diff options
author | Kristaps Dzonsons <kristaps@bsd.lv> | 2009-03-08 11:41:22 +0000 |
---|---|---|
committer | Kristaps Dzonsons <kristaps@bsd.lv> | 2009-03-08 11:41:22 +0000 |
commit | 056d0c68fbd56e3ce25fe3ad8c58ade0ab30a1b4 (patch) | |
tree | a2b977b6c334b960dc3df02e4f402667e9314030 | |
parent | 7024ee7ca35a8033dee9ec0936cd2f3e0251aca2 (diff) | |
download | mandoc-056d0c68fbd56e3ce25fe3ad8c58ade0ab30a1b4.tar.gz |
BROKEN CHECK-IN. Wide-ranging updates.
Removed libmdoc internal argv limit (all dynamic).
Removed "data" union in mdoc_node (to be selectively re-added).
-rw-r--r-- | action.c | 191 | ||||
-rw-r--r-- | argv.c | 314 | ||||
-rw-r--r-- | hash.c | 5 | ||||
-rw-r--r-- | macro.c | 144 | ||||
-rw-r--r-- | mdoc.c | 304 | ||||
-rw-r--r-- | mdoc.h | 52 | ||||
-rw-r--r-- | mmain.c | 33 | ||||
-rw-r--r-- | private.h | 20 | ||||
-rw-r--r-- | validate.c | 623 | ||||
-rw-r--r-- | xstd.c | 2 |
10 files changed, 837 insertions, 851 deletions
@@ -33,12 +33,24 @@ * children have been filled in (post-fix order). */ +enum merr { + ENOWIDTH +}; + +enum mwarn { + WBADSEC, + WNOWIDTH, + WBADDATE +}; + struct actions { int (*post)(struct mdoc *); }; -/* Per-macro action routines. */ - +static int nwarn(struct mdoc *, + const struct mdoc_node *, enum mwarn); +static int nerr(struct mdoc *, + const struct mdoc_node *, enum merr); static int post_ar(struct mdoc *); static int post_bl(struct mdoc *); static int post_bl_width(struct mdoc *); @@ -49,11 +61,8 @@ static int post_nm(struct mdoc *); static int post_os(struct mdoc *); static int post_sh(struct mdoc *); static int post_ex(struct mdoc *); - static int post_prologue(struct mdoc *); -/* Array of macro action routines. */ - const struct actions mdoc_actions[MDOC_MAX] = { { NULL }, /* \" */ { post_dd }, /* Dd */ @@ -165,6 +174,53 @@ const struct actions mdoc_actions[MDOC_MAX] = { }; +#define merr(m, t) nerr((m), (m)->last, (t)) +static int +nerr(struct mdoc *m, const struct mdoc_node *n, enum merr type) +{ + char *p; + + p = NULL; + + switch (type) { + case (ENOWIDTH): + p = "missing width argument"; + break; + } + + assert(p); + return(mdoc_nerr(m, n, p)); +} + + +#define mwarn(m, t) nwarn((m), (m)->last, (t)) +static int +nwarn(struct mdoc *m, const struct mdoc_node *n, enum mwarn type) +{ + char *p; + int c; + + p = NULL; + c = WARN_SYNTAX; + + switch (type) { + case (WBADSEC): + p = "inappropriate document section in manual section"; + c = WARN_COMPAT; + break; + case (WNOWIDTH): + p = "cannot determine default width"; + break; + case (WBADDATE): + p = "malformed date syntax"; + break; + } + + assert(p); + return(mdoc_nwarn(m, n, c, p)); +} + + static int post_ex(struct mdoc *mdoc) { @@ -174,19 +230,22 @@ post_ex(struct mdoc *mdoc) * our name (if it's been set). */ - if (0 == mdoc->last->data.elem.argc) + if (NULL == mdoc->last->args) return(1); - if (mdoc->last->data.elem.argv[0].sz) + if (mdoc->last->args->argv[0].sz) return(1); assert(mdoc->meta.name); mdoc_msg(mdoc, "writing %s argument: %s", - mdoc_argnames[MDOC_Std], mdoc->meta.name); + mdoc_argnames[MDOC_Std], + mdoc->meta.name); - mdoc->last->data.elem.argv[0].sz = 1; - mdoc->last->data.elem.argv[0].value = xcalloc(1, sizeof(char *)); - mdoc->last->data.elem.argv[0].value[0] = xstrdup(mdoc->meta.name); + assert(1 == mdoc->last->args->argv[0].sz); + + mdoc->last->args->argv[0].sz = 1; + mdoc->last->args->argv[0].value = xcalloc(1, sizeof(char *)); + mdoc->last->args->argv[0].value[0] = xstrdup(mdoc->meta.name); return(1); } @@ -240,9 +299,7 @@ post_sh(struct mdoc *mdoc) case (9): break; default: - return(mdoc_warn(mdoc, WARN_COMPAT, - "inappropriate section for " - "manual section")); + return(mwarn(mdoc, WBADSEC)); } break; default: @@ -288,7 +345,7 @@ post_dt(struct mdoc *mdoc) * --> title = TITLE, volume = local, msec = 0, arch = NULL */ - mdoc->meta.title = xstrdup(n->data.text.string); + mdoc->meta.title = xstrdup(n->string); mdoc_msg(mdoc, "title: %s", mdoc->meta.title); if (NULL == (n = n->next)) { @@ -306,14 +363,14 @@ post_dt(struct mdoc *mdoc) * arch = NULL */ - if ((cp = mdoc_a2msec(n->data.text.string))) { + if ((cp = mdoc_a2msec(n->string))) { mdoc->meta.vol = xstrdup(cp); errno = 0; - lval = strtol(n->data.text.string, &ep, 10); - if (n->data.text.string[0] != '\0' && *ep == '\0') + lval = strtol(n->string, &ep, 10); + if (n->string[0] != '\0' && *ep == '\0') mdoc->meta.msec = (int)lval; } else - mdoc->meta.vol = xstrdup(n->data.text.string); + mdoc->meta.vol = xstrdup(n->string); if (NULL == (n = n->next)) { mdoc_msg(mdoc, "volume: %s", mdoc->meta.vol); @@ -329,15 +386,15 @@ post_dt(struct mdoc *mdoc) * VOL */ - if ((cp = mdoc_a2vol(n->data.text.string))) { + if ((cp = mdoc_a2vol(n->string))) { free(mdoc->meta.vol); mdoc->meta.vol = xstrdup(cp); n = n->next; } else { - cp = mdoc_a2arch(n->data.text.string); + cp = mdoc_a2arch(n->string); if (NULL == cp) { free(mdoc->meta.vol); - mdoc->meta.vol = xstrdup(n->data.text.string); + mdoc->meta.vol = xstrdup(n->string); } else mdoc->meta.arch = xstrdup(cp); } @@ -385,7 +442,6 @@ static int post_bl_tagwidth(struct mdoc *mdoc) { struct mdoc_node *n; - struct mdoc_block *b; int sz; char buf[32]; @@ -394,30 +450,25 @@ post_bl_tagwidth(struct mdoc *mdoc) * to intuit our width from the first body element. */ - b = &mdoc->last->data.block; - - if (NULL == (n = b->body->child)) + if (NULL == (n = mdoc->last->body->child)) return(1); - assert(MDOC_It == n->tok); /* * Use the text width, if a text node, or the default macro * width if a macro. */ - if ((n = n->data.block.head->child)) { + if ((n = n->head->child)) { if (MDOC_TEXT != n->type) { - if (0 == (sz = mdoc_macro2len(n->tok))) + if (0 == (sz = (int)mdoc_macro2len(n->tok))) sz = -1; } else - sz = (int)strlen(n->data.text.string) + 1; + sz = (int)strlen(n->string) + 1; } else sz = -1; if (-1 == sz) { - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "cannot determine default %s", - mdoc_argnames[MDOC_Width])) + if ( ! mwarn(mdoc, WNOWIDTH)) return(0); sz = 10; } @@ -429,38 +480,53 @@ post_bl_tagwidth(struct mdoc *mdoc) * We're guaranteed that a MDOC_Width doesn't already exist. */ - (b->argc)++; - b->argv = xrealloc(b->argv, b->argc * sizeof(struct mdoc_arg)); + if (NULL == mdoc->last->args) { + mdoc->last->args = xcalloc + (1, sizeof(struct mdoc_arg)); + mdoc->last->args->refcnt = 1; + } + + n = mdoc->last; + sz = (int)n->args->argc; + + (n->args->argc)++; + + n->args->argv = xrealloc(n->args->argv, + n->args->argc * sizeof(struct mdoc_arg)); - b->argv[b->argc - 1].arg = MDOC_Width; - b->argv[b->argc - 1].line = mdoc->last->line; - b->argv[b->argc - 1].pos = mdoc->last->pos; - b->argv[b->argc - 1].sz = 1; - b->argv[b->argc - 1].value = xcalloc(1, sizeof(char *)); - b->argv[b->argc - 1].value[0] = xstrdup(buf); + n->args->argv[sz - 1].arg = MDOC_Width; + n->args->argv[sz - 1].line = mdoc->last->line; + n->args->argv[sz - 1].pos = mdoc->last->pos; + n->args->argv[sz - 1].sz = 1; + n->args->argv[sz - 1].value = xcalloc(1, sizeof(char *)); + n->args->argv[sz - 1].value[0] = xstrdup(buf); - mdoc_msg(mdoc, "adding %s argument: %dn", - mdoc_argnames[MDOC_Width], sz); + mdoc_msg(mdoc, "adding %s argument: %s", + mdoc_argnames[MDOC_Width], buf); return(1); } static int -post_bl_width(struct mdoc *mdoc) +post_bl_width(struct mdoc *m) { size_t width; int i, tok; char buf[32]; char **p; - for (i = 0; i < (int)mdoc->last->data.block.argc; i++) - if (MDOC_Width == mdoc->last->data.block.argv[i].arg) + if (NULL == m->last->args) + return(merr(m, ENOWIDTH)); + + for (i = 0; i < (int)m->last->args->argc; i++) + if (MDOC_Width == m->last->args->argv[i].arg) break; - assert(i < (int)mdoc->last->data.block.argc); - assert(1 == mdoc->last->data.block.argv[i].sz); - p = &mdoc->last->data.block.argv[i].value[0]; + if (i == (int)m->last->args->argc) + return(merr(m, ENOWIDTH)); + + p = &m->last->args->argv[i].value[0]; /* * If the value to -width is a macro, then we re-write it to be @@ -469,14 +535,12 @@ post_bl_width(struct mdoc *mdoc) if (xstrcmp(*p, "Ds")) width = 8; - else if (MDOC_MAX == (tok = mdoc_find(mdoc, *p))) + else if (MDOC_MAX == (tok = mdoc_tokhash_find(m->htab, *p))) return(1); else if (0 == (width = mdoc_macro2len(tok))) - return(mdoc_warn(mdoc, WARN_SYNTAX, - "%s macro has no length", - mdoc_argnames[MDOC_Width])); + return(mwarn(m, WNOWIDTH)); - mdoc_msg(mdoc, "re-writing %s argument: %s -> %zun", + mdoc_msg(m, "re-writing %s argument: %s -> %zun", mdoc_argnames[MDOC_Width], *p, width); /* The value already exists: free and reallocate it. */ @@ -484,7 +548,7 @@ post_bl_width(struct mdoc *mdoc) (void)snprintf(buf, sizeof(buf), "%zun", width); free(*p); - *p = strdup(buf); + *p = xstrdup(buf); return(1); } @@ -493,7 +557,7 @@ post_bl_width(struct mdoc *mdoc) static int post_bl(struct mdoc *mdoc) { - int i, r; + int i, r, len; if (MDOC_BLOCK != mdoc->last->type) return(1); @@ -506,10 +570,12 @@ post_bl(struct mdoc *mdoc) * rewritten into real lengths). */ - for (r = i = 0; i < (int)mdoc->last->data.block.argc; i++) { - if (MDOC_Tag == mdoc->last->data.block.argv[i].arg) + len = (int)(mdoc->last->args ? mdoc->last->args->argc : 0); + + for (r = i = 0; i < len; i++) { + if (MDOC_Tag == mdoc->last->args->argv[i].arg) r |= 1 << 0; - if (MDOC_Width == mdoc->last->data.block.argv[i].arg) + if (MDOC_Width == mdoc->last->args->argv[i].arg) r |= 1 << 1; } @@ -556,8 +622,11 @@ post_dd(struct mdoc *mdoc) (void)xstrlcpys(buf, mdoc->last->child, sizeof(buf)); - if (0 == (mdoc->meta.date = mdoc_atotime(buf))) - return(mdoc_err(mdoc, "invalid date syntax")); + if (0 == (mdoc->meta.date = mdoc_atotime(buf))) { + if ( ! mwarn(mdoc, WBADDATE)) + return(0); + mdoc->meta.date = time(NULL); + } mdoc_msg(mdoc, "date: %u", mdoc->meta.date); return(post_prologue(mdoc)); @@ -29,6 +29,8 @@ * Routines to parse arguments of macros. Arguments follow the syntax * of `-arg [val [valN...]]'. Arguments come in all types: quoted * arguments, multiple arguments per value, no-value arguments, etc. + * + * There's no limit to the number or arguments that may be allocated. */ #define ARGS_QUOTED (1 << 0) @@ -41,32 +43,31 @@ #define ARGV_MULTI (1 << 2) #define ARGV_OPT_SINGLE (1 << 3) +enum mwarn { + WQUOTPARM, + WARGVPARM, + WCOLEMPTY, + WTAILWS +}; + +enum merr { + EQUOTTERM, + EARGVAL +}; + static int argv_a2arg(int, const char *); static int args(struct mdoc *, int, int *, char *, int, char **); static int argv(struct mdoc *, int, int, - struct mdoc_arg *, int *, char *); + struct mdoc_argv *, int *, char *); static int argv_single(struct mdoc *, int, - struct mdoc_arg *, int *, char *); + struct mdoc_argv *, int *, char *); static int argv_opt_single(struct mdoc *, int, - struct mdoc_arg *, int *, char *); + struct mdoc_argv *, int *, char *); static int argv_multi(struct mdoc *, int, - struct mdoc_arg *, int *, char *); -static int pwarn(struct mdoc *, int, int, int); -static int perr(struct mdoc *, int, int, int); - -/* Warning messages. */ - -#define WQUOTPARM (0) -#define WARGVPARM (1) -#define WCOLEMPTY (2) -#define WTAILWS (3) - -/* Error messages. */ - -#define EQUOTTERM (0) -#define EARGVAL (1) -#define EARGMANY (2) + struct mdoc_argv *, int *, char *); +static int pwarn(struct mdoc *, int, int, enum mwarn); +static int perr(struct mdoc *, int, int, enum merr); /* Per-argument flags. */ @@ -209,59 +210,158 @@ static int mdoc_argflags[MDOC_MAX] = { }; +/* + * Parse an argument from line text. This comes in the form of -key + * [value0...], which may either have a single mandatory value, at least + * one mandatory value, an optional single value, or no value. + */ +int +mdoc_argv(struct mdoc *mdoc, int line, int tok, + struct mdoc_arg **v, int *pos, char *buf) +{ + int i; + char *p; + struct mdoc_argv tmp; + struct mdoc_arg *arg; + + if (0 == buf[*pos]) + return(ARGV_EOLN); + + assert( ! isspace((u_char)buf[*pos])); + + if ('-' != buf[*pos]) + return(ARGV_WORD); + + i = *pos; + p = &buf[++(*pos)]; + + assert(*pos > 0); + + /* LINTED */ + while (buf[*pos]) { + if (isspace((u_char)buf[*pos])) + if ('\\' != buf[*pos - 1]) + break; + (*pos)++; + } + + if (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. + */ + + if (MDOC_ARG_MAX == (tmp.arg = argv_a2arg(tok, p))) { + if ( ! pwarn(mdoc, line, i, WARGVPARM)) + return(ARGV_ERROR); + return(ARGV_WORD); + } + + while (buf[*pos] && isspace((u_char)buf[*pos])) + (*pos)++; + + /* FIXME: whitespace if no value. */ + + if ( ! argv(mdoc, tok, line, &tmp, pos, buf)) + return(ARGV_ERROR); + + if (NULL == (arg = *v)) { + v = xcalloc(1, sizeof(struct mdoc_arg)); + arg = *v; + } + + arg->argc++; + arg->argv = xrealloc(arg->argv, arg->argc * + sizeof(struct mdoc_arg)); + + (void)memcpy(&arg->argv[(int)arg->argc - 1], + &tmp, sizeof(struct mdoc_argv)); + + return(ARGV_ARG); +} + + +void +mdoc_argv_free(struct mdoc_arg *p) +{ + int i, j; + + if (p->refcnt && --(p->refcnt) > 0) + return; + + /* LINTED */ + for (i = 0; i < (int)p->argc; i++) { + if (0 == p->argv[i].sz) + continue; + /* 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); +} + + + static int -perr(struct mdoc *mdoc, int line, int pos, int code) +perr(struct mdoc *mdoc, int line, int pos, enum merr code) { - int c; + char *p; + + p = NULL; switch (code) { case (EQUOTTERM): - c = mdoc_perr(mdoc, line, pos, - "unterminated quoted parameter"); + p = "unterminated quoted parameter"; break; case (EARGVAL): - c = mdoc_perr(mdoc, line, pos, - "argument requires a value"); + p = "argument requires a value"; break; - case (EARGMANY): - c = mdoc_perr(mdoc, line, pos, - "too many values for argument"); - break; - default: - abort(); - /* NOTREACHED */ } - return(c); + + assert(p); + return(mdoc_perr(mdoc, line, pos, p)); } static int -pwarn(struct mdoc *mdoc, int line, int pos, int code) +pwarn(struct mdoc *mdoc, int line, int pos, enum mwarn code) { + char *p; int c; + p = NULL; + c = WARN_SYNTAX; + switch (code) { case (WQUOTPARM): - c = mdoc_pwarn(mdoc, line, pos, WARN_SYNTAX, - "unexpected quoted parameter"); + p = "unexpected quoted parameter"; break; case (WARGVPARM): - c = mdoc_pwarn(mdoc, line, pos, WARN_SYNTAX, - "argument-like parameter"); + p = "argument-like parameter"; break; case (WCOLEMPTY): - c = mdoc_pwarn(mdoc, line, pos, WARN_SYNTAX, - "last list column is empty"); + p = "last list column is empty"; + c = WARN_COMPAT; break; case (WTAILWS): - c = mdoc_pwarn(mdoc, line, pos, WARN_COMPAT, - "trailing whitespace"); + p = "trailing whitespace"; + c = WARN_COMPAT; break; - default: - abort(); - /* NOTREACHED */ } - return(c); + + assert(p); + return(mdoc_pwarn(mdoc, line, pos, c, p)); } @@ -287,7 +387,7 @@ mdoc_args(struct mdoc *mdoc, int line, break; assert(n); - c = (int)n->data.block.argc; + c = (int)(n->args ? n->args->argc : 0); assert(c > 0); /* @@ -298,7 +398,7 @@ mdoc_args(struct mdoc *mdoc, int line, /* LINTED */ for (i = 0; i < c; i++) { - switch (n->data.block.argv[i].arg) { + switch (n->args->argv[i].arg) { case (MDOC_Column): fl |= ARGS_TABSEP; fl &= ~ARGS_DELIM; @@ -312,6 +412,7 @@ mdoc_args(struct mdoc *mdoc, int line, break; } } + break; default: break; } @@ -621,48 +722,43 @@ argv_a2arg(int tok, const char *argv) static int argv_multi(struct mdoc *mdoc, int line, - struct mdoc_arg *v, int *pos, char *buf) + struct mdoc_argv *v, int *pos, char *buf) { int c, ppos; char *p; - v->sz = 0; - v->value = xcalloc(MDOC_LINEARG_MAX, sizeof(char *)); - ppos = *pos; - for (v->sz = 0; v->sz < MDOC_LINEARG_MAX; v->sz++) { + for (v->sz = 0; ; v->sz++) { if ('-' == buf[*pos]) break; c = args(mdoc, line, pos, buf, ARGS_QUOTED, &p); - if (ARGS_ERROR == c) { - free(v->value); + if (ARGS_ERROR == c) return(0); - } else if (ARGS_EOLN == c) + else if (ARGS_EOLN == c) break; + + if (0 == v->sz % 5) + v->value = xrealloc(v->value, + (v->sz + 5) * sizeof(char *)); + v->value[(int)v->sz] = p; } - if (0 < v->sz && v->sz < MDOC_LINEARG_MAX) + if (v->sz) return(1); - free(v->value); - if (0 == v->sz) - return(perr(mdoc, line, ppos, EARGVAL)); - - return(perr(mdoc, line, ppos, EARGMANY)); + return(perr(mdoc, line, ppos, EARGVAL)); } static int argv_opt_single(struct mdoc *mdoc, int line, - struct mdoc_arg *v, int *pos, char *buf) + struct mdoc_argv *v, int *pos, char *buf) { - int c, ppos; + int c; char *p; - ppos = *pos; - if ('-' == buf[*pos]) return(1); @@ -684,7 +780,7 @@ argv_opt_single(struct mdoc *mdoc, int line, */ static int argv_single(struct mdoc *mdoc, int line, - struct mdoc_arg *v, int *pos, char *buf) + struct mdoc_argv *v, int *pos, char *buf) { int c, ppos; char *p; @@ -711,7 +807,7 @@ argv_single(struct mdoc *mdoc, int line, */ static int argv(struct mdoc *mdoc, int tok, int line, - struct mdoc_arg *v, int *pos, char *buf) + struct mdoc_argv *v, int *pos, char *buf) { int fl; @@ -745,85 +841,3 @@ argv(struct mdoc *mdoc, int tok, int line, return(1); } - - -/* - * Parse an argument from line text. This comes in the form of -key - * [value0...], which may either have a single mandatory value, at least - * one mandatory value, an optional single value, or no value. - */ -int -mdoc_argv(struct mdoc *mdoc, int line, int tok, - struct mdoc_arg *v, int *pos, char *buf) -{ - int i; - char *p; - - (void)memset(v, 0, sizeof(struct mdoc_arg)); - - if (0 == buf[*pos]) - return(ARGV_EOLN); - - assert( ! isspace((u_char)buf[*pos])); - - if ('-' != buf[*pos]) - return(ARGV_WORD); - - i = *pos; - p = &buf[++(*pos)]; - - v->line = line; - v->pos = *pos; - - assert(*pos > 0); - - /* LINTED */ - while (buf[*pos]) { - if (isspace((u_char)buf[*pos])) - if ('\\' != buf[*pos - 1]) - break; - (*pos)++; - } - - if (buf[*pos]) - buf[(*pos)++] = 0; - - /* - * We now parse out the per-macro arguments. XXX - this can be - * made much cleaner using per-argument tables. See argv_a2arg - * for details. - */ - - if (MDOC_ARG_MAX == (v->arg = argv_a2arg(tok, p))) { - if ( ! pwarn(mdoc, line, i, WARGVPARM)) - return(ARGV_ERROR); - return(ARGV_WORD); - } - - while (buf[*pos] && isspace((u_char)buf[*pos])) - (*pos)++; - - /* FIXME: whitespace if no value. */ - - if ( ! argv(mdoc, tok, line, v, pos, buf)) - return(ARGV_ERROR); - - return(ARGV_ARG); -} - - -void -mdoc_argv_free(int sz, struct mdoc_arg *arg) -{ - int i; - - for (i = 0; i < sz; i++) { - if (0 == arg[i].sz) { - assert(NULL == arg[i].value); - continue; - } - assert(arg[i].value); - free(arg[i].value); - } -} - @@ -128,6 +128,11 @@ mdoc_tokhash_find(const void *arg, const char *tmp) assert(0 == (size_t)slot % sizeof(struct mdoc_macro)); slot /= sizeof(struct mdoc_macro); + /* + * FIXME: is this necessary, or do we only need to check the + * remaining characters (2+)? + */ + if (0 != strcmp(mdoc_macronames[slot], tmp)) return(MDOC_MAX); return(slot); @@ -31,11 +31,6 @@ * macro. */ -/* - * FIXME: don't use static mdoc_argv values, as they require us to do a - * complicated copy-over when actually assigning them to dynamic memory. - */ - #include "private.h" static int macro_obsolete(MACRO_PROT_ARGS); @@ -71,7 +66,6 @@ static int perr(struct mdoc *, int, int, int); #define ENOCTX (1) #define ENOPARMS (2) -#define EARGVLIM (3) /* Central table of library: who gets parsed how. */ @@ -204,11 +198,6 @@ perr(struct mdoc *mdoc, int line, int pos, int type) c = mdoc_perr(mdoc, line, pos, "macro doesn't expect parameters"); break; - case (EARGVLIM): - c = mdoc_perr(mdoc, line, pos, - "argument hard-limit %d reached", - MDOC_LINEARG_MAX); - break; default: abort(); /* NOTREACHED */ @@ -243,7 +232,7 @@ lookup(struct mdoc *mdoc, int line, int pos, int from, const char *p) { int res; - res = mdoc_find(mdoc, p); + res = mdoc_tokhash_find(mdoc->htab, p); if (MDOC_PARSED & mdoc_macros[from].flags) return(res); if (MDOC_MAX == res) @@ -731,16 +720,17 @@ macro_scoped_close(MACRO_PROT_ARGS) static int macro_text(MACRO_PROT_ARGS) { - int la, lastpunct, c, w, argc; - struct mdoc_arg argv[MDOC_LINEARG_MAX]; + int la, lastpunct, c, w; + struct mdoc_arg *arg; char *p; la = ppos; lastpunct = 0; + arg = NULL; - for (argc = 0; argc < MDOC_LINEARG_MAX; argc++) { + for (;;) { la = *pos; - c = mdoc_argv(mdoc, line, tok, &argv[argc], pos, buf); + c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); if (ARGV_EOLN == c) break; if (ARGV_WORD == c) { @@ -748,23 +738,12 @@ macro_text(MACRO_PROT_ARGS) break; } else if (ARGV_ARG == c) continue; - - mdoc_argv_free(argc, argv); + mdoc_argv_free(arg); return(0); } - if (MDOC_LINEARG_MAX == argc) { - mdoc_argv_free(argc - 1, argv); - return(perr(mdoc, line, ppos, EARGVLIM)); - } - - c = mdoc_elem_alloc(mdoc, line, ppos, - tok, (size_t)argc, argv); - - if (0 == c) { - mdoc_argv_free(argc, argv); + if ( ! mdoc_elem_alloc(mdoc, line, ppos, tok, arg)) return(0); - } mdoc->next = MDOC_NEXT_CHILD; @@ -774,11 +753,8 @@ macro_text(MACRO_PROT_ARGS) w = mdoc_args(mdoc, line, pos, buf, tok, &p); assert(ARGS_PHRASE != c); - if (ARGS_ERROR == w) { - mdoc_argv_free(argc, argv); + if (ARGS_ERROR == w) return(0); - } - if (ARGS_EOLN == w) break; if (ARGS_PUNCT == w) @@ -788,37 +764,29 @@ macro_text(MACRO_PROT_ARGS) lookup(mdoc, line, la, tok, p); if (MDOC_MAX != c && -1 != c) { - if (0 == lastpunct && ! rewind_elem(mdoc, tok)) { - mdoc_argv_free(argc, argv); + if (0 == lastpunct && ! rewind_elem(mdoc, tok)) return(0); - } - mdoc_argv_free(argc, argv); c = mdoc_macro(mdoc, c, line, la, pos, buf); if (0 == c) return(0); if (ppos > 1) return(1); return(append_delims(mdoc, line, pos, buf)); - } else if (-1 == c) { - mdoc_argv_free(argc, argv); + } else if (-1 == c) return(0); - } /* FIXME: .Fl and .Ar handling of `|'. */ if (ARGS_QWORD != w && mdoc_isdelim(p)) { - if (0 == lastpunct && ! rewind_elem(mdoc, tok)) { - mdoc_argv_free(argc, argv); + if (0 == lastpunct && ! rewind_elem(mdoc, tok)) return(0); - } lastpunct = 1; } else if (lastpunct) { - c = mdoc_elem_alloc(mdoc, line, ppos, - tok, (size_t)argc, argv); - if (0 == c) { - mdoc_argv_free(argc, argv); + c = mdoc_elem_alloc(mdoc, line, ppos, tok, arg); + + if (0 == c) return(0); - } + mdoc->next = MDOC_NEXT_CHILD; lastpunct = 0; } @@ -828,8 +796,6 @@ macro_text(MACRO_PROT_ARGS) mdoc->next = MDOC_NEXT_SIBLING; } - mdoc_argv_free(argc, argv); - if (0 == lastpunct && ! rewind_elem(mdoc, tok)) return(0); if (ppos > 1) @@ -865,8 +831,8 @@ macro_text(MACRO_PROT_ARGS) static int macro_scoped(MACRO_PROT_ARGS) { - int c, lastarg, argc; - struct mdoc_arg argv[MDOC_LINEARG_MAX]; + int c, lastarg; + struct mdoc_arg *arg; char *p; assert ( ! (MDOC_CALLABLE & mdoc_macros[tok].flags)); @@ -881,10 +847,12 @@ macro_scoped(MACRO_PROT_ARGS) } /* Parse arguments. */ + + arg = NULL; - for (argc = 0; argc < MDOC_LINEARG_MAX; argc++) { + for (;;) { lastarg = *pos; - c = mdoc_argv(mdoc, line, tok, &argv[argc], pos, buf); + c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); if (ARGV_EOLN == c) break; if (ARGV_WORD == c) { @@ -892,20 +860,11 @@ macro_scoped(MACRO_PROT_ARGS) break; } else if (ARGV_ARG == c) continue; - mdoc_argv_free(argc, argv); + mdoc_argv_free(arg); return(0); } - if (MDOC_LINEARG_MAX == argc) { - mdoc_argv_free(argc - 1, argv); - return(perr(mdoc, line, ppos, EARGVLIM)); - } - - c = mdoc_block_alloc(mdoc, line, ppos, - tok, (size_t)argc, argv); - mdoc_argv_free(argc, argv); - - if (0 == c) + if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, arg)) return(0); mdoc->next = MDOC_NEXT_CHILD; @@ -998,7 +957,7 @@ macro_scoped_line(MACRO_PROT_ARGS) int lastarg, c; char *p; - if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, 0, NULL)) + if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, NULL)) return(0); mdoc->next = MDOC_NEXT_CHILD; @@ -1082,7 +1041,7 @@ macro_constant_scoped(MACRO_PROT_ARGS) break; } - if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, 0, NULL)) + if ( ! mdoc_block_alloc(mdoc, line, ppos, tok, NULL)) return(0); mdoc->next = MDOC_NEXT_CHILD; @@ -1185,14 +1144,13 @@ macro_constant_scoped(MACRO_PROT_ARGS) static int macro_constant_delimited(MACRO_PROT_ARGS) { - int lastarg, flushed, j, c, maxargs, argc, + int lastarg, flushed, j, c, maxargs, igndelim, ignargs; - struct mdoc_arg argv[MDOC_LINEARG_MAX]; + struct mdoc_arg *arg; char *p; lastarg = ppos; flushed = 0; - /* * Maximum arguments per macro. Some of these have none and @@ -1241,13 +1199,12 @@ macro_constant_delimited(MACRO_PROT_ARGS) break; } - argc = 0; + arg = NULL; if ( ! ignargs) - for ( ; argc < MDOC_LINEARG_MAX; argc++) { + for (;;) { lastarg = *pos; - c = mdoc_argv(mdoc, line, tok, - &argv[argc], pos, buf); + c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); if (ARGV_EOLN == c) break; if (ARGV_WORD == c) { @@ -1255,21 +1212,11 @@ macro_constant_delimited(MACRO_PROT_ARGS) break; } else if (ARGV_ARG == c) continue; - mdoc_argv_free(argc, argv); + mdoc_argv_free(arg); return(0); } - if (MDOC_LINEARG_MAX == argc) { - mdoc_argv_free(argc - 1, argv); - return(perr(mdoc, line, ppos, EARGVLIM)); - } - - c = mdoc_elem_alloc(mdoc, line, ppos, - tok, (size_t)argc, argv); - - mdoc_argv_free(argc, argv); - - if (0 == c) + if ( ! mdoc_elem_alloc(mdoc, line, ppos, tok, arg)) return(0); mdoc->next = MDOC_NEXT_CHILD; @@ -1331,15 +1278,17 @@ macro_constant_delimited(MACRO_PROT_ARGS) static int macro_constant(MACRO_PROT_ARGS) { - int c, w, la, argc; - struct mdoc_arg argv[MDOC_LINEARG_MAX]; + int c, w, la; + struct mdoc_arg *arg; char *p; assert( ! (MDOC_CALLABLE & mdoc_macros[tok].flags)); - for (argc = 0; argc < MDOC_LINEARG_MAX; argc++) { + arg = NULL; + + for (;;) { la = *pos; - c = mdoc_argv(mdoc, line, tok, &argv[argc], pos, buf); + c = mdoc_argv(mdoc, line, tok, &arg, pos, buf); if (ARGV_EOLN == c) break; if (ARGV_WORD == c) { @@ -1347,22 +1296,11 @@ macro_constant(MACRO_PROT_ARGS) break; } else if (ARGV_ARG == c) continue; - - mdoc_argv_free(argc, argv); + mdoc_argv_free(arg); return(0); } - if (MDOC_LINEARG_MAX == argc) { - mdoc_argv_free(argc - 1, argv); - return(perr(mdoc, line, ppos, EARGVLIM)); - } - - c = mdoc_elem_alloc(mdoc, line, ppos, - tok, (size_t)argc, argv); - - mdoc_argv_free(argc, argv); - - if (0 == c) + if ( ! mdoc_elem_alloc(mdoc, line, ppos, tok, arg)) return(0); mdoc->next = MDOC_NEXT_CHILD; @@ -32,16 +32,12 @@ * in macro.c and validate.c. */ -static struct mdoc_arg *argdup(size_t, const struct mdoc_arg *); -static void argfree(size_t, struct mdoc_arg *); -static void argcpy(struct mdoc_arg *, - const struct mdoc_arg *); - static struct mdoc_node *mdoc_node_alloc(const struct mdoc *); static int mdoc_node_append(struct mdoc *, struct mdoc_node *); -static void mdoc_elem_free(struct mdoc_elem *); -static void mdoc_text_free(struct mdoc_text *); + +static int parsetext(struct mdoc *, int, char *); +static int parsemacro(struct mdoc *, int, char *); const char *const __mdoc_macronames[MDOC_MAX] = { @@ -173,76 +169,20 @@ mdoc_endparse(struct mdoc *mdoc) /* - * Main line-parsing routine. If the line is a macro-line (started with - * a '.' control character), then pass along to the parser, which parses - * subsequent macros until the end of line. If normal text, simply - * append the entire line to the chain. + * Main parse routine. Parses a single line -- really just hands off to + * the macro or text parser. */ int -mdoc_parseln(struct mdoc *mdoc, int line, char *buf) +mdoc_parseln(struct mdoc *m, int ln, char *buf) { - int c, i; - char tmp[5]; - - if (MDOC_HALT & mdoc->flags) - return(0); - - mdoc->linetok = 0; - - if ('.' != *buf) { - /* - * Free-form text. Not allowed in the prologue. - */ - if (SEC_PROLOGUE == mdoc->lastnamed) - return(mdoc_perr(mdoc, line, 0, - "no text in prologue")); - if ( ! mdoc_word_alloc(mdoc, line, 0, buf)) - return(0); - mdoc->next = MDOC_NEXT_SIBLING; - return(1); - } - - /* - * Control-character detected. Begin the parsing sequence. - */ - - if (buf[1] && '\\' == buf[1]) - if (buf[2] && '\"' == buf[2]) - return(1); - - i = 1; - while (buf[i] && ! isspace((u_char)buf[i]) && - i < (int)sizeof(tmp)) - i++; - - if (i == (int)sizeof(tmp)) { - mdoc->flags |= MDOC_HALT; - return(mdoc_perr(mdoc, line, 1, "unknown macro")); - } else if (i <= 2) { - mdoc->flags |= MDOC_HALT; - return(mdoc_perr(mdoc, line, 1, "unknown macro")); - } - - i--; - - (void)memcpy(tmp, buf + 1, (size_t)i); - tmp[i++] = 0; - - if (MDOC_MAX == (c = mdoc_find(mdoc, tmp))) { - mdoc->flags |= MDOC_HALT; - return(mdoc_perr(mdoc, line, 1, "unknown macro")); - } - - while (buf[i] && isspace((u_char)buf[i])) - i++; + /* If in error-mode, then we parse no more. */ - if ( ! mdoc_macro(mdoc, c, line, 1, &i, buf)) { - mdoc->flags |= MDOC_HALT; + if (MDOC_HALT & m->flags) return(0); - } - return(1); + return('.' == *buf ? parsemacro(m, ln, buf) : + parsetext(m, ln, buf)); } @@ -297,21 +237,26 @@ mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, int -mdoc_macro(struct mdoc *mdoc, int tok, - int ln, int ppos, int *pos, char *buf) +mdoc_macro(struct mdoc *m, int tok, + int ln, int pp, int *pos, char *buf) { - assert(mdoc_macros[tok].fp); + /* FIXME - these should happen during validation. */ if (MDOC_PROLOGUE & mdoc_macros[tok].flags && - SEC_PROLOGUE != mdoc->lastnamed) - return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document body")); + SEC_PROLOGUE != m->lastnamed) + return(mdoc_perr(m, ln, pp, + "disallowed in document body")); + if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && - SEC_PROLOGUE == mdoc->lastnamed) - return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document prologue")); - if (1 != ppos && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) - return(mdoc_perr(mdoc, ln, ppos, "macro not callable")); - return((*mdoc_macros[tok].fp)(mdoc, tok, ln, ppos, pos, buf)); + SEC_PROLOGUE == m->lastnamed) + return(mdoc_perr(m, ln, pp, + "disallowed in prologue")); + + if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) + return(mdoc_perr(m, ln, pp, "not callable")); + + return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); } @@ -323,27 +268,6 @@ mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p) assert(mdoc->first); assert(MDOC_ROOT != p->type); - /* See if we exceed the suggest line-max. */ - - switch (p->type) { - case (MDOC_TEXT): - /* FALLTHROUGH */ - case (MDOC_ELEM): - /* FALLTHROUGH */ - case (MDOC_BLOCK): - mdoc->linetok++; - break; - default: - break; - } - - /* This sort-of works (re-opening of text macros...). */ - if (mdoc->linetok > MDOC_LINEARG_SOFTMAX) - if ( ! mdoc_nwarn(mdoc, p, WARN_COMPAT, - "suggested %d tokens per line exceeded (has %d)", - MDOC_LINEARG_SOFTMAX, mdoc->linetok)) - return(0); - switch (mdoc->next) { case (MDOC_NEXT_SIBLING): mdoc->last->next = p; @@ -365,15 +289,15 @@ mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p) switch (p->type) { case (MDOC_HEAD): assert(MDOC_BLOCK == p->parent->type); - p->parent->data.block.head = p; + p->parent->head = p; break; case (MDOC_TAIL): assert(MDOC_BLOCK == p->parent->type); - p->parent->data.block.tail = p; + p->parent->tail = p; break; case (MDOC_BODY): assert(MDOC_BLOCK == p->parent->type); - p->parent->data.block.body = p; + p->parent->body = p; break; default: break; @@ -468,7 +392,7 @@ mdoc_root_alloc(struct mdoc *mdoc) int mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, - int tok, size_t argsz, const struct mdoc_arg *args) + int tok, struct mdoc_arg *args) { struct mdoc_node *p; @@ -478,8 +402,10 @@ mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, p->line = line; p->type = MDOC_BLOCK; p->tok = tok; - p->data.block.argc = argsz; - p->data.block.argv = argdup(argsz, args); + p->args = args; + + if (args) + (args->refcnt)++; return(mdoc_node_append(mdoc, p)); } @@ -487,7 +413,7 @@ mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, int mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, - int tok, size_t argsz, const struct mdoc_arg *args) + int tok, struct mdoc_arg *args) { struct mdoc_node *p; @@ -497,8 +423,10 @@ mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, p->pos = pos; p->type = MDOC_ELEM; p->tok = tok; - p->data.elem.argc = argsz; - p->data.elem.argv = argdup(argsz, args); + p->args = args; + + if (args) + (args->refcnt)++; return(mdoc_node_append(mdoc, p)); } @@ -515,77 +443,20 @@ mdoc_word_alloc(struct mdoc *mdoc, p->line = line; p->pos = pos; p->type = MDOC_TEXT; - p->data.text.string = xstrdup(word); + p->string = xstrdup(word); return(mdoc_node_append(mdoc, p)); } -static void -argfree(size_t sz, struct mdoc_arg *p) -{ - int i, j; - - if (0 == sz) - return; - - assert(p); - /* LINTED */ - for (i = 0; i < (int)sz; i++) - if (p[i].sz > 0) { - assert(p[i].value); - /* LINTED */ - for (j = 0; j < (int)p[i].sz; j++) - free(p[i].value[j]); - free(p[i].value); - } - free(p); -} - - -static void -mdoc_elem_free(struct mdoc_elem *p) -{ - - argfree(p->argc, p->argv); -} - - -static void -mdoc_block_free(struct mdoc_block *p) -{ - - argfree(p->argc, p->argv); -} - - -static void -mdoc_text_free(struct mdoc_text *p) -{ - - if (p->string) - free(p->string); -} - - void mdoc_node_free(struct mdoc_node *p) { - switch (p->type) { - case (MDOC_TEXT): - mdoc_text_free(&p->data.text); - break; - case (MDOC_ELEM): - mdoc_elem_free(&p->data.elem); - break; - case (MDOC_BLOCK): - mdoc_block_free(&p->data.block); - break; - default: - break; - } - + if (p->string) + free(p->string); + if (p->args) + mdoc_argv_free(p->args); free(p); } @@ -603,43 +474,78 @@ mdoc_node_freelist(struct mdoc_node *p) } -int -mdoc_find(const struct mdoc *mdoc, const char *key) +/* + * Parse free-form text, that is, a line that does not begin with the + * control