/usr/share/doc/libargtable2-docs/argtable2-intro.html is in libargtable2-docs 13-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 | <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="CONTENT-TYPE" CONTENT="text/html; charset=us-ascii">
<TITLE>Introduction to argtable</TITLE>
<META NAME="GENERATOR" CONTENT="OpenOffice.org 1.1.4 (FreeBSD)">
<META NAME="CREATED" CONTENT="20030925;11311700">
<META NAME="CHANGED" CONTENT="20051003;18290000">
<META NAME="CLASSIFICATION" CONTENT="ANSI C command line parser">
<META NAME="KEYWORDS" CONTENT="command line parser">
<STYLE>
<!--
@page { size: 21.01cm 29.69cm }
H1 { margin-top: 2.01cm }
-->
</STYLE>
</HEAD>
<BODY LANG="en-AU" DIR="LTR">
<H1 ALIGN=CENTER>Introduction to argtable-2.x<BR><FONT SIZE=3>An ANSI
C library for parsing GNU style command line arguments</FONT></H1>
<P ALIGN=CENTER STYLE="margin-bottom: 0.99cm">Stewart
Heitmann<BR><I>sheitmann@users.sourceforge.net</I></P>
<P>Parsing a program's command line arguments has always been a
distraction from the main programming task at hand. The argtable
library simplifies the job by enabling the programmer to define the
command line options directly in the source code as a static array of
structs and then pass that array to argtable library functions which
parse the command line accordingly. The values extracted from the
comand line are deposited directly into user-defined program
variables where they can be accessed by the main program. Argtable
can also generate descriptions of the command line syntax from that
same array for display as on-line help. The software is freely
available under the terms of the GNU Library General Public License
(LGPL).</P>
<P ALIGN=CENTER><FONT FACE="Courier, monospace"><FONT SIZE=3>myprog
-abc --scalar=7 --verbose -o myfile <file> [<file>]</FONT></FONT></P>
<P ALIGN=LEFT>Argtable uses GNUgetopt to perform the actual parsing
so it is 100% GNU compatable. It supports both GNU short options (as
in -abc, -o myfile) GNU long options (as in –scalar=7,
–verbose), as well as untagged arguments (as in <file>
[<file>]). It does not support non-GNU command line formats,
such as DOS /X /Y /Z style options.</P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>How
it works</FONT></FONT></P>
<P ALIGN=LEFT>Argtable provides a set of arg_<I>xxx</I> structs, one
for each type of argument (literal, integer, double, string,
filename, etc) that it supports and each struct is capable of
handling multiple occurrences of that argument on the command line.
Furthermore, each option can be given alternative short option (ie:
-c) or long option (ie: --scalar) forms that can be used
interchangeably. It fact, each option can even take multiple
alternative short or long options, or both. Options can also be
defined to have no option tag at all (ie: <file>) in which case
they are identifed by their position on the command line (tagged
options can appear anywhere on the command line).</P>
<P ALIGN=LEFT>To define a program's the command line options the user
creates an arg_<I>xxx</I> struct for each type of argument required
and collates them into an array which we call the argument table. The
order of the structs in the argument table defines the order in which
the command line options are expected, although the parsing order
really only matters for untagged options. The argument table itself
is just an array of void pointers, and by convention each arg_<I>xxx</I>
struct has a known arg_hdr struct as its first entry that the
argtable functions use to identify the structure.
