector<std::string>::iterator libIt = Libraries.begin();
730 unsigned libPos = 0, filePos = 0;
732 if ( libIt != Libraries.end() )
733 libPos = Libraries.getPosition( libIt - Libraries.begin() );
736 if ( fileIt != Files.end() )
737 filePos = Files.getPosition( fileIt - Files.begin() );
741 if ( filePos != 0 && (libPos == 0 || filePos < libPos) ) {
742 // Source File Is next
745 else if ( libPos != 0 && (filePos == 0 || libPos < filePos) ) {
750 break; // we're done with the list
754 Note that, for compatibility reasons, the ``cl::opt`` also supports an
755 ``unsigned getPosition()`` option that will provide the absolute position of
756 that option. You can apply the same approach as above with a ``cl::opt`` and a
757 ``cl::list`` option as you can with two lists.
759 .. _interpreter style options:
760 .. _cl::ConsumeAfter:
761 .. _this section for more information:
763 The ``cl::ConsumeAfter`` modifier
764 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
766 The ``cl::ConsumeAfter`` `formatting option`_ is used to construct programs that
767 use "interpreter style" option processing. With this style of option
768 processing, all arguments specified after the last positional argument are
769 treated as special interpreter arguments that are not interpreted by the command
772 As a concrete example, lets say we are developing a replacement for the standard
773 Unix Bourne shell (``/bin/sh``). To run ``/bin/sh``, first you specify options
774 to the shell itself (like ``-x`` which turns on trace output), then you specify
775 the name of the script to run, then you specify arguments to the script. These
776 arguments to the script are parsed by the Bourne shell command line option
777 processor, but are not interpreted as options to the shell itself. Using the
778 CommandLine library, we would specify this as:
782 cl::opt<string> Script(cl::Positional, cl::desc("<input script>"), cl::init("-"));
783 cl::list<string> Argv(cl::ConsumeAfter, cl::desc("<program arguments>..."));
784 cl::opt<bool> Trace("x", cl::desc("Enable trace output"));
786 which automatically provides the help output:
790 USAGE: spiffysh [options] <input script> <program arguments>...
793 -help - display available options (-help-hidden for more)
794 -x - Enable trace output
796 At runtime, if we run our new shell replacement as ```spiffysh -x test.sh -a -x
797 -y bar``', the ``Trace`` variable will be set to true, the ``Script`` variable
798 will be set to "``test.sh``", and the ``Argv`` list will contain ``["-a", "-x",
799 "-y", "bar"]``, because they were specified after the last positional argument
800 (which is the script name).
802 There are several limitations to when ``cl::ConsumeAfter`` options can be
803 specified. For example, only one ``cl::ConsumeAfter`` can be specified per
804 program, there must be at least one `positional argument`_ specified, there must
805 not be any `cl::list`_ positional arguments, and the ``cl::ConsumeAfter`` option
806 should be a `cl::list`_ option.
809 .. _Internal vs External Storage:
811 Internal vs External Storage
812 ----------------------------
814 By default, all command line options automatically hold the value that they
815 parse from the command line. This is very convenient in the common case,
816 especially when combined with the ability to define command line options in the
817 files that use them. This is called the internal storage model.
819 Sometimes, however, it is nice to separate the command line option processing
820 code from the storage of the value parsed. For example, lets say that we have a
821 '``-debug``' option that we would like to use to enable debug information across
822 the entire body of our program. In this case, the boolean value controlling the
823 debug code should be globally accessible (in a header file, for example) yet the
824 command line option processing code should not be exposed to all of these
825 clients (requiring lots of .cpp files to ``#include CommandLine.h``).
827 To do this, set up your .h file with your option, like this for example:
831 // DebugFlag.h - Get access to the '-debug' command line option
834 // DebugFlag - This boolean is set to true if the '-debug' command line option
835 // is specified. This should probably not be referenced directly, instead, use
836 // the DEBUG macro below.
838 extern bool DebugFlag;
840 // DEBUG macro - This macro should be used by code to emit debug information.
841 // In the '-debug' option is specified on the command line, and if this is a
842 // debug build, then the code specified as the option to the macro will be
843 // executed. Otherwise it will not be.
