=================================================================== RCS file: /cvs/mandoc/Attic/argv.c,v retrieving revision 1.39 retrieving revision 1.52 diff -u -p -r1.39 -r1.52 --- mandoc/Attic/argv.c 2009/03/08 11:41:22 1.39 +++ mandoc/Attic/argv.c 2009/03/16 22:19:19 1.52 @@ -1,4 +1,4 @@ -/* $Id: argv.c,v 1.39 2009/03/08 11:41:22 kristaps Exp $ */ +/* $Id: argv.c,v 1.52 2009/03/16 22:19:19 kristaps Exp $ */ /* * Copyright (c) 2008 Kristaps Dzonsons * @@ -16,6 +16,8 @@ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR * PERFORMANCE OF THIS SOFTWARE. */ +#include + #include #include #include @@ -43,6 +45,8 @@ #define ARGV_MULTI (1 << 2) #define ARGV_OPT_SINGLE (1 << 3) +#define MULTI_STEP 5 + enum mwarn { WQUOTPARM, WARGVPARM, @@ -58,7 +62,7 @@ enum merr { static int argv_a2arg(int, const char *); static int args(struct mdoc *, int, int *, char *, int, char **); -static int argv(struct mdoc *, int, int, +static int argv(struct mdoc *, int, struct mdoc_argv *, int *, char *); static int argv_single(struct mdoc *, int, struct mdoc_argv *, int *, char *); @@ -92,10 +96,11 @@ static int mdoc_argvflags[MDOC_ARG_MAX] = { ARGV_MULTI, /* MDOC_Column */ ARGV_SINGLE, /* MDOC_Width */ ARGV_NONE, /* MDOC_Compact */ - ARGV_SINGLE, /* MDOC_Std */ + ARGV_OPT_SINGLE, /* MDOC_Std */ ARGV_NONE, /* MDOC_Filled */ ARGV_NONE, /* MDOC_Words */ ARGV_NONE, /* MDOC_Emphasis */ + ARGV_NONE, /* MDOC_Symbolic */ ARGV_NONE /* MDOC_Symbolic */ }; @@ -207,6 +212,16 @@ static int mdoc_argflags[MDOC_MAX] = { 0, /* Fr */ 0, /* Ud */ 0, /* Lb */ + 0, /* Ap */ + ARGS_DELIM, /* Lp */ + ARGS_DELIM | ARGS_QUOTED, /* Lk */ + ARGS_DELIM | ARGS_QUOTED, /* Mt */ + ARGS_DELIM, /* Brq */ + 0, /* Bro */ + ARGS_DELIM, /* Brc */ + ARGS_QUOTED, /* %C */ + 0, /* Es */ + 0, /* En */ }; @@ -220,18 +235,20 @@ mdoc_argv(struct mdoc *mdoc, int line, int tok, struct mdoc_arg **v, int *pos, char *buf) { int i; - char *p; + char *p, sv; struct mdoc_argv tmp; struct mdoc_arg *arg; if (0 == buf[*pos]) return(ARGV_EOLN); - assert( ! isspace((u_char)buf[*pos])); + assert(' ' != buf[*pos]); - if ('-' != buf[*pos]) + if ('-' != buf[*pos] || ARGS_ARGVLIKE & mdoc_argflags[tok]) return(ARGV_WORD); + /* Parse through to the first unescaped space. */ + i = *pos; p = &buf[++(*pos)]; @@ -239,47 +256,49 @@ mdoc_argv(struct mdoc *mdoc, int line, int tok, /* LINTED */ while (buf[*pos]) { - if (isspace((u_char)buf[*pos])) + if (' ' == buf[*pos]) if ('\\' != buf[*pos - 1]) break; (*pos)++; } - if (buf[*pos]) + /* XXX - save zeroed byte, if not an argument. */ + + sv = 0; + if (buf[*pos]) { + sv = buf[*pos]; buf[(*pos)++] = 0; + } (void)memset(&tmp, 0, sizeof(struct mdoc_argv)); tmp.line = line; tmp.pos = *pos; - /* - * We now parse out the per-macro arguments. XXX - this can be - * made much cleaner using per-argument tables. See argv_a2arg - * for details. - */ + /* See if our token accepts the argument. */ if (MDOC_ARG_MAX == (tmp.arg = argv_a2arg(tok, p))) { + /* XXX - restore saved zeroed byte. */ + if (sv) + buf[*pos - 1] = sv; if ( ! pwarn(mdoc, line, i, WARGVPARM)) return(ARGV_ERROR); return(ARGV_WORD); } - while (buf[*pos] && isspace((u_char)buf[*pos])) + while (buf[*pos] && ' ' == buf[*pos]) (*pos)++; - /* FIXME: whitespace if no value. */ - - if ( ! argv(mdoc, tok, line, &tmp, pos, buf)) + if ( ! argv(mdoc, line, &tmp, pos, buf)) return(ARGV_ERROR); if (NULL == (arg = *v)) { - v = xcalloc(1, sizeof(struct mdoc_arg)); + *v = xcalloc(1, sizeof(struct mdoc_arg)); arg = *v; - } + } arg->argc++; arg->argv = xrealloc(arg->argv, arg->argc * - sizeof(struct mdoc_arg)); + sizeof(struct mdoc_argv)); (void)memcpy(&arg->argv[(int)arg->argc - 1], &tmp, sizeof(struct mdoc_argv)); @@ -293,9 +312,17 @@ mdoc_argv_free(struct mdoc_arg *p) { int i, j; - if (p->refcnt && --(p->refcnt) > 0) + if (NULL == p) return; + if (p->refcnt) { + --(p->refcnt); + if (p->refcnt) + return; + } + + assert(p->argc); + /* LINTED */ for (i = 0; i < (int)p->argc; i++) { if (0 == p->argv[i].sz) @@ -303,11 +330,11 @@ mdoc_argv_free(struct mdoc_arg *p) /* LINTED */ for (j = 0; j < (int)p->argv[i].sz; j++) free(p->argv[i].value[j]); + free(p->argv[i].value); } - if (p->argc) - free(p->argv); + free(p->argv); free(p); } @@ -417,8 +444,6 @@ mdoc_args(struct mdoc *mdoc, int line, break; } - /* Continue parsing the arguments themselves... */ - return(args(mdoc, line, pos, buf, fl, v)); } @@ -455,10 +480,10 @@ args(struct mdoc *mdoc, int line, break; i++; /* There must be at least one space... */ - if (0 == buf[i] || ! isspace((u_char)buf[i])) + if (0 == buf[i] || ' ' != buf[i]) break; i++; - while (buf[i] && isspace((u_char)buf[i])) + while (buf[i] && ' ' == buf[i]) i++; } if (0 == buf[i]) { @@ -474,24 +499,16 @@ args(struct mdoc *mdoc, int line, /* * Thar be dragons here! If we're tab-separated, search - * ahead for either a tab or the `Ta' macro. If a tab - * is detected, it mustn't be escaped; if a `Ta' is - * detected, it must be space-buffered before and after. - * If either of these hold true, then prune out the + * ahead for either a tab or the `Ta' macro. + * If a `Ta' is detected, it must be space-buffered before and + * after. If either of these hold true, then prune out the * extra spaces and call it an argument. */ if (ARGS_TABSEP & fl) { /* Scan ahead to unescaped tab. */ - for (p = *v; ; p++) { - if (NULL == (p = strchr(p, '\t'))) - break; - if (p == *v) - break; - if ('\\' != *(p - 1)) - break; - } + p = strchr(*v, '\t'); /* Scan ahead to unescaped `Ta'. */ @@ -547,7 +564,7 @@ args(struct mdoc *mdoc, int line, return(0); if (p) - return(ARGS_WORD); + return(ARGS_PHRASE); /* Configure the eoln case, too. */ @@ -559,14 +576,14 @@ args(struct mdoc *mdoc, int line, return(0); *pos += (int)(p - *v); - return(ARGS_WORD); + return(ARGS_PHRASE); } - + /* Do non-tabsep look-ahead here. */ if ( ! (ARGS_TABSEP & fl)) while (buf[*pos]) { - if (isspace((u_char)buf[*pos])) + if (' ' == buf[*pos]) if ('\\' != buf[*pos - 1]) break; (*pos)++; @@ -581,7 +598,7 @@ args(struct mdoc *mdoc, int line, return(ARGS_WORD); if ( ! (ARGS_TABSEP & fl)) - while (buf[*pos] && isspace((u_char)buf[*pos])) + while (buf[*pos] && ' ' == buf[*pos]) (*pos)++; if (buf[*pos]) @@ -613,7 +630,7 @@ args(struct mdoc *mdoc, int line, if (0 == buf[*pos]) return(ARGS_QWORD); - while (buf[*pos] && isspace((u_char)buf[*pos])) + while (buf[*pos] && ' ' == buf[*pos]) (*pos)++; if (buf[*pos]) @@ -659,6 +676,8 @@ argv_a2arg(int tok, const char *argv) return(MDOC_File); else if (xstrcmp(argv, "offset")) return(MDOC_Offset); + else if (xstrcmp(argv, "compact")) + return(MDOC_Compact); break; case (MDOC_Bf): @@ -704,6 +723,8 @@ argv_a2arg(int tok, const char *argv) return(MDOC_Offset); else if (xstrcmp(argv, "compact")) return(MDOC_Compact); + else if (xstrcmp(argv, "nested")) + return(MDOC_Nested); break; case (MDOC_Rv): @@ -738,11 +759,11 @@ argv_multi(struct mdoc *mdoc, int line, else if (ARGS_EOLN == c) break; - if (0 == v->sz % 5) + if (0 == v->sz % MULTI_STEP) v->value = xrealloc(v->value, - (v->sz + 5) * sizeof(char *)); + (v->sz + MULTI_STEP) * sizeof(char *)); - v->value[(int)v->sz] = p; + v->value[(int)v->sz] = xstrdup(p); } if (v->sz) @@ -770,7 +791,7 @@ argv_opt_single(struct mdoc *mdoc, int line, v->sz = 1; v->value = xcalloc(1, sizeof(char *)); - v->value[0] = p; + v->value[0] = xstrdup(p); return(1); } @@ -795,7 +816,7 @@ argv_single(struct mdoc *mdoc, int line, v->sz = 1; v->value = xcalloc(1, sizeof(char *)); - v->value[0] = p; + v->value[0] = xstrdup(p); return(1); } @@ -806,28 +827,14 @@ argv_single(struct mdoc *mdoc, int line, * multiple parameters. */ static int -argv(struct mdoc *mdoc, int tok, int line, +argv(struct mdoc *mdoc, int line, struct mdoc_argv *v, int *pos, char *buf) { - int fl; v->sz = 0; v->value = NULL; - fl = mdoc_argvflags[v->arg]; - /* - * Override the default per-argument value. - */ - - switch (tok) { - case (MDOC_Ex): - fl = ARGV_OPT_SINGLE; - break; - default: - break; - } - - switch (fl) { + switch (mdoc_argvflags[v->arg]) { case (ARGV_SINGLE): return(argv_single(mdoc, line, v, pos, buf)); case (ARGV_MULTI):