</P>
<P ALIGN=LEFT>By way of example, let us consider the arg_int struct
which is used for command line options taking integer arguments, as
in –scalar=7.</P>
<P ALIGN=LEFT STYLE="margin-left: 1.91cm; text-indent: -0.99cm"><FONT FACE="Courier, monospace">struct
arg_int<BR>{<BR>struct arg_hdr hdr;<BR>int count;<BR>int *ival;<BR>};</FONT></P>
<P ALIGN=LEFT>The struct's first data member, <FONT FACE="Courier, monospace">hdr</FONT>,
holds the “private” data used by the argtable library
functions. It contains things like the argument's tag string and so
on. Direct access to this data is openly permitted, but it rarely
necessary to do so. The <FONT FACE="Courier, monospace">ival</FONT>
member variable points to an array of integers that hold the values
extracted from the command line and <FONT FACE="Courier, monospace">count</FONT>
gives the number of values held in the array. The storage for the
<FONT FACE="Courier, monospace">ival</FONT> array is allocated when
the arg_int is constructed. This must done with one of the dedicated
arg_int constructor functions:</P>
<P ALIGN=LEFT STYLE="margin-left: 7.21cm; text-indent: -6.22cm"><FONT FACE="Courier, monospace">struct
arg_int* arg_int0(const char* shortopts,<BR>const char*
longopts,<BR>const char* datatype,<BR>const char* glossary);</FONT></P>
<P ALIGN=LEFT STYLE="margin-left: 7.19cm; text-indent: -6.17cm"><FONT FACE="Courier, monospace">struct
arg_int* arg_int1(const char* shortopts,<BR>const char*
longopts,<BR>const char* datatype,<BR>const char *glossary);</FONT></P>
<P STYLE="margin-left: 7.24cm; text-indent: -6.17cm"><FONT FACE="Courier, monospace">struct
arg_int* arg_intn(const char* shortopts,<BR>const char*
longopts,<BR>const char *datatype,<BR>int mincount,<BR>int
maxcount,<BR>const char *glossary);</FONT></P>
<P ALIGN=LEFT>All the argtable constructor functions work in the same
manner; they allocate a block of memory that contains an arg_<I>xxx</I>
struct at its head followed by storage for the local data for that
structure, in this case the contents of the <FONT FACE="Courier, monospace">ival</FONT>
array. For this reason, you should never manually instantiate any
arg_<I>xxx</I> struct yourself. Always use the constructor functions
provided to allocate the structure and deallocate it using <FONT FACE="Courier, monospace">free</FONT>
when you are finished.</P>
<P ALIGN=LEFT>The three forms of the <FONT FACE="Courier, monospace">arg_int</FONT>
constructors represent the three most common uses of command line
arguments: <FONT FACE="Courier, monospace">arg_int0</FONT> is for
options that appear zero-or-once on the command line, <FONT FACE="Courier, monospace">arg_int1</FONT>
is for options that appear exactly once on the command line, and
<FONT FACE="Courier, monospace">arg_intn</FONT> is for options that
appear any number of times (within a given range). Thus <FONT FACE="Courier, monospace">arg_int0</FONT>
and <FONT FACE="Courier, monospace">arg_int1</FONT> will both
allocate one element to the <FONT FACE="Courier, monospace">ival</FONT>
array of the resulting structure, whereas <FONT FACE="Courier, monospace">arg_intn</FONT>
will allocate sufficient space for <FONT FACE="Courier, monospace">ival</FONT>
to store up to <FONT FACE="Courier, monospace">maxcount</FONT>
elements. The former are just specialised forms of the latter and are
provided for convenience. Notice the arg_<I>xxx</I>0, arg_<I>xxx</I>1,
and arg_<I>xxx</I>n naming convention applies likewise to all
argtable constructor functions.</P>
<P ALIGN=LEFT>Continuing with our <FONT FACE="Courier, monospace">arg_int</FONT>
example, the following code fragment will construct a integer type
option of the form <FONT FACE="Courier, monospace">--scalar=<n></FONT>
that must appear on the command line between 3 and 5 times inclusive.
</P>
<P ALIGN=LEFT STYLE="margin-left: 1.02cm; margin-right: 0.84cm"><FONT FACE="Courier, monospace">struct
arg_int *s;<BR>s = arg_intn(NULL,”scalar”,”<n>”,3,5,“foo
value”);</FONT></P>
<P ALIGN=LEFT>Upon completion <FONT FACE="Courier, monospace">s</FONT>
will point to a memory block containing the <FONT FACE="Courier, monospace">arg_int</FONT>
struct followed by the <FONT FACE="Courier, monospace">ival</FONT>
array of 5 elements.</P>
<P ALIGN=LEFT STYLE="margin-left: 1.02cm; margin-bottom: 0cm; line-height: 100%">
<IMG SRC="arg_int.gif" NAME="Graphic1" ALIGN=LEFT WIDTH=488 HEIGHT=442 BORDER=0><BR CLEAR=LEFT><BR>
</P>
<P ALIGN=LEFT>As shown in the diagram, the <FONT FACE="Courier, monospace">s->hdr