847 #define DEBUG(X) do { if (DebugFlag) { X; } } while (0)
850 This allows clients to blissfully use the ``DEBUG()`` macro, or the
851 ``DebugFlag`` explicitly if they want to. Now we just need to be able to set
852 the ``DebugFlag`` boolean when the option is set. To do this, we pass an
853 additional argument to our command line argument processor, and we specify where
854 to fill in with the `cl::location`_ attribute:
858 bool DebugFlag; // the actual value
859 static cl::opt<bool, true> // The parser
860 Debug("debug", cl::desc("Enable debug output"), cl::Hidden, cl::location(DebugFlag));
862 In the above example, we specify "``true``" as the second argument to the
863 `cl::opt`_ template, indicating that the template should not maintain a copy of
864 the value itself. In addition to this, we specify the `cl::location`_
865 attribute, so that ``DebugFlag`` is automatically set.
870 This section describes the basic attributes that you can specify on options.
872 * The option name attribute (which is required for all options, except
873 `positional options`_) specifies what the option name is. This option is
874 specified in simple double quotes:
878 cl::opt<**bool**> Quiet("quiet");
882 * The **cl::desc** attribute specifies a description for the option to be
883 shown in the ``-help`` output for the program.
887 * The **cl::value_desc** attribute specifies a string that can be used to
888 fine tune the ``-help`` output for a command line option. Look `here`_ for an
893 * The **cl::init** attribute specifies an initial value for a `scalar`_
894 option. If this attribute is not specified then the command line option value
895 defaults to the value created by the default constructor for the
900 If you specify both **cl::init** and **cl::location** for an option, you
901 must specify **cl::location** first, so that when the command-line parser
902 sees **cl::init**, it knows where to put the initial value. (You will get an
903 error at runtime if you don't put them in the right order.)
907 * The **cl::location** attribute where to store the value for a parsed command
908 line option if using external storage. See the section on `Internal vs
909 External Storage`_ for more information.
913 * The **cl::aliasopt** attribute specifies which option a `cl::alias`_ option is
918 * The **cl::values** attribute specifies the string-to-value mapping to be used
919 by the generic parser. It takes a **clEnumValEnd terminated** list of
920 (option, value, description) triplets that specify the option name, the value
921 mapped to, and the description shown in the ``-help`` for the tool. Because
922 the generic parser is used most frequently with enum values, two macros are
925 #. The **clEnumVal** macro is used as a nice simple way to specify a triplet
926 for an enum. This macro automatically makes the option name be the same as
927 the enum name. The first option to the macro is the enum, the second is
928 the description for the command line option.
930 #. The **clEnumValN** macro is used to specify macro options where the option
931 name doesn't equal the enum name. For this macro, the first argument is
932 the enum value, the second is the flag name, and the second is the
935 You will get a compile time error if you try to use cl::values with a parser
936 that does not support it.
940 * The **cl::multi_val** attribute specifies that this option takes has multiple
941 values (example: ``-sectalign segname sectname sectvalue``). This attribute
942 takes one unsigned argument - the number of values for the option. This
943 attribute is valid only on ``cl::list`` options (and will fail with compile
944 error if you try to use it with other option types). It is allowed to use all
945 of the usual modifiers on multi-valued options (besides
946 ``cl::ValueDisallowed``, obviously).
951 Option modifiers are the flags and expressions that you pass into the
952 constructors for `cl::opt`_ and `cl::list`_. These modifiers give you the
953 ability to tweak how options are parsed and how ``-help`` output is generated to
954 fit your application well.
956 These options fall into five main categories:
958 #. Hiding an option from ``-help`` output
960 #. Controlling the number of occurrences required and allowed
962 #. Controlling whether or not a value must be specified
964 #. Controlling other formatting options
966 #. Miscellaneous option modifiers
968 It is not possible to specify two options from the same category (you'll get a
969 runtime error) to a single option, except for options in the miscellaneous
970 category. The CommandLine library specifies defaults for all of these settings
971 that are the most useful in practice and the most common, which mean that you
972 usually shouldn't have to worry about these.
974 Hiding an option from ``-help`` output
975 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
977 The ``cl::NotHidden``, ``cl::Hidden``, and ``cl::ReallyHidden`` modifiers are
978 used to control whether or not an option appears in the ``-help`` and
979 ``-help-hidden`` output for the compiled program:
983 * The **cl::NotHidden** modifier (which is the default for `cl::opt`_ and
984 `cl::list`_ options) indicates the option is to appear in both help
989 * The **cl::Hidden** modifier (which is the default for `cl::alias`_ options)
990 indicates that the option should not appear in the ``-help`` output, but
991 should appear in the ``-help-hidden`` output.