character. + */ +static int +parsetext(struct mdoc *mdoc, int line, char *buf) { - return(mdoc_tokhash_find(mdoc->htab, key)); + if (SEC_PROLOGUE == mdoc->lastnamed) + return(mdoc_perr(mdoc, line, 0, + "text disallowed in prologue")); + + if ( ! mdoc_word_alloc(mdoc, line, 0, buf)) + return(0); + + mdoc->next = MDOC_NEXT_SIBLING; + return(1); } -static void -argcpy(struct mdoc_arg *dst, const struct mdoc_arg *src) +/* + * Parse a macro line, that is, a line beginning with the control + * character. + */ +int +parsemacro(struct mdoc *m, int ln, char *buf) { - int i; + int i, c; + char mac[5]; - dst->line = src->line; - dst->pos = src->pos; - dst->arg = src->arg; - if (0 == (dst->sz = src->sz)) - return; - dst->value = xcalloc(dst->sz, sizeof(char *)); - for (i = 0; i < (int)dst->sz; i++) - dst->value[i] = xstrdup(src->value[i]); -} + /* Comments are quickly ignored. */ + + if (buf[1] && '\\' == buf[1]) + if (buf[2] && '\"' == buf[2]) + return(1); + /* Copy the first word into a nil-terminated buffer. */ -static struct mdoc_arg * -argdup(size_t argsz, const struct mdoc_arg *args) -{ - struct mdoc_arg *pp; - int i; + for (i = 1; i < 5; i++) { + if (0 == (mac[i - 1] = buf[i])) + break; + else if (isspace((unsigned char)buf[i])) + break; + } - if (0 == argsz) - return(NULL); + mac[i - 1] = 0; + + if (i == 5 || i <= 2) { + (void)mdoc_perr(m, ln, 1, "unknown macro: %s%s", + mac, i == 5 ? "..." : ""); + goto err; + } + + if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) { + (void)mdoc_perr(m, ln, 1, "unknown macro: %s", mac); + goto err; + } - pp = xcalloc((size_t)argsz, sizeof(struct mdoc_arg)); - for (i = 0; i < (int)argsz; i++) - argcpy(&pp[i], &args[i]); + /* The macro is sane. Jump to the next word. */ - return(pp); -} + while (buf[i] && isspace((unsigned char)buf[i])) + i++; + /* Begin recursive parse sequence. */ + + if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) + goto err; + + return(1); + +err: /* Error out. */ + + m->flags |= MDOC_HALT; + return(0); +} @@ -173,15 +173,6 @@ enum mdoc_warn { WARN_COMPAT }; -/* An argument to a macro (multiple values = `It -column'). */ -struct mdoc_arg { - int arg; - int line; - int pos; - size_t sz; - char **value; -}; - /* Type of a syntax node. */ enum mdoc_type { MDOC_TEXT, @@ -229,31 +220,19 @@ struct mdoc_meta { char *name; }; -/* Text-only node. */ -struct mdoc_text { - char *string; -}; - -/* Block (scoped) node. */ -struct mdoc_block { - size_t argc; - struct mdoc_arg *argv; - struct mdoc_node *head; - struct mdoc_node *body; - struct mdoc_node *tail; +/* An argument to a macro (multiple values = `It -column'). */ +struct mdoc_argv { + int arg; + int line; + int pos; + size_t sz; + char **value; }; -/* In-line element node. */ -struct mdoc_elem { +struct mdoc_arg { size_t argc; - struct mdoc_arg *argv; -}; - -/* Typed nodes of an AST node. */ -union mdoc_data { - struct mdoc_text text; - struct mdoc_elem elem; - struct mdoc_block block; + struct mdoc_argv *argv; + unsigned int refcnt; }; /* Node in AST. */ @@ -269,10 +248,16 @@ struct mdoc_node { #define MDOC_VALID (1 << 0) #define MDOC_ACTED (1 << 1) enum mdoc_type type; - union mdoc_data data; enum mdoc_sec sec; + struct mdoc_arg *args; /* BLOCK/ELEM */ + struct mdoc_node *head; /* BLOCK */ + struct mdoc_node *body; /* BLOCK */ + struct mdoc_node *tail; /* BLOCK */ + char *string; /* TEXT */ }; +#define MDOC_IGN_SCOPE (1 << 0) + /* Call-backs for parse messages. */ struct mdoc_cb { void (*mdoc_msg)(void *, int, int, const char *); @@ -297,6 +282,9 @@ void mdoc_free(struct mdoc *); /* Allocate a new parser instance. */ struct mdoc *mdoc_alloc(void *data, const struct mdoc_cb *); +/* Set parse options. */ +void mdoc_setflags(struct mdoc *, int); + /* Parse a single line in a stream (boolean retval). */ int mdoc_parseln(struct mdoc *, int, char *buf); @@ -44,11 +44,13 @@ struct mmain { u_long bufsz; /* Input buffer size. */ char *in; /* Input file name. */ int fdin; /* Input file desc. */ + int pflags; /* Parse flags. */ }; extern char *__progname; -static int getsopts(struct mmain *, char *); +static int optswarn(struct mmain *, char *); +static int optsopt(struct mmain *, char *); static int parse(struct mmain *); static void msg_msg(void *, int, int, const char *); static int msg_err(void *, int, int, const char *); @@ -109,7 +111,7 @@ mmain_getopt(struct mmain *p, int argc, char *argv[], extern int optind; - sz = strlcpy(opts, "VvW:", 32); + sz = strlcpy(opts, "VvW:f:", 32); assert(sz < 32); if (u) { @@ -122,6 +124,10 @@ mmain_getopt(struct mmain *p, int argc, char *argv[], /* LINTED */ while (-1 != (c = getopt(argc, argv, opts))) switch (c) { + case ('f'): + if ( ! optsopt(p, optarg)) + return(-1); + break; case ('v'): p->dbg++; break; @@ -129,7 +135,7 @@ mmain_getopt(struct mmain *p, int argc, char *argv[], (void)printf("%s %s\n", __progname, VERSION); return(0); case ('W'): - if ( ! getsopts(p, optarg)) + if ( ! optswarn(p, optarg)) return(-1); break; case ('?'): @@ -208,7 +214,26 @@ mmain_mdoc(struct mmain *p) static int -getsopts(struct mmain *p, char *arg) +optsopt(struct mmain *p, char *arg) +{ + char *v; + char *toks[] = { "ignore-scope", NULL }; + + while (*arg) + switch (getsubopt(&arg, toks, &v)) { + case (0): + p->pflags |= MDOC_IGN_SCOPE; + break; + default: + return(0); + } + + return(1); +} + + +static int +optswarn(struct mmain *p, char *arg) { char *v; char *toks[] = { "all", "compat", @@ -33,6 +33,7 @@ struct mdoc { int linetok; int flags; #define MDOC_HALT (1 << 0) + int pflags; enum mdoc_next next; struct mdoc_node *last; struct mdoc_node *first; @@ -42,14 +43,6 @@ struct mdoc { }; -/* Hard-limit of macro arguments. */ - -#define MDOC_LINEARG_MAX 9 - -/* Suggested limit of macro arguments. */ - -#define MDOC_LINEARG_SOFTMAX 9 - #define MACRO_PROT_ARGS struct mdoc *mdoc, int tok, int line, \ int ppos, int *pos, char *buf @@ -106,13 +99,12 @@ int mdoc_verr(struct mdoc *, int, int, const char *, ...); int mdoc_macro(MACRO_PROT_ARGS); -int mdoc_find(const struct mdoc *, const char *); int mdoc_word_alloc(struct mdoc *, int, int, const char *); int mdoc_elem_alloc(struct mdoc *, int, int, - int, size_t, const struct mdoc_arg *); + int, struct mdoc_arg *); int mdoc_block_alloc(struct mdoc *, int, int, - int, size_t, const struct mdoc_arg *); + int, struct mdoc_arg *); int mdoc_root_alloc(struct mdoc *); int mdoc_head_alloc(struct mdoc *, int, int, int); int mdoc_tail_alloc(struct mdoc *, int, int, int); @@ -141,13 +133,13 @@ int mdoc_valid_post(struct mdoc *); int mdoc_action_pre(struct mdoc *, struct mdoc_node *); int mdoc_action_post(struct mdoc *); -int mdoc_argv(struct mdoc *, int, int, - struct mdoc_arg *, int *, char *); +int mdoc_argv(struct mdoc *, int, int, + struct mdoc_arg **, int *, char *); #define ARGV_ERROR (-1) #define ARGV_EOLN (0) #define ARGV_ARG (1) #define ARGV_WORD (2) -void mdoc_argv_free(int, struct mdoc_arg *); +void mdoc_argv_free(struct mdoc_arg *); int mdoc_args(struct mdoc *, int, int *, char *, int, char **); #define ARGS_ERROR (-1) @@ -24,6 +24,7 @@ #include "private.h" /* FIXME: .Bl -diag can't have non-text children in HEAD. */ +/* TODO: ignoring Pp (it's superfluous in some invocations). */ /* * Pre- and post-validate macros as they're parsed. Pre-validation @@ -36,11 +37,41 @@ #define PRE_ARGS struct mdoc *mdoc, const struct mdoc_node *n #define POST_ARGS struct mdoc *mdoc +enum merr { + ENODATA, + ENOPROLOGUE, + ELINE, + EATT, + ENAME, + ELISTTYPE, + EDISPTYPE, + EMULTIDISP, + EMULTILIST, + EARGREP, + EBOOL, + ENESTDISP +}; + +enum mwarn { + WWRONGMSEC, + WSECOOO, + WSECREP, + WBADSTAND, + WNAMESECINC, + WNOMULTILINE, + WMULTILINE, + WLINE, + WNOLINE, + WPROLOOO, + WPROLREP, + WARGVAL, + WBADSEC, + WBADMSEC +}; + typedef int (*v_pre)(PRE_ARGS); typedef int (*v_post)(POST_ARGS); -/* TODO: ignoring Pp (it's superfluous in some invocations). */ - struct valids { v_pre *pre; v_post *post; @@ -48,34 +79,32 @@ struct valids { /* Utility checks. */ +static int nwarn(struct mdoc *, + const struct mdoc_node *, enum mwarn); +static int nerr(struct mdoc *, + const struct mdoc_node *, enum merr); static int check_parent(PRE_ARGS, int, enum mdoc_type); static int check_msec(PRE_ARGS, ...); static int check_sec(PRE_ARGS, ...); static int check_stdarg(PRE_ARGS); - static int check_text(struct mdoc *, int, int, const