</FONT>structure keeps, among other things, references back to the
string parameters of the constructor function. The <FONT FACE="Courier, monospace">s->count</FONT>
variable is initialised to zero as it represents the number of valid
values that are stored in the <FONT FACE="Courier, monospace">s->ival</FONT>
array <U>after</U> parsing the command line. The size of the <FONT FACE="Courier, monospace">s->ival</FONT>
array is instead given by <FONT FACE="Courier, monospace">s->hdr.maxcount</FONT>.</P>
<P ALIGN=LEFT>In this example we omitted a short option form by
passing a NULL shortopts parameter to the constructor function. If
instead we passed shortops as, say, “k”</P>
<P ALIGN=LEFT STYLE="margin-left: 1.02cm; margin-right: 0.84cm"><FONT FACE="Courier, monospace">s
= arg_intn(“k”,<FONT COLOR="#000000">”scalar”,<FONT COLOR="#000000">”<n>”,</FONT>3,5,“foo
value”</FONT>);</FONT></P>
<P ALIGN=LEFT>then the resulting structure would be the same but the
option could be accepted on the command line as either <FONT FACE="Courier, monospace">-k<n></FONT>
or –<FONT FACE="Courier, monospace">scalar=<n></FONT>
equivalently. Indeed, we can go even further and define multiple
alternative forms for both the short and long options. Alternative
short options are given a string of single characters, whereas long
options are given as a comma separated string. For instance,
</P>
<P ALIGN=LEFT STYLE="margin-left: 1.02cm; margin-right: 0.84cm"><FONT FACE="Courier, monospace">s
= arg_intn(“kKx”,”scalar,foo”<FONT COLOR="#000000">,<FONT COLOR="#000000">”<n>”,</FONT>3,5,“foo
value”</FONT>);</FONT></P>
<P ALIGN=LEFT>will accept any of the following alternative forms on
the command line: <FONT FACE="Courier, monospace">-k<n> -K<n>
-x<n> --scalar=<n> --foo=<n></FONT></P>
<DL>
<DD STYLE="margin-left: 0cm; margin-bottom: 0.51cm">Apart from
<FONT FACE="Courier, monospace">arg_int</FONT>, other arg_<I>xxx</I>
structs of interest are:
</DD></DL>
<TABLE WIDTH=745 BORDER=0 CELLPADDING=10 CELLSPACING=0>
<COL WIDTH=332>
<COL WIDTH=373>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm; font-style: normal">
<FONT FACE="Courier, monospace">struct arg_lit<BR>{<BR>struct
arg_hdr hdr;<BR>int count;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm">For literal options (ie: no
arguments).<BR>eg: <FONT FACE="Courier, monospace">-v, --verbose</FONT></DD></DL>
</TD>
</TR>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm; font-style: normal">
<FONT FACE="Courier, monospace">struct arg_dbl<BR>{<BR>struct
arg_hdr hdr;<BR>int count;<BR>double *dval;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm">For options taking real
arguments.<BR>eg: <FONT FACE="Courier, monospace">-x1.0e-6,
--pi=3.1415</FONT></DD></DL>
</TD>
</TR>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm; font-style: normal">
<FONT FACE="Courier, monospace">struct arg_str<BR>{<BR>struct
arg_hdr hdr;<BR>int count;<BR>const char **sval;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm; margin-bottom: 0.51cm">For options
taking string arguments.<BR>eg: <FONT FACE="Courier, monospace">-Dmacro,
--title=”hello world”</FONT></DD><DD STYLE="margin-left: 0cm">
<BR>
</DD></DL>
</TD>
</TR>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm; font-style: normal">
<FONT FACE="Courier, monospace">struct arg_rex<BR>{<BR>struct
arg_hdr hdr;<BR>int count;<BR>const char **sval;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm; margin-bottom: 0.51cm">For options
taking string arguments that must match some regular expression
template.<BR>eg: <FONT FACE="Courier, monospace">-Dmacro,
--title=”hello world”</FONT></DD><DD STYLE="margin-left: 0cm">
<FONT SIZE=1 STYLE="font-size: 8pt">Note: arg_rex is not
supported under Microsoft Windows.</FONT></DD></DL>
</TD>
</TR>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm"><FONT FACE="Courier, monospace">struct
arg_file<BR>{<BR>struct arg_hdr hdr;<BR>int count;<BR>const char
**filename;<BR>const char **basename;<BR>const char
**extension;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm">For options taking filename
arguments.<BR>This option returns not only the filename itself,
but separates out its basename and extension as well.<BR>eg: <FONT FACE="Courier, monospace">-o
myfile, --infile=memo.txt</FONT></DD></DL>
</TD>
</TR>
<TR VALIGN=TOP>
<TD WIDTH=332>
<P STYLE="margin-left: 1.27cm; text-indent: -0.96cm"><FONT FACE="Courier, monospace">struct
arg_date<BR>{<BR>struct arg_hdr hdr;<BR>const char *format;<BR>int
count;<BR>struct tm *tm_val;<BR>};</FONT></P>