993 .. _cl::ReallyHidden:
995 * The **cl::ReallyHidden** modifier indicates that the option should not appear
998 Controlling the number of occurrences required and allowed
999 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1001 This group of options is used to control how many time an option is allowed (or
1002 required) to be specified on the command line of your program. Specifying a
1003 value for this setting allows the CommandLine library to do error checking for
1006 The allowed values for this option group are:
1010 * The **cl::Optional** modifier (which is the default for the `cl::opt`_ and
1011 `cl::alias`_ classes) indicates that your program will allow either zero or
1012 one occurrence of the option to be specified.
1016 * The **cl::ZeroOrMore** modifier (which is the default for the `cl::list`_
1017 class) indicates that your program will allow the option to be specified zero
1022 * The **cl::Required** modifier indicates that the specified option must be
1023 specified exactly one time.
1027 * The **cl::OneOrMore** modifier indicates that the option must be specified at
1030 * The **cl::ConsumeAfter** modifier is described in the `Positional arguments
1033 If an option is not specified, then the value of the option is equal to the
1034 value specified by the `cl::init`_ attribute. If the ``cl::init`` attribute is
1035 not specified, the option value is initialized with the default constructor for
1038 If an option is specified multiple times for an option of the `cl::opt`_ class,
1039 only the last value will be retained.
1041 Controlling whether or not a value must be specified
1042 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1044 This group of options is used to control whether or not the option allows a
1045 value to be present. In the case of the CommandLine library, a value is either
1046 specified with an equal sign (e.g. '``-index-depth=17``') or as a trailing
1047 string (e.g. '``-o a.out``').
1049 The allowed values for this option group are:
1051 .. _cl::ValueOptional:
1053 * The **cl::ValueOptional** modifier (which is the default for ``bool`` typed
1054 options) specifies that it is acceptable to have a value, or not. A boolean
1055 argument can be enabled just by appearing on the command line, or it can have
1056 an explicit '``-foo=true``'. If an option is specified with this mode, it is
1057 illegal for the value to be provided without the equal sign. Therefore
1058 '``-foo true``' is illegal. To get this behavior, you must use
1059 the `cl::ValueRequired`_ modifier.
1061 .. _cl::ValueRequired:
1063 * The **cl::ValueRequired** modifier (which is the default for all other types
1064 except for `unnamed alternatives using the generic parser`_) specifies that a
1065 value must be provided. This mode informs the command line library that if an
1066 option is not provides with an equal sign, that the next argument provided
1067 must be the value. This allows things like '``-o a.out``' to work.
1069 .. _cl::ValueDisallowed:
1071 * The **cl::ValueDisallowed** modifier (which is the default for `unnamed
1072 alternatives using the generic parser`_) indicates that it is a runtime error
1073 for the user to specify a value. This can be provided to disallow users from
1074 providing options to boolean options (like '``-foo=true``').
1076 In general, the default values for this option group work just like you would
1077 want them to. As mentioned above, you can specify the `cl::ValueDisallowed`_
1078 modifier to a boolean argument to restrict your command line parser. These
1079 options are mostly useful when `extending the library`_.
1081 .. _formatting option:
1083 Controlling other formatting options
1084 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1086 The formatting option group is used to specify that the command line option has
1087 special abilities and is otherwise different from other command line arguments.
1088 As usual, you can only specify one of these arguments at most.
1090 .. _cl::NormalFormatting:
1092 * The **cl::NormalFormatting** modifier (which is the default all options)
1093 specifies that this option is "normal".
1097 * The **cl::Positional** modifier specifies that this is a positional argument
1098 that does not have a command line option associated with it. See the
1099 `Positional Arguments`_ section for more information.
1101 * The **cl::ConsumeAfter** modifier specifies that this option is used to
1102 capture "interpreter style" arguments. See `this section for more
1108 * The **cl::Prefix** modifier specifies that this option prefixes its value.