char *); static int check_argv(struct mdoc *, const struct mdoc_node *, - const struct mdoc_arg *); - + const struct mdoc_argv *); +static int check_args(struct mdoc *, + const struct mdoc_node *); static int err_child_lt(struct mdoc *, const char *, int); static int warn_child_lt(struct mdoc *, const char *, int); static int err_child_gt(struct mdoc *, const char *, int); static int warn_child_gt(struct mdoc *, const char *, int); static int err_child_eq(struct mdoc *, const char *, int); static int warn_child_eq(struct mdoc *, const char *, int); - -/* Utility auxiliaries. */ - static inline int count_child(struct mdoc *); static inline int warn_count(struct mdoc *, const char *, int, const char *, int); static inline int err_count(struct mdoc *, const char *, int, const char *, int); - -/* Specific pre-child-parse routines. */ - static int pre_an(PRE_ARGS); static int pre_bd(PRE_ARGS); static int pre_bl(PRE_ARGS); @@ -93,14 +122,10 @@ static int pre_prologue(PRE_ARGS); static int pre_rv(PRE_ARGS); static int pre_sh(PRE_ARGS); static int pre_ss(PRE_ARGS); - -/* Specific post-child-parse routines. */ - static int herr_ge1(POST_ARGS); static int hwarn_le1(POST_ARGS); static int herr_eq0(POST_ARGS); static int eerr_eq0(POST_ARGS); -static int eerr_le1(POST_ARGS); static int eerr_le2(POST_ARGS); static int eerr_eq1(POST_ARGS); static int eerr_ge1(POST_ARGS); @@ -110,7 +135,6 @@ static int bwarn_ge1(POST_ARGS); static int hwarn_eq1(POST_ARGS); static int ewarn_ge1(POST_ARGS); static int ebool(POST_ARGS); - static int post_an(POST_ARGS); static int post_at(POST_ARGS); static int post_bf(POST_ARGS); @@ -124,8 +148,6 @@ static int post_sh_body(POST_ARGS); static int post_sh_head(POST_ARGS); static int post_st(POST_ARGS); -/* Collections of pre-child-parse routines. */ - static v_pre pres_an[] = { pre_an, NULL }; static v_pre pres_bd[] = { pre_display, pre_bd, NULL }; static v_pre pres_bl[] = { pre_bl, NULL }; @@ -142,9 +164,6 @@ static v_pre pres_os[] = { pre_prologue, pre_os, NULL }; static v_pre pres_rv[] = { pre_rv, NULL }; static v_pre pres_sh[] = { pre_sh, NULL }; static v_pre pres_ss[] = { pre_ss, NULL }; - -/* Collections of post-child-parse routines. */ - static v_post posts_bool[] = { eerr_eq1, ebool, NULL }; static v_post posts_bd[] = { herr_eq0, bwarn_ge1, NULL }; static v_post posts_text[] = { eerr_ge1, NULL }; @@ -160,7 +179,7 @@ static v_post posts_pf[] = { eerr_eq1, NULL }; static v_post posts_lb[] = { eerr_eq1, NULL }; static v_post posts_st[] = { eerr_eq1, post_st, NULL }; static v_post posts_pp[] = { ewarn_eq0, NULL }; -static v_post posts_ex[] = { eerr_le1, post_ex, NULL }; +static v_post posts_ex[] = { eerr_eq0, post_ex, NULL }; static v_post posts_an[] = { post_an, NULL }; static v_post posts_at[] = { post_at, NULL }; static v_post posts_xr[] = { eerr_ge1, eerr_le2, NULL }; @@ -171,8 +190,6 @@ static v_post posts_fo[] = { hwarn_eq1, bwarn_ge1, NULL }; static v_post posts_bk[] = { herr_eq0, bwarn_ge1, NULL }; static v_post posts_fd[] = { ewarn_ge1, NULL }; -/* Per-macro pre- and post-child-check routine collections. */ - const struct valids mdoc_valids[MDOC_MAX] = { { NULL, NULL }, /* \" */ { pres_dd, posts_text }, /* Dd */ @@ -289,39 +306,18 @@ mdoc_valid_pre(struct mdoc *mdoc, const struct mdoc_node *node) { v_pre *p; - struct mdoc_arg *argv; - size_t argc; - int line, pos, i, j; + int line, pos; const char *tp; if (MDOC_TEXT == node->type) { - tp = node->data.text.string; + tp = node->string; line = node->line; pos = node->pos; return(check_text(mdoc, line, pos, tp)); } - if (MDOC_BLOCK == node->type || MDOC_ELEM == node->type) { - argv = MDOC_BLOCK == node->type ? - node->data.block.argv : - node->data.elem.argv; - argc = MDOC_BLOCK == node->type ? - node->data.block.argc : - node->data.elem.argc; - - for (i = 0; i < (int)argc; i++) { - for (j = 0; j < (int)argv[i].sz; j++) { - tp = argv[i].value[j]; - line = argv[i].line; - pos = argv[i].pos; - if ( ! check_text(mdoc, line, pos, tp)) - return(0); - } - if ( ! check_argv(mdoc, node, &argv[i])) - return(0); - } - } - + if ( ! check_args(mdoc, node)) + return(0); if (NULL == mdoc_valids[node->tok].pre) return(1); for (p = mdoc_valids[node->tok].pre; *p; p++) @@ -363,6 +359,123 @@ mdoc_valid_post(struct mdoc *mdoc) } +#define merr(m, t) nerr((m), (m)->last, (t)) +static int +nerr(struct mdoc *m, const struct mdoc_node *n, enum merr type) +{ + char *p; + + p = NULL; + + switch (type) { + case (ENESTDISP): + p = "displays may not be nested"; + break; + case (EBOOL): + p = "expected boolean value"; + break; + case (EARGREP): + p = "argument repeated"; + break; + case (EMULTIDISP): + p = "multiple display types specified"; + break; + case (EMULTILIST): + p = "multiple list types specified"; + break; + case (ELISTTYPE): + p = "missing list type"; + break; + case (EDISPTYPE): + p = "missing display type"; + break; + case (ELINE): + p = "expected line arguments"; + break; + case (ENOPROLOGUE): + p = "document has no prologue"; + break; + case (ENODATA): + p = "document has no data"; + break; + case (EATT): + p = "expected valid AT&T symbol"; + break; + case (ENAME): + p = "default name not yet set"; + break; + } + + assert(p); + return(mdoc_nerr(m, n, p)); +} + + +#define mwarn(m, t) nwarn((m), (m)->last, (t)) +static int +nwarn(struct mdoc *m, const struct mdoc_node *n, enum mwarn type) +{ + char *p; + enum mdoc_warn c; + + c = WARN_SYNTAX; + p = NULL; + + switch (type) { + case (WBADMSEC): + p = "inappropriate manual section"; + c = WARN_COMPAT; + break; + case (WBADSEC): + p = "inappropriate document section"; + c = WARN_COMPAT; + break; + case (WARGVAL): + p = "argument value suggested"; + c = WARN_COMPAT; + break; + case (WPROLREP): + p = "prologue macros repeated"; + c = WARN_COMPAT; + break; + case (WPROLOOO): + p = "prologue macros out-of-order"; + c = WARN_COMPAT; + break; + case (WNOLINE): + p = "suggested no line arguments"; + break; + case (WLINE): + p = "suggested line arguments"; + break; + case (WMULTILINE): + p = "suggested multi-line arguments"; + break; + case (WNOMULTILINE): + p = "suggested no multi-line arguments"; + break; + case (WWRONGMSEC): + p = "document section in wrong manual section"; + c = WARN_COMPAT; + break; + case (WSECOOO): + p = "document section out of conventional order"; + break; + case (WSECREP): + p = "document section repeated"; + break; + case (WBADSTAND): + p = "unknown standard"; + break; + case (WNAMESECINC): + p = "NAME section contents incomplete/badly-ordered"; + break; + } + assert(p); + return(mdoc_nwarn(m, n, c, p)); +} + + static inline int warn_count(struct mdoc *m, const char *k, @@ -370,8 +483,7 @@ warn_count(struct mdoc *m, const char *k, { return(mdoc_warn(m, WARN_SYNTAX, - "suggests %s %s %d (has %d)", - v, k, want, has)); + "suggests %s %s %d (has %d)", v, k, want, has)); } @@ -380,8 +492,8 @@ err_count(struct mdoc *m, const char *k, int want, const char *v, int has) { - return(mdoc_err(m, "requires %s %s %d (has %d)", - v, k, want, has)); + return(mdoc_err(m, + "requires %s %s %d (has %d)", v, k, want, has)); } @@ -420,7 +532,7 @@ b##lvl##_##name(POST_ARGS) \ { \ if (MDOC_BODY != mdoc->last->type) \ return(1); \ - return(func(mdoc, "multiline parameters", (num))); \ + return(func(mdoc, "multi-line arguments", (num))); \ } #define CHECK_ELEM_DEFN(name, lvl, func, num) \ @@ -428,7 +540,7 @@ static int \ e##lvl##_##name(POST_ARGS) \ { \ assert(MDOC_ELEM == mdoc->last->type); \ - return(func(mdoc, "line parameters", (num))); \ + return(func(mdoc, "line arguments", (num))); \ } #define CHECK_HEAD_DEFN(name, lvl, func, num) \ @@ -437,7 +549,7 @@ h##lvl##_##name(POST_ARGS) \ { \ if (MDOC_HEAD != mdoc->last->type) \ return(1); \ - return(func(mdoc, "line parameters", (num))); \ + return(func(mdoc, "line arguments", (num))); \ } @@ -453,7 +565,6 @@ CHECK_ELEM_DEFN(eq0, warn, warn_child_eq, 0) /* ewarn_eq0() */ CHECK_ELEM_DEFN(ge1, warn, warn_child_gt, 0) /* ewarn_gt1() */ CHECK_ELEM_DEFN(eq1, err, err_child_eq, 1) /* eerr_eq1() */ CHECK_ELEM_DEFN(le2, err, err_child_lt, 3) /* eerr_le2() */ -CHECK_ELEM_DEFN(le1, err, err_child_lt, 2) /* eerr_le1() */ CHECK_ELEM_DEFN(eq0, err, err_child_eq, 0) /* eerr_eq0() */ CHECK_ELEM_DEFN(ge1, err, err_child_gt, 0) /* eerr_ge1() */ CHECK_HEAD_DEFN(eq0, err, err_child_eq, 0) /* herr_eq0() */ @@ -466,12 +577,10 @@ static