</TD>
<TD WIDTH=373>
<DL>
<DD STYLE="margin-left: 0cm; margin-bottom: 0.51cm">For options
taking date/time arguments.<BR>eg: <FONT FACE="Courier, monospace">12/31/04,
-d 1982-11-28,<BR>--time 23:59</FONT></DD><DD STYLE="margin-left: 0cm">
<FONT FACE="Bitstream Vera Sans"><FONT SIZE=1 STYLE="font-size: 8pt">Note:
arg_date is not supported under Microsoft Windows.</FONT></FONT></DD></DL>
</TD>
</TR>
</TABLE>
<DL>
<DD STYLE="margin-left: 0cm; margin-top: 1.5cm; margin-bottom: 0.51cm">
<FONT FACE="Albany, sans-serif"><FONT SIZE=4>The argument table</FONT></FONT></DD></DL>
<P ALIGN=LEFT>
Having constructed our arg_xxx structs we collate them into an
argument table, as in the following example which defines the command
line arguments:<FONT FACE="Courier, monospace">[-a] [-b] [-c]
[--scalar=<n>] [-v|--verbose] [-o myfile] <file> [<file>]</FONT></P>
<P STYLE="margin-left: 0.99cm; margin-bottom: 0cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>struct
arg_lit *a = arg_lit0(“a”, NULL, ”the -a
option”);<BR>struct arg_lit *b = arg_lit0(“b”,
NULL, ”the -b option”);<BR>struct arg_lit *c =
arg_lit0(“c”, NULL, ”the -c option”);<BR>struct
arg_int *scal = arg_int0(NULL, ”scalar”,”<n>”,
”foo value”);<BR>struct arg_lit *verb = arg_lit0(“v”,
”verbose, ”verbose output”);<BR>struct arg_file *o
= arg_file0(“o”, NULL,”myfile”, ”output
file”);<BR>struct arg_file *file =
arg_filen(NULL,NULL,”<file>”,1,2, ”input
files”);<BR>struct arg_end *end = arg_end(20);<BR>void
*argtable[] = {a,b,c,scal,verb,o,file,end};</FONT></FONT></P>
<P STYLE="margin-left: 1.02cm; margin-bottom: 0cm"><BR>
</P>
<P>The “-a”, “-b”, “-c” and
“-v|--verbose” options do not take argument values so we
use <FONT FACE="Courier, monospace">arg_lit</FONT> structs for them.
We use the <FONT FACE="Courier, monospace">arg_lit0</FONT> form of
the constructor function because these particular options only appear
on the command line once or not at all.</P>
<P>The “--scalar=<n>” option takes an integer
argument so its uses an <FONT FACE="Courier, monospace">arg_int</FONT>
struct. It too appears either once or not at all so we use the
<FONT FACE="Courier, monospace">arg_int0</FONT> constructor function.</P>
<P>The “-o myfile” and “<file>” options
both refer to filenames so we use the <FONT FACE="Courier, monospace">arg_file</FONT>
struct for them. We use the <FONT FACE="Courier, monospace">arg_file0</FONT>
constructor function for the former because it appears either once or
not at all, but the latter must appear either once or twice so we use
the more general <FONT FACE="Courier, monospace">arg_filen</FONT>
constructor function for that. Notice that it is an untagged option
as it does not take either short or long option strings.</P>
<P>The <FONT FACE="Courier, monospace">arg_end</FONT> struct is a
special one as it doesn't represent any command line option.
Primarily it marks the end of the argtable array, but it also stores
any parser errors encountered when processing the command line
arguments. The integer parameter passed to the <FONT FACE="Courier, monospace">arg_end</FONT>
constructor is the maximum number of errors that it will store, in
this case 20, any further errors are discarded and replaced with the
single error message “too many errors”.</P>
<P>We will see how to use <FONT FACE="Courier, monospace">arg_end</FONT>
in error reporting soon but first we must ensure that all of the
argument table entries were successfully allocated by their
constructor functions. If they were'nt then there will be NULL
entries in the argtable array which will cause trouble. We can use
the arg_nullcheck function to check argtable for NULL entries in one
step. It returns non-zero if any NULL entries were encountered up
until the end of the table as marked by the <FONT FACE="Courier, monospace">arg_end</FONT>
structure.</P>
<P STYLE="margin-left: 1.98cm; text-indent: -1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>if
(arg_nullcheck(argtable) != 0)<BR>printf("error: insufficient
memory\n");</FONT></FONT></P>
<P>Presuming that went well, we may now initate any default values we
wish to assign our optional arguments. We simply write our desired
values directly into the <FONT FACE="Courier, monospace">arg_<I>xxx</I></FONT>
structs knowing that argtable will only overwrite them if valid
command line values are given in their place. Here we set the default
values of 3 and “-” for the <FONT FACE="Courier, monospace">repeat</FONT>
and <FONT FACE="Courier, monospace">outfile</FONT> arguments
respectively.