1109 With 'Prefix' options, the equal sign does not separate the value from the
1110 option name specified. Instead, the value is everything after the prefix,
1111 including any equal sign if present. This is useful for processing odd
1112 arguments like ``-lmalloc`` and ``-L/usr/lib`` in a linker tool or
1113 ``-DNAME=value`` in a compiler tool. Here, the '``l``', '``D``' and '``L``'
1114 options are normal string (or list) options, that have the **cl::Prefix**
1115 modifier added to allow the CommandLine library to recognize them. Note that
1116 **cl::Prefix** options must not have the **cl::ValueDisallowed** modifier
1122 * The **cl::Grouping** modifier is used to implement Unix-style tools (like
1123 ``ls``) that have lots of single letter arguments, but only require a single
1124 dash. For example, the '``ls -labF``' command actually enables four different
1125 options, all of which are single letters. Note that **cl::Grouping** options
1128 The CommandLine library does not restrict how you use the **cl::Prefix** or
1129 **cl::Grouping** modifiers, but it is possible to specify ambiguous argument
1130 settings. Thus, it is possible to have multiple letter options that are prefix
1131 or grouping options, and they will still work as designed.
1133 To do this, the CommandLine library uses a greedy algorithm to parse the input
1134 option into (potentially multiple) prefix and grouping options. The strategy
1135 basically looks like this:
1139 parse(string OrigInput) {
1141 1. string input = OrigInput;
1142 2. if (isOption(input)) return getOption(input).parse(); // Normal option
1143 3. while (!isOption(input) && !input.empty()) input.pop_back(); // Remove the last letter
1144 4. if (input.empty()) return error(); // No matching option
1145 5. if (getOption(input).isPrefix())
1146 return getOption(input).parse(input);
1147 6. while (!input.empty()) { // Must be grouping options
1148 getOption(input).parse();
1149 OrigInput.erase(OrigInput.begin(), OrigInput.begin()+input.length());
1151 while (!isOption(input) && !input.empty()) input.pop_back();
1153 7. if (!OrigInput.empty()) error();
1157 Miscellaneous option modifiers
1158 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1160 The miscellaneous option modifiers are the only flags where you can specify more
1161 than one flag from the set: they are not mutually exclusive. These flags
1162 specify boolean properties that modify the option.
1164 .. _cl::CommaSeparated:
1166 * The **cl::CommaSeparated** modifier indicates that any commas specified for an
1167 option's value should be used to split the value up into multiple values for
1168 the option. For example, these two options are equivalent when
1169 ``cl::CommaSeparated`` is specified: "``-foo=a -foo=b -foo=c``" and
1170 "``-foo=a,b,c``". This option only makes sense to be used in a case where the
1171 option is allowed to accept one or more values (i.e. it is a `cl::list`_
1174 .. _cl::PositionalEatsArgs:
1176 * The **cl::PositionalEatsArgs** modifier (which only applies to positional
1177 arguments, and only makes sense for lists) indicates that positional argument
1178 should consume any strings after it (including strings that start with a "-")
1179 up until another recognized positional argument. For example, if you have two
1180 "eating" positional arguments, "``pos1``" and "``pos2``", the string "``-pos1
1181 -foo -bar baz -pos2 -bork``" would cause the "``-foo -bar -baz``" strings to
1182 be applied to the "``-pos1``" option and the "``-bork``" string to be applied
1183 to the "``-pos2``" option.
1187 * The **cl::Sink** modifier is used to handle unknown options. If there is at
1188 least one option with ``cl::Sink`` modifier specified, the parser passes
1189 unrecognized option strings to it as values instead of signaling an error. As
1190 with ``cl::CommaSeparated``, this modifier only makes sense with a `cl::list`_
1193 So far, these are the only three miscellaneous option modifiers.
1200 Some systems, such as certain variants of Microsoft Windows and some older
1201 Unices have a relatively low limit on command-line length. It is therefore
1202 customary to use the so-called 'response files' to circumvent this
1203 restriction. These files are mentioned on the command-line (using the "@file")
1204 syntax. The program reads these files and inserts the contents into argv,
1205 thereby working around the command-line length limits. Response files are
1206 enabled by an optional fourth argument to `cl::ParseEnvironmentOptions`_ and
1207 `cl::ParseCommandLineOptions`_.
1209 Top-Level Classes and Functions
1210 -------------------------------
1212 Despite all of the built-in flexibility, the CommandLine option library really
1213 only consists of one function `cl::ParseCommandLineOptions`_) and three main
1214 classes: `cl::opt`_, `cl::list`_, and `cl::alias`_. This section describes
1215 these three classes in detail.
1217 .. _cl::ParseCommandLineOptions:
1219 The ``cl::ParseCommandLineOptions`` function
1220 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1222 The ``cl::ParseCommandLineOptions`` function is designed to be called directly
1223 from ``main``, and is used to fill in the values of all of the command line
1224 option variables once ``argc`` and ``argv`` are available.