int check_stdarg(PRE_ARGS) { - if (MDOC_Std == n->data.elem.argv[0].arg && - 1 == n->data.elem.argc) - return(1); - - return(mdoc_nwarn(mdoc, n, WARN_COMPAT, - "one argument suggested")); + if (n->args && 1 == n->args->argc) + if (MDOC_Std == n->args->argv[0].arg) + return(1); + return(nwarn(mdoc, n, WARGVAL)); } @@ -484,6 +593,7 @@ check_sec(PRE_ARGS, ...) va_start(ap, n); for (;;) { + /* LINTED */ sec = (enum mdoc_sec)va_arg(ap, int); if (SEC_CUSTOM == sec) break; @@ -494,8 +604,7 @@ check_sec(PRE_ARGS, ...) } va_end(ap); - return(mdoc_nwarn(mdoc, n, WARN_COMPAT, - "inappropriate document section for macro")); + return(nwarn(mdoc, n, WBADSEC)); } @@ -507,6 +616,7 @@ check_msec(PRE_ARGS, ...) va_start(ap, n); for (;;) { + /* LINTED */ if (0 == (msec = va_arg(ap, int))) break; if (msec != mdoc->meta.msec) @@ -516,44 +626,41 @@ check_msec(PRE_ARGS, ...) } va_end(ap); - return(mdoc_nwarn(mdoc, n, WARN_COMPAT, - "inappropriate manual section for macro")); + return(nwarn(mdoc, n, WBADMSEC)); } -/* - * Check over an argument. When this has more stuff in it, make this - * into a table-driven function; until then, a switch is fine. - */ static int -check_argv(struct mdoc *mdoc, - const struct mdoc_node *node, - const struct mdoc_arg *argv) +check_args(struct mdoc *m, const struct mdoc_node *n) { + int i; + if (NULL == n->args) + return(1); - switch (argv->arg) { - case (MDOC_Std): - switch (node->tok) { - case (MDOC_Ex): - /* - * If the -std does not have an argument, then - * set it with the default name (if set). This - * only happens with MDOC_Ex. - */ - if (1 == argv->sz) - return(1); - assert(0 == argv->sz); - if (mdoc->meta.name) - return(1); - return(mdoc_nerr(mdoc, node, - "default name not yet set")); - default: - break; - } - break; - default: - break; + for (i = 0; i < (int)n->args->argc; i++) + if ( ! check_argv(m, n, &n->args->argv[i])) + return(0); + + return(1); +} + + +static int +check_argv(struct mdoc *m, const struct mdoc_node *n, + const struct mdoc_argv *v) +{ + int i; + + for (i = 0; i < (int)v->sz; i++) + if ( ! check_text(m, v->line, v->pos, v->value[i])) + return(0); + + if (MDOC_Std == v->arg && MDOC_Ex == n->tok) { + /* `Nm' name must be set. */ + if (v->sz || m->meta.name) + return(1); + return(nerr(m, n, ENAME)); } return(1); @@ -570,7 +677,7 @@ check_text(struct mdoc *mdoc, int line, int pos, const char *p) for ( ; *p; p++) { if ( ! isprint((u_char)*p) && '\t' != *p) return(mdoc_perr(mdoc, line, pos, - "invalid non-printing characters")); + "invalid non-printing character")); if ('\\' != *p) continue; if ((c = mdoc_isescape(p))) { @@ -620,31 +727,27 @@ pre_display(PRE_ARGS) if (NULL == node) return(1); - return(mdoc_nerr(mdoc, n, "displays may not be nested")); + return(nerr(mdoc, n, ENESTDISP)); } static int pre_bl(PRE_ARGS) { - int type, i, width, offset; - struct mdoc_arg *argv; - size_t argc; + int i, type, width, offset; if (MDOC_BLOCK != n->type) return(1); - - argc = n->data.block.argc; + if (NULL == n->args) + return(nerr(mdoc, n, ELISTTYPE)); /* Make sure that only one type of list is specified. */ type = offset = width = -1; /* LINTED */ - for (i = 0; i < (int)argc; i++) { - argv = &n->data.block.argv[i]; - - switch (argv->arg) { + for (i = 0; i < (int)n->args->argc; i++) + switch (n->args->argv[i].arg) { case (MDOC_Bullet): /* FALLTHROUGH */ case (MDOC_Dash): @@ -667,34 +770,28 @@ pre_bl(PRE_ARGS) /* FALLTHROUGH */ case (MDOC_Column): if (-1 == type) { - type = argv->arg; + type = n->args->argv[i].arg; break; } - return(mdoc_perr(mdoc, argv->line, argv->pos, - "multiple types specified")); + return(nerr(mdoc, n, EMULTILIST)); case (MDOC_Width): if (-1 == width) { - width = argv->arg; + width = n->args->argv[i].arg; break; } - return(mdoc_perr(mdoc, argv->line, argv->pos, - "multiple -%s arguments", - mdoc_argnames[MDOC_Width])); + return(nerr(mdoc, n, EARGREP)); case (MDOC_Offset): if (-1 == offset) { - offset = argv->arg; + offset = n->args->argv[i].arg; break; } - return(mdoc_perr(mdoc, argv->line, argv->pos, - "multiple -%s arguments", - mdoc_argnames[MDOC_Offset])); + return(nerr(mdoc, n, EARGREP)); default: break; } - } if (-1 == type) - return(mdoc_err(mdoc, "no type specified")); + return(nerr(mdoc, n, ELISTTYPE)); switch (type) { case (MDOC_Column): @@ -707,14 +804,13 @@ pre_bl(PRE_ARGS) if (-1 == width) break; return(mdoc_nwarn(mdoc, n, WARN_SYNTAX, - "superfluous -%s argument", + "superfluous %s argument", mdoc_argnames[MDOC_Width])); case (MDOC_Tag): - if (-1 == width && ! mdoc_nwarn(mdoc, n, WARN_SYNTAX, - "suggest -%s argument", - mdoc_argnames[MDOC_Width])) - return(0); - break; + if (-1 != width) + break; + return(mdoc_nerr(mdoc, n, "missing %s argument", + mdoc_argnames[MDOC_Width])); default: break; } @@ -726,22 +822,19 @@ pre_bl(PRE_ARGS) static int pre_bd(PRE_ARGS) { - int type, err, i; - struct mdoc_arg *argv; - size_t argc; + int i, type, err; if (MDOC_BLOCK != n->type) return(1); - - argc = n->data.block.argc; + if (NULL == n->args) + return(nerr(mdoc, n, EDISPTYPE)); /* Make sure that only one type of display is specified. */ /* LINTED */ - for (i = 0, err = type = 0; ! err && i < (int)argc; i++) { - argv = &n->data.block.argv[i]; - - switch (argv->arg) { + for (i = 0, err = type = 0; ! err && + i < (int)n->args->argc; i++) + switch (n->args->argv[i].arg) { case (MDOC_Ragged): /* FALLTHROUGH */ case (MDOC_Unfilled): @@ -753,16 +846,14 @@ pre_bd(PRE_ARGS) case (MDOC_File): if (0 == type++) break; - return(mdoc_perr(mdoc, argv->line, argv->pos, - "multiple types specified")); + return(nerr(mdoc, n, EMULTIDISP)); default: break; } - } if (type) return(1); - return(mdoc_err(mdoc, "no type specified")); + return(nerr(mdoc, n, EDISPTYPE)); } @@ -800,9 +891,9 @@ static int pre_an(PRE_ARGS) { - if (1 >= n->data.elem.argc) + if (n->args && 1 >= n->args->argc) return(1); - return(mdoc_nerr(mdoc, n, "one argument allowed")); + return(mdoc_nerr(mdoc, n, "only one argument allowed")); } @@ -863,12 +954,10 @@ pre_dt(PRE_ARGS) { if (0 == mdoc->meta.date || mdoc->meta.os) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "out-of-order prologue")) + if ( ! nwarn(mdoc, n, WPROLOOO)) return(0); if (mdoc->meta.title) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "prologue re-invoked")) + if ( ! nwarn(mdoc, n, WPROLREP)) return(0); return(1); } @@ -879,12 +968,10 @@ pre_os(PRE_ARGS) { if (NULL == mdoc->meta.title || 0 == mdoc->meta.date) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "out-of-order prologue")) + if ( ! nwarn(mdoc, n, WPROLOOO)) return(0); if (mdoc->meta.os) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "prologue re-invoked")) + if ( ! nwarn(mdoc, n, WPROLREP)) return(0); return(1); } @@ -895,12 +982,10 @@ pre_dd(PRE_ARGS) { if (mdoc->meta.title || mdoc->meta.os) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "out-of-order prologue")) + if ( ! nwarn(mdoc, n, WPROLOOO)) return(0); if (mdoc->meta.date) - if ( ! mdoc_nwarn(mdoc, n, WARN_COMPAT, - "prologue re-invoked")) + if ( ! nwarn(mdoc, n, WPROLREP)) return(0); return(1); } @@ -915,13 +1000,14 @@ post_bf(POST_ARGS) if (MDOC_BLOCK != mdoc->last->type) return(1); - head = mdoc->last->data.block.head; + head = mdoc->last->head; - if (0 == mdoc->last->data.block.argc) { - if (NULL == head->child) - return(mdoc_err(mdoc, "argument expected")); + if (NULL == mdoc->last->args) { + if (NULL == head->child || + MDOC_TEXT != head->child->type) + return(mdoc_err(mdoc, "text argument expected")); - p = head->child->data.text.string; + p = head->child->string; if (xstrcmp(p, "Em")) return(1); else if (xstrcmp(p, "Li")) @@ -932,11 +1018,9 @@ post_bf(POST_ARGS) } if (head->child) - return(mdoc_err(mdoc, "argument expected")); + return(mdoc_err(mdoc, "one argument expected")); - if (1 == mdoc->last->data.block.argc) - return(1); - return(mdoc_err(mdoc, "argument expected")); + return(1); } @@ -948,7 +1032,7 @@ post_nm(POST_ARGS) return(1); if (mdoc->meta.name) return(1); - return(mdoc_err(mdoc, "not yet invoked with name")); + return(merr(mdoc, ENAME)); } @@ -958,9 +1042,11 @@ post_at(POST_ARGS) if (NULL == mdoc->last->child) return(1); - if (mdoc_a2att(mdoc->last->child->data.text.string)) + if (MDOC_TEXT != mdoc->last->child->type) + return(merr(mdoc, EATT)); + if (mdoc_a2att(mdoc->last->child->string)) return(1); - return(mdoc_err(mdoc, "require valid