</P>
<P STYLE="margin-left: 1.07cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>repeat->ival[0]=3;<BR>outfile->filename[0]=”-”;</FONT></FONT></P>
<P>Argtable does not require we initialise any default values, it is
simply more convenient for our program if we pre-load defaults prior
to parsing rather than retro-fit defaults to missing values later.
However, you may prefer the latter.</P>
<P STYLE="margin-top: 1.47cm; margin-bottom: 0.51cm"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Parsing
the command line</FONT></FONT></P>
<P>Now our argument table is complete, we can use it to parse the
command line arguments. We use the <FONT FACE="Courier, monospace">arg_parse</FONT>
function to do that, and it returns the number of parse errors it
encountered.</P>
<P STYLE="margin-left: 2.01cm; text-indent: -0.99cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>nerrors
= arg_parse(argc,argv,argtable);</FONT></FONT></P>
<P>If there were no errors then we have successfully parsed the
command line and we can proceed with our main processing task, using
the values to be found in our program's arg_<I>xxx</I> structs.</P>
<P STYLE="margin-left: 2.01cm; text-indent: -0.99cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>if
(nerrors==0)<BR>{<BR>int i;<BR>printf(“-a = %d\n”,
a->count);<BR>printf(“-b = %d\n”,
b->count);<BR>printf(“-c = %d\n”,
c->count);<BR>printf(“--verbose = %d\n”,
verb->count);<BR>if (scal->count >
0)<BR>printf(“--scalar=%d\n”,scal->ival[0]);<BR>if
(o->count > 0)<BR>printf(“-o %s\n”,o->filename[0]);<BR>for
(i=0; i<file->count;
i++)<BR>printf(“file[%d]=%s\n”,i,file->filename[i]);<BR>};</FONT></FONT></P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Error
processing</FONT></FONT></P>
<P>If the <FONT FACE="Courier, monospace">arg_parse</FONT> function
reported errors then we need to display them as <FONT FACE="Courier, monospace">arg_parse</FONT>
does not do so itself. As mentioned earlier, the <FONT FACE="Courier, monospace">arg_parse</FONT>
function stores the errors it encounters in the argument table's
<FONT FACE="Courier, monospace">arg_end</FONT> struct. We dont need
to know the internal details of the <FONT FACE="Courier, monospace">arg_end</FONT>
struct, we simply call the <FONT FACE="Courier, monospace">arg_print_errors</FONT>
function to print those errors in the order they were encountered.
</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>void
arg_print_errors(FILE* fp, struct arg_end* end, const char*
progname);</FONT></FONT></P>
<P>We pass the function a pointer to the argument table's <FONT FACE="Courier, monospace">arg_end</FONT>
struct as well as the name of the program which is prependend to each
error message. The program name can be NULL if not required.</P>
<P STYLE="margin-left: 2.01cm; text-indent: -1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>If
(nerrors > 0)<BR>arg_print_errors(stdout,end,”myprog”);</FONT></FONT></P>
<P STYLE="font-weight: medium">This example illustrates the results
of invoking our example program with incorrect command line options:</P>
<P STYLE="margin-left: 1.07cm"><FONT FACE="Courier, monospace"><FONT SIZE=2><B>$
./myprog -x -y -z --scalar=hello --verby</B><BR>myprog: invalid
option "-x"<BR>myprog: invalid option "-y"<BR>myprog:
invalid option "-z"<BR>myprog: invalid argument "hello"
to option --scalar=<n><BR>myprog: invalid option
"--verby"<BR>myprog: missing option <file></FONT></FONT></P>
<P>The reason <FONT FACE="Courier, monospace">arg_parse</FONT>
function doesnt print error messages is so it can be called multiple
times to parse the command line with alternative argument tables
without having extraneous error messages displayed prematurely. Thus
we may define separate argument tables for those programs that have
mutually exclusive sets of command line options, trying each argument
table in turn until we find a successful candidate. Should all
argument tables fail to satisfy then we can choose to print the error
messages from all of them, or perhaps only show the errors form the
one that matched the closest. In any event, we control which messages
are displayed.</P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Displaying
the option syntax</FONT></FONT></P>
<P>If you want your program to display on-line help you can use the
<FONT FACE="Courier, monospace">arg_print_syntax</FONT> function to
display the exact command line syntax as derived from an argument
table. There are actually two forms of the function:</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>void
arg_print_syntax(FILE *fp, void **argtable, const char *suffix);<BR>void
arg_print_syntaxv(FILE *fp, void **argtable, const char *suffix);</FONT></FONT></P>
<P>The latter displays a more verbose form of output, and is
distinguished by the “v” at the end of the function name.