1226 The ``cl::ParseCommandLineOptions`` function requires two parameters (``argc``
1227 and ``argv``), but may also take an optional third parameter which holds
1228 `additional extra text`_ to emit when the ``-help`` option is invoked, and a
1229 fourth boolean parameter that enables `response files`_.
1231 .. _cl::ParseEnvironmentOptions:
1233 The ``cl::ParseEnvironmentOptions`` function
1234 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1236 The ``cl::ParseEnvironmentOptions`` function has mostly the same effects as
1237 `cl::ParseCommandLineOptions`_, except that it is designed to take values for
1238 options from an environment variable, for those cases in which reading the
1239 command line is not convenient or desired. It fills in the values of all the
1240 command line option variables just like `cl::ParseCommandLineOptions`_ does.
1242 It takes four parameters: the name of the program (since ``argv`` may not be
1243 available, it can't just look in ``argv[0]``), the name of the environment
1244 variable to examine, the optional `additional extra text`_ to emit when the
1245 ``-help`` option is invoked, and the boolean switch that controls whether
1246 `response files`_ should be read.
1248 ``cl::ParseEnvironmentOptions`` will break the environment variable's value up
1249 into words and then process them using `cl::ParseCommandLineOptions`_.
1250 **Note:** Currently ``cl::ParseEnvironmentOptions`` does not support quoting, so
1251 an environment variable containing ``-option "foo bar"`` will be parsed as three
1252 words, ``-option``, ``"foo``, and ``bar"``, which is different from what you
1253 would get from the shell with the same input.
1255 The ``cl::SetVersionPrinter`` function
1256 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1258 The ``cl::SetVersionPrinter`` function is designed to be called directly from
1259 ``main`` and *before* ``cl::ParseCommandLineOptions``. Its use is optional. It
1260 simply arranges for a function to be called in response to the ``--version``
1261 option instead of having the ``CommandLine`` library print out the usual version
1262 string for LLVM. This is useful for programs that are not part of LLVM but wish
1263 to use the ``CommandLine`` facilities. Such programs should just define a small
1264 function that takes no arguments and returns ``void`` and that prints out
1265 whatever version information is appropriate for the program. Pass the address of
1266 that function to ``cl::SetVersionPrinter`` to arrange for it to be called when
1267 the ``--version`` option is given by the user.
1272 The ``cl::opt`` class
1273 ^^^^^^^^^^^^^^^^^^^^^
1275 The ``cl::opt`` class is the class used to represent scalar command line
1276 options, and is the one used most of the time. It is a templated class which
1277 can take up to three arguments (all except for the first have default values
1283 template <class DataType, bool ExternalStorage = false,
1284 class ParserClass = parser<DataType> >
1288 The first template argument specifies what underlying data type the command line
1289 argument is, and is used to select a default parser implementation. The second
1290 template argument is used to specify whether the option should contain the
1291 storage for the option (the default) or whether external storage should be used
1292 to contain the value parsed for the option (see `Internal vs External Storage`_
1293 for more information).
1295 The third template argument specifies which parser to use. The default value
1296 selects an instantiation of the ``parser`` class based on the underlying data
1297 type of the option. In general, this default works well for most applications,
1298 so this option is only used when using a `custom parser`_.
1300 .. _lists of arguments:
1303 The ``cl::list`` class
1304 ^^^^^^^^^^^^^^^^^^^^^^
1306 The ``cl::list`` class is the class used to represent a list of command line
1307 options. It too is a templated class which can take up to three arguments:
1312 template <class DataType, class Storage = bool,
1313 class ParserClass = parser<DataType> >
1317 This class works the exact same as the `cl::opt`_ class, except that the second
1318 argument is the **type** of the external storage, not a boolean value. For this
1319 class, the marker type '``bool``' is used to indicate that internal storage
1324 The ``cl::bits`` class
1325 ^^^^^^^^^^^^^^^^^^^^^^
1327 The ``cl::bits`` class is the class used to represent a list of command line
1328 options in the form of a bit vector. It is also a templated class which can
1329 take up to three arguments:
1334 template <class DataType, class Storage = bool,
1335 class ParserClass = parser<DataType> >
1339 This class works the exact same as the `cl::list`_ class, except that the second
1340 argument must be of **type** ``unsigned`` if external storage is used.