AT&T symbol")); + return(merr(mdoc, EATT)); } @@ -968,15 +1054,15 @@ static int post_an(POST_ARGS) { - if (0 != mdoc->last->data.elem.argc) { + if (mdoc->last->args) { if (NULL == mdoc->last->child) return(1); - return(mdoc_err(mdoc, "argument(s) expected")); + return(merr(mdoc, ELINE)); } if (mdoc->last->child) return(1); - return(mdoc_err(mdoc, "argument(s) expected")); + return(merr(mdoc, ELINE)); } @@ -984,47 +1070,31 @@ static int post_ex(POST_ARGS) { - if (0 == mdoc->last->data.elem.argc) { - if (mdoc->last->child) - return(1); - return(mdoc_err(mdoc, "argument(s) expected")); - } - if (mdoc->last->child) - return(mdoc_err(mdoc, "argument(s) expected")); - if (1 != mdoc->last->data.elem.argc) - return(mdoc_err(mdoc, "argument(s) expected")); - if (MDOC_Std != mdoc->last->data.elem.argv[0].arg) - return(mdoc_err(mdoc, "argument(s) expected")); - - return(1); + if (mdoc->last->args) + return(1); + return(merr(mdoc, ELINE)); } static int post_it(POST_ARGS) { - int type, sv, i; -#define TYPE_NONE (0) -#define TYPE_BODY (1) -#define TYPE_HEAD (2) -#define TYPE_OHEAD (3) - size_t argc; - struct mdoc_node *n; + int type, i, cols; + struct mdoc_node *n, *c; if (MDOC_BLOCK != mdoc->last->type) return(1); n = mdoc->last->parent->parent; + if (NULL == n->args) + return(merr(mdoc, ELISTTYPE)); - argc = n->data.block.argc; - type = TYPE_NONE; - sv = -1; - /* Some types require block-head, some not. */ /* LINTED */ - for (i = 0; TYPE_NONE == type && i < (int)argc; i++) - switch (n->data.block.argv[i].arg) { + for (cols = type = -1, i = 0; -1 == type && + i < (int)n->args->argc; i++) + switch (n->args->argv[i].arg) { case (MDOC_Tag): /* FALLTHROUGH */ case (MDOC_Diag): @@ -1034,9 +1104,7 @@ post_it(POST_ARGS) case (MDOC_Ohang): /* FALLTHROUGH */ case (MDOC_Inset): - type = TYPE_HEAD; - sv = n->data.block.argv[i].arg; - break; + /* FALLTHROUGH */ case (MDOC_Bullet): /* FALLTHROUGH */ case (MDOC_Dash): @@ -1046,74 +1114,76 @@ post_it(POST_ARGS) case (MDOC_Hyphen): /* FALLTHROUGH */ case (MDOC_Item): - type = TYPE_BODY; - sv = n->data.block.argv[i].arg; + type = n->args->argv[i].arg; break; case (MDOC_Column): - type = TYPE_OHEAD; - sv = n->data.block.argv[i].arg; + type = n->args->argv[i].arg; + cols = (int)n->args->argv[i].sz; break; default: break; } - assert(TYPE_NONE != type); - - n = mdoc->last->data.block.head; + if (-1 == type) + return(merr(mdoc, ELISTTYPE)); - if (TYPE_HEAD == type) { - if (NULL == n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "argument(s) suggested")) + switch (type) { + case (MDOC_Tag): + if (NULL == mdoc->last->head->child) + if ( ! mwarn(mdoc, WLINE)) return(0); - - n = mdoc->last->data.block.body; - if (NULL == n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "multiline body suggested")) + break; + case (MDOC_Hang): + /* FALLTHROUGH */ + case (MDOC_Ohang): + /* FALLTHROUGH */ + case (MDOC_Inset): + /* FALLTHROUGH */ + case (MDOC_Diag): + if (NULL == mdoc->last->head->child) + if ( ! mwarn(mdoc, WLINE)) return(0); - - } else if (TYPE_BODY == type) { - if (n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "no argument suggested")) + if (NULL == mdoc->last->body->child) + if ( ! mwarn(mdoc, WMULTILINE)) return(0); - - n = mdoc->last->data.block.body; - if (NULL == n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "multiline body suggested")) + break; + case (MDOC_Bullet): + /* FALLTHROUGH */ + case (MDOC_Dash): + /* FALLTHROUGH */ + case (MDOC_Enum): + /* FALLTHROUGH */ + case (MDOC_Hyphen): + /* FALLTHROUGH */ + case (MDOC_Item): + if (mdoc->last->head->child) + if ( ! mwarn(mdoc, WNOLINE)) return(0); - } else { - if (NULL == n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "argument(s) suggested")) + if (NULL == mdoc->last->body->child) + if ( ! mwarn(mdoc, WMULTILINE)) return(0); - - n = mdoc->last->data.block.body; - if (n->child) - if ( ! mdoc_warn(mdoc, WARN_SYNTAX, - "no multiline body suggested")) + break; + case (MDOC_Column): + if (NULL == mdoc->last->head->child) + if ( ! mwarn(mdoc, WLINE)) return(0); + if (mdoc->last->body->child) + if ( ! mwarn(mdoc, WNOMULTILINE)) + return(0); + c = mdoc->last->head->child; + for (i = 0; c; c = c->next) + i++; + if (i == cols) + break; + if ( ! mdoc_warn(mdoc, WARN_SYNTAX, + "column mismatch (have %d, want %d)", i, cols)) + return(0); + break; + default: + break; } - if (MDOC_Column != sv) - return(1); - - argc = mdoc->last->parent->parent->data.block.argv->sz; - n = mdoc->last->data.block.head->child; - - for (i = 0; n; n = n->next) - i++; - - if (i == (int)argc) - return(1); - - return(mdoc_err(mdoc, "need %zu columns (have %d)", argc, i)); -#undef TYPE_NONE -#undef TYPE_BODY -#undef TYPE_HEAD -#undef TYPE_OHEAD + return(1); } @@ -1124,20 +1194,14 @@ post_bl(POST_ARGS) if (MDOC_BODY != mdoc->last->type) return(1); - if (NULL == (mdoc->last->child)) + if (NULL == mdoc->last->child) return(1); - /* - * Only allow `It' macros to be the immediate descendants of the - * `Bl' list. - */ - /* LINTED */ for (n = mdoc->last->child; n; n = n->next) { if (MDOC_BLOCK == n->type) if (MDOC_It == n->tok) continue; - return(mdoc_nerr(mdoc, n, "bad child of parent %s", mdoc_macronames[mdoc->last->tok])); } @@ -1155,16 +1219,16 @@ ebool(struct mdoc *mdoc) for (n = mdoc->last->child; n; n = n->next) { if (MDOC_TEXT != n->type) break; - if (xstrcmp(n->data.text.string, "on")) + if (xstrcmp(n->string, "on")) continue; - if (xstrcmp(n->data.text.string, "off")) + if (xstrcmp(n->string, "off")) continue; break; } if (NULL == n) return(1); - return(mdoc_nerr(mdoc, n, "expected boolean")); + return(nerr(mdoc, n, EBOOL)); } @@ -1173,16 +1237,14 @@ post_root(POST_ARGS) { if (NULL == mdoc->first->child) - return(mdoc_err(mdoc, "document lacks data")); + return(merr(mdoc, ENODATA)); if (SEC_PROLOGUE == mdoc->lastnamed) - return(mdoc_err(mdoc, "document lacks prologue")); + return(merr(mdoc, ENOPROLOGUE)); if (MDOC_BLOCK != mdoc->first->child->type) - return(mdoc_err(mdoc, "lacking post-prologue %s", - mdoc_macronames[MDOC_Sh])); + return(merr(mdoc, ENODATA)); if (MDOC_Sh != mdoc->first->child->tok) - return(mdoc_err(mdoc, "lacking post-prologue %s", - mdoc_macronames[MDOC_Sh])); + return(merr(mdoc, ENODATA)); return(1); } @@ -1192,10 +1254,9 @@ static int post_st(POST_ARGS) { - if (mdoc_a2st(mdoc->last->child->data.text.string)) + if (mdoc_a2st(mdoc->last->child->string)) return(1); - - return(mdoc_warn(mdoc, WARN_SYNTAX, "invalid standard")); + return(mwarn(mdoc, WBADSTAND)); } @@ -1227,28 +1288,20 @@ post_sh_body(POST_ARGS) */ if (NULL == (n = mdoc->last->child)) - return(mdoc_warn(mdoc, WARN_SYNTAX, - "section should have %s and %s", - mdoc_macronames[MDOC_Nm], - mdoc_macronames[MDOC_Nd])); + return(mwarn(mdoc, WNAMESECINC)); for ( ; n && n->next; n = n->next) { if (MDOC_ELEM == n->type && MDOC_Nm == n->tok) continue; if (MDOC_TEXT == n->type) continue; - if ( ! (mdoc_nwarn(mdoc, n, WARN_SYNTAX, - "section should have %s first", - mdoc_macronames[MDOC_Nm]))) + if ( ! mwarn(mdoc, WNAMESECINC)) return(0); } if (MDOC_ELEM == n->type && MDOC_Nd == n->tok) return(1); - - return(mdoc_warn(mdoc, WARN_SYNTAX, - "section should have %s last", - mdoc_macronames[MDOC_Nd])); + return(mwarn(mdoc, WNAMESECINC)); } @@ -1274,19 +1327,16 @@ post_sh_head(POST_ARGS) /* The NAME section should always be first. */ if (SEC_BODY == mdoc->lastnamed && SEC_NAME != sec) - return(mdoc_warn(mdoc, WARN_SYNTAX, - "section NAME should be first")); + return(mwarn(mdoc, WSECOOO)); if (SEC_CUSTOM == sec) return(1); /* Check for repeated or out-of-order sections. */ if (sec == mdoc->lastnamed) - return(mdoc_warn(mdoc, WARN_SYNTAX, - "section repeated")); + return(mwarn(mdoc, WSECREP)); if (sec < mdoc->lastnamed) - return(mdoc_warn(mdoc, WARN_SYNTAX, - "section out of conventional order")); + return(mwarn(mdoc, WSECOOO)); /* Check particular section/manual section conventions. */ @@ -1298,8 +1348,7 @@ post_sh_head(POST_ARGS) case (3): break; default: - return(mdoc_warn(mdoc, WARN_COMPAT, - "section in wrong manual section")); + return(mwarn(mdoc, WWRONGMSEC)); } break; default: @@ -103,7 +103,7 @@ xstrlcpys(char *buf, const struct mdoc_node *n, size_t sz) for ( ; n; n = n->next) { assert(MDOC_TEXT == n->type); - p = n->data.text.string; + p = n->string; if ( ! xstrlcat(buf, p, sz)) return(0); if (n->next && ! xstrlcat(buf, " ", sz)) |