Both functions display the syntax for an entire argument table, with
the <FONT FACE="Courier, monospace">suffix</FONT> parameter provided
as a convenience for appending newline characters or any other string
onto the end of the output. In the verbose form, each argument table
entry displays its alternative short and long options separated by
the “|” character followed by its datatype string. For
instance,
</P>
<P ALIGN=LEFT STYLE="margin-left: 1.02cm; margin-right: 0.84cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>arg_int0(“kKx”,”scalar,foo”<FONT COLOR="#000000">,<FONT COLOR="#000000">”<n>”,</FONT>“foo
value”</FONT>);</FONT></FONT></P>
<P>will be displayed in verbose form as
“<FONT FACE="Courier, monospace">[-k|-K|-x|--scalar|--foo=<n>]</FONT>”.
Whereas the standard form abbreviates the output by displaying only
the first option of each argument table entry, as in “<FONT FACE="Courier, monospace">[-k
<n>]</FONT>”. The standard form also concatentates all
short options in the argument table into a single option string at
the head of the display in standard GNU style (eg: -a -b -c is
displayed as -abc). The argument table from our previous example
would thus be displayed in standard form as:</P>
<P STYLE="margin-left: 0.99cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>[-abcv]
[--scalar=<n>] [-o myfile] <file> [<file>]</FONT></FONT></P>
<P>and in verbose form as:</P>
<P STYLE="margin-left: 0.99cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>[-a]
[-b] [-c] [--scalar=<n>] [-o myfile] [-v|--verbose] <file>
[<file>]</FONT></FONT></P>
<P>Notice that optional entries are automatically enclosed in square
brackets whereas mandatory arguments are not. Futhermore arguments
that accept multiple instances are displayed once per instance, as in
“<FONT FACE="Courier, monospace"><file> [<file>]</FONT>”.
This occurs up to a maximum of three instances after which the
repetition is replaced by an elipisis, as in “<FONT FACE="Courier, monospace">[<file>]...</FONT>”.</P>
<P>The <FONT FACE="Courier, monospace">arg_print_syntax</FONT>
functions safely ignore NULL short and long option strings, whereas a
NULL datatype string is automatically replaced by the default
datatype for that arg_<I>xxx</I> struct. The default datatype can be
suppressed by using an empty datatype string instead of a NULL.</P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Displaying
the option glossary</FONT></FONT></P>
<P>The individual entries of the argument table can be displayed in a
glossary layout by the arg_print_glossary function. It displays the
full syntax of each argument table entry followed by each table
entry's glossary string – the glossary string is the last
parameter passed to the arg_<I>xxx</I> constructor functions. Table
entries with NULL glossary strings are not displayed.</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>void
arg_print_glossary(FILE *fp, void **argtable, const char *format);</FONT></FONT></P>
<P>The <FONT FACE="Courier, monospace">format</FONT> string passed to
the arg_print_glossary function is actually a printf style format
string. It should contain exactly two “%s” format
parameters, the first is used to control the printf format of the
option's syntax string and the second is for the argument's glossary
string. A typical format string would be <FONT FACE="Courier, monospace">"
%-25s %s\n". </FONT>The format string allows fine control over
the display formatting but demands dilligence as any unexpected
parameters in it will cause unpredictable results. Here is the
results of calling <FONT FACE="Courier, monospace">arg_print_glossary</FONT>
on our earlier example argument table:</P>
<DL>
<DD>
<TABLE WIDTH=617 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<COL WIDTH=162>
<COL WIDTH=455>
<TR VALIGN=TOP>
<TD WIDTH=162>
<P><FONT FACE="Courier, monospace"><FONT SIZE=2>-a<BR>-b<BR>-c<BR>--scalar=<n><BR>-v,
--verbose<BR>-o myfile<BR><file></FONT></FONT></P>
</TD>
<TD WIDTH=455>
<P><FONT FACE="Courier, monospace"><FONT SIZE=2>the -a option<BR>the
-b option<BR>the -c option<BR>foo value<BR>verbose option<BR>output
file<BR>input files</FONT></FONT></P>
</TD>
</TR>
</TABLE>
</DL>
<P STYLE="margin-top: 0.51cm">Sometimes you will wish to add extra
lines of text to the glossary, or even put your own text into the
syntax string generated by <FONT FACE="Courier, monospace">arg_print_syntax</FONT>.