1344 The ``cl::alias`` class
1345 ^^^^^^^^^^^^^^^^^^^^^^^
1347 The ``cl::alias`` class is a nontemplated class that is used to form aliases for
1356 The `cl::aliasopt`_ attribute should be used to specify which option this is an
1357 alias for. Alias arguments default to being `cl::Hidden`_, and use the aliased
1358 options parser to do the conversion from string to data.
1362 The ``cl::extrahelp`` class
1363 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
1365 The ``cl::extrahelp`` class is a nontemplated class that allows extra help text
1366 to be printed out for the ``-help`` option.
1374 To use the extrahelp, simply construct one with a ``const char*`` parameter to
1375 the constructor. The text passed to the constructor will be printed at the
1376 bottom of the help message, verbatim. Note that multiple ``cl::extrahelp``
1377 **can** be used, but this practice is discouraged. If your tool needs to print
1378 additional help information, put all that help into a single ``cl::extrahelp``
1385 cl::extrahelp("\nADDITIONAL HELP:\n\n This is the extra help\n");
1387 .. _different parser:
1388 .. _discussed previously:
1393 Parsers control how the string value taken from the command line is translated
1394 into a typed value, suitable for use in a C++ program. By default, the
1395 CommandLine library uses an instance of ``parser<type>`` if the command line
1396 option specifies that it uses values of type '``type``'. Because of this,
1397 custom option processing is specified with specializations of the '``parser``'
1400 The CommandLine library provides the following builtin parser specializations,
1401 which are sufficient for most applications. It can, however, also be extended to
1402 work with new data types and new ways of interpreting the same data. See the
1403 `Writing a Custom Parser`_ for more details on this type of library extension.
1408 * The generic ``parser<t>`` parser can be used to map strings values to any data
1409 type, through the use of the `cl::values`_ property, which specifies the
1410 mapping information. The most common use of this parser is for parsing enum
1411 values, which allows you to use the CommandLine library for all of the error
1412 checking to make sure that only valid enum values are specified (as opposed to
1413 accepting arbitrary strings). Despite this, however, the generic parser class
1414 can be used for any data type.
1419 * The **parser<bool> specialization** is used to convert boolean strings to a
1420 boolean value. Currently accepted strings are "``true``", "``TRUE``",
1421 "``True``", "``1``", "``false``", "``FALSE``", "``False``", and "``0``".
1423 * The **parser<boolOrDefault> specialization** is used for cases where the value
1424 is boolean, but we also need to know whether the option was specified at all.
1425 boolOrDefault is an enum with 3 values, BOU_UNSET, BOU_TRUE and BOU_FALSE.
1426 This parser accepts the same strings as **``parser<bool>``**.
1430 * The **parser<string> specialization** simply stores the parsed string into the
1431 string value specified. No conversion or modification of the data is
1437 * The **parser<int> specialization** uses the C ``strtol`` function to parse the
1438 string input. As such, it will accept a decimal number (with an optional '+'
1439 or '-' prefix) which must start with a non-zero digit. It accepts octal
1440 numbers, which are identified with a '``0``' prefix digit, and hexadecimal
1441 numbers with a prefix of '``0x``' or '``0X``'.
1447 * The **parser<double>** and **parser<float> specializations** use the standard
1448 C ``strtod`` function to convert floating point strings into floating point
1449 values. As such, a broad range of string formats is supported, including
1450 exponential notation (ex: ``1.7e15``) and properly supports locales.
1452 .. _Extension Guide:
1453 .. _extending the library:
1458 Although the CommandLine library has a lot of functionality built into it
1459 already (as discussed previously), one of its true strengths lie in its
1460 extensibility. This section discusses how the CommandLine library works under
1461 the covers and illustrates how to do some simple, common, extensions.
1465 .. _Writing a Custom Parser:
1467 Writing a custom parser
1468 -----------------------
1470 One of the simplest and most common extensions is the use of a custom parser.
1471 As `discussed previously`_, parsers are the portion of the CommandLine library
1472 that turns string input from the user into a particular parsed data type,
1473 validating the input in the process.
1475 There are two ways to use a new parser:
1477 #. Specialize the `cl::parser`_ template for your custom data type.
1479 This approach has the advantage that users of your custom data type will
1480 automatically use your custom parser whenever they define an option with a
1481 value type of your data type. The disadvantage of this approach is that it
1482 doesn't work if your fundamental data type is something that is already
1485 #. Write an independent class, using it explicitly from options that need it.
1487 This approach works well in situations where you would line to parse an
1488 option using special syntax for a not-very-special data-type. The drawback
1489 of this approach is that users of your parser have to be aware that they are
1490 using your parser instead of the builtin ones.
1492 To guide the discussion, we will discuss a custom parser that accepts file
1493 sizes, specified with an optional unit after the numeric size. For example, we
1494 would like to parse "102kb", "41M", "1G" into the appropriate integer value. In
1495 this case, the underlying data type we want to parse into is '``unsigned``'. We
1496 choose approach #2 above because we don't want to make this the default for all
1497 ``unsigned`` options.
1499 To start out, we declare our new ``FileSizeParser`` class:
1503 struct FileSizeParser : public cl::basic_parser<unsigned> {
1504 // parse - Return true on error.
1505 bool parse(cl::Option &O, const char *ArgName, const std::string &ArgValue,
1509 Our new class inherits from the ``cl::basic_parser`` template class to fill in
1510 the default, boiler plate code for us. We give it the data type that we parse
1511 into, the last argument to the ``parse`` method, so that clients of our custom
1512 parser know what object type to pass in to the parse method. (Here we declare
1513 that we parse into '``unsigned``' variables.)
1515 For most purposes, the only method that must be implemented in a custom parser
1516 is the ``parse`` method. The ``parse`` method is called whenever the option is
1517 invoked, passing in the option itself, the option name, the string to parse, and
1518 a reference to a return value. If the string to parse is not well-formed, the
1519 parser should output an error message and return true. Otherwise it should
1520 return false and set '``Val``' to the parsed value. In our example, we
1521 implement ``parse`` as:
1525 bool FileSizeParser::parse(cl::Option &O, const char *ArgName,
1526 const std::string &Arg, unsigned &Val) {
1527 const char *ArgStart = Arg.c_str();
1530 // Parse integer part, leaving 'End' pointing to the first non-integer char
1531 Val = (unsigned)strtol(ArgStart, &End, 0);
1535 case 0: return false; // No error
1536 case 'i': // Ignore the 'i' in KiB if people use that
1537 case 'b': case 'B': // Ignore B suffix
1540 case 'g': case 'G': Val *= 1024*1024*1024; break;
1541 case 'm': case 'M': Val *= 1024*1024; break;
1542 case 'k': case 'K': Val *= 1024; break;
1545 // Print an error message if unrecognized character!
1546 return O.error("'" + Arg + "' value invalid for file size argument!");
1551 This function implements a very simple parser for the kinds of strings we are
1552 interested in. Although it has some holes (it allows "``123KKK``" for example),
1553 it is good enough for this example. Note that we use the option itself to print
1554 out the error message (the ``error`` method always returns true) in order to get
1555 a nice error message (shown below). Now that we have our parser class, we can
1560 static cl::opt<unsigned, false, FileSizeParser>
1561 MFS("max-file-size", cl::desc("Maximum file size to accept"),
1562 cl::value_desc("size"));
1564 Which adds this to the output of our program:
1569 -help - display available options (-help-hidden for more)
1571 -max-file-size=<size> - Maximum file size to accept
1573 And we can test that our parse works correctly now (the test program just prints
1574 out the max-file-size argument value):
1580 $ ./test -max-file-size=123MB
1582 $ ./test -max-file-size=3G
1584 $ ./test -max-file-size=dog
1585 -max-file-size option: 'dog' value invalid for file size argument!
1587 It looks like it works. The error message that we get is nice and helpful, and
1588 we seem to accept reasonable file sizes. This wraps up the "custom parser"
1591 Exploiting external storage
1592 ---------------------------
1594 Several of the LLVM libraries define static ``cl::opt`` instances that will
1595 automatically be included in any program that links with that library. This is
1596 a feature. However, sometimes it is necessary to know the value of the command
1597 line option outside of the library. In these cases the library does or should
1598 provide an external storage location that is accessible to users of the
1599 library. Examples of this include the ``llvm::DebugFlag`` exported by the
1600 ``lib/Support/Debug.cpp`` file and the ``llvm::TimePassesIsEnabled`` flag
1601 exported by the ``lib/VMCore/PassManager.cpp`` file.
1605 TODO: complete this section
1607 .. _dynamically loaded options:
1609 Dynamically adding command line options
1613 TODO: fill in this section