You can add newline characters to your argument table strings if you
wish, but it soon gets ugly. A better way is to add <FONT FACE="Courier, monospace">arg_rem</FONT>
structs to your argument table. They are dummy argument table entries
in the sense that they do not alter the argument parsing but their
datatype and glossary strings do appear in the output generated by
the <FONT FACE="Courier, monospace">arg_print_syntax</FONT> and
<FONT FACE="Courier, monospace">arg_print_glossary</FONT> functions.
The name <FONT FACE="Courier, monospace">arg_rem</FONT> is for
“remark” and is inspired by the REM statement used in the
BASIC language.</P>
<P STYLE="margin-top: 1.5cm"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Cleaning
up</FONT></FONT></P>
<P>At the end of our program we need to deallocate the memory
allocated to each of our arg_<I>xxx</I> structs. We could do that by
calling <FONT FACE="Courier, monospace">free</FONT> on each of them
individually, but the <FONT FACE="Courier, monospace">arg_freetable</FONT>
function can do it for us more conveniently.
</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace"><FONT SIZE=2>arg_freetable(argtable,sizeof(argtable)/sizeof(argtable[0]));</FONT></FONT></P>
<P>It will step through an argument table and call <FONT FACE="Courier, monospace">free</FONT>
on each of its elements on our behalf. Note the second parameter,
<FONT FACE="Courier, monospace">sizeof(argtable)/sizeof(argtable[0])</FONT>,
merely represents the number of elements in our argtable array. Upon
completion of this function, all of the argtable array entries will
have been set to NULL.</P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Hint:
declaring global arg_<I>xxx</I> variables</FONT></FONT></P>
<P>ANSI C wont allow the the arg_<I>xxx</I> constructor functions to
be placed in the global namespace, so if you wish to make your
arg_<I>xxx</I> structs global you must initialiase them elsewhere.
Here's a programming trick for using global arg_<I>xxx</I> structs
while stull declaring your argtable statically.</P>
<P STYLE="margin-left: 1.27cm; margin-bottom: 0cm; line-height: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2>#include <argtable2.h></FONT></FONT></P>
<P STYLE="margin-left: 1.27cm; margin-bottom: 0cm; line-height: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2><I>/* global arg_xxx
structs */</I><BR>struct arg_lit *a, *b, *c, *verb;<BR>struct arg_int
*scal;<BR>struct arg_file *o, *file;<BR>struct arg_end *end;</FONT></FONT></P>
<P STYLE="margin-left: 2.54cm; text-indent: -1.27cm; margin-bottom: 0cm; line-height: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2>int main(int argc, char
**argv)<BR>{<BR><I>/* the global arg_xxx structs are initialised
within the argtable */</I></FONT></FONT></P>
<P STYLE="margin-left: 3.81cm; text-indent: -1.27cm; margin-bottom: 0cm; line-height: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2>void *argtable[] ={<BR>a
= arg_lit0(“a”, NULL, ”the -a option”),<BR>b
= arg_lit0(“b”, NULL, ”the -b option”),<BR>c
= arg_lit0(“c”, NULL, ”the -c option”),<BR>scal
= arg_int0(NULL, ”scalar”,”<n>”, ”foo
value”),<BR>verb = arg_lit0(“v”, ”verbose,
”verbose output”),<BR>o = arg_file0(“o”,
NULL,”myfile”, ”output file”),<BR>file =
arg_filen(NULL,NULL,”<file>”,1,2, ”input
files”),<BR>end = arg_end(20),<BR>};</FONT></FONT></P>
<P STYLE="margin-left: 2.54cm; margin-bottom: 0cm; line-height: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2>...<BR>return 0;<BR>};</FONT></FONT></P>
<P>See the <B>ls.c</B> program included in the argtable distribution
for an example of using this declaration style.</P>
<P STYLE="margin-top: 1.5cm"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Compiling
programs with argtable library</FONT></FONT></P>
<P>All source code that uses the argtable library must include the
“argtable2.h” header function and link against the
argtable2 library. A typical unix makefile for doing this would be:</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace">CC =
gcc<BR>CFLAGS = -I/usr/local/include -Wall -ansi<BR>LDFLAGS =
-L/usr/local/lib<BR>LDLIBS = -largtable2</FONT></P>
<P STYLE="margin-left: 1.98cm; text-indent: -1.02cm"><FONT FACE="Courier, monospace">myprog.o:
myprog.c<BR>$(CC) -c $(CFLAGS) -o myprog.o myprog.c</FONT></P>
<P STYLE="margin-left: 1.98cm; text-indent: -0.99cm"><FONT FACE="Courier, monospace">myprog:
myprog.o<BR>$(CC) $(LDFLAGS) myprog.o -o myprog $(LDLIBS)</FONT></P>
<P>Though the last two make rules are redundant as the default make
rules suffice, so we could simplify the entire Makefile to just:</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace">CC =
gcc<BR>CFLAGS = -I/usr/local/include -Wall -ansi<BR>LDFLAGS =
-L/usr/local/lib<BR>LDLIBS = -largtable2</FONT></P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace">myprog:
myprog.o</FONT></P>
<P>If you would prefer to statically link all the libraries to your
application then add the “-static” switch to CFLAGS.</P>
<P STYLE="margin-left: 1.02cm"><FONT FACE="Courier, monospace">CFLAGS
= -I/usr/local/include -Wall -ansi -static</FONT></P>
<P>If you wish to statically link only the argtable2 library while
linking all other libraries dynamically, then treat the argtable
static library as if it were an object file. The linker will
statically link the object files within the library to your
application.</P>
<P STYLE="margin-left: 1.98cm; text-indent: -0.99cm"><FONT FACE="Courier, monospace">myprog:
myprog.o<BR>$(CC) myprog.o /usr/local/lib/libargtable2.a -o myprog</FONT></P>
<P STYLE="margin-top: 1.5cm; page-break-after: avoid"><FONT FACE="Albany, sans-serif"><FONT SIZE=4>Example
code</FONT></FONT></P>
<P>The argtable distribution comes with example programs that
implement complete GNU compatable command line options for several
common unix commands. See the argtable-2.x/example/ directory for the
source code of the following programs:</P>
<DL>
<DD><FONT FACE="Courier, monospace"><FONT SIZE=2>echo [-neE]
[--help] [--version] [STRING]...</FONT></FONT></DD><DD>
<BR>
</DD><DD>
<FONT FACE="Courier, monospace"><FONT SIZE=2>ls
[-aAbBcCdDfFgGhHiklLmnNopqQrRsStuUvxX1] [--author]
[--block-size=SIZE] [--color=[WHEN]] [--format=WORD] [--full-time]
[--si] [--dereference-command-line-symlink-to-dir]
[--indicator-style=WORD] [-I PATTERN] [--show-control-chars]
[--quoting-style=WORD] [--sort=WORD] [--time=WORD]
[--time-style=STYLE] [-T COLS] [-w COLS] [--help] [--version]
[FILE]...</FONT></FONT></DD><DD>
<BR>
</DD><DD>
<FONT FACE="Courier, monospace"><FONT SIZE=2>mv [-bfiuv]
[--backup=[CONTROL]] [--reply={yes,no,query}]
[--strip-trailing-slashes] [-S SUFFIX]
[--target-directory=DIRECTORY] [--help] [--version] SOURCE
[SOURCE]... DEST|DIRECTORY</FONT></FONT></DD><DD>
<BR>
</DD><DD>
<FONT FACE="Courier, monospace"><FONT SIZE=2>rm [-dfirv] [--help]
[--version] <file> [<file>]...</FONT></FONT></DD><DD>
<BR>
</DD><DD STYLE="margin-bottom: 0.51cm">
<FONT FACE="Courier, monospace"><FONT SIZE=2>uname [-asnrvmpio]
[--help] [--version]</FONT></FONT></DD></DL>
<P STYLE="margin-top: 0.43cm; page-break-after: avoid">
<BR><BR>
</P>
<P><BR><BR>
</P>
</BODY>
</HTML>
|