=================================================================== RCS file: /cvs/docbook2mdoc/docbook2mdoc.c,v retrieving revision 1.23 retrieving revision 1.55 diff -u -p -r1.23 -r1.55 --- docbook2mdoc/docbook2mdoc.c 2014/03/30 17:46:17 1.23 +++ docbook2mdoc/docbook2mdoc.c 2019/03/22 16:35:41 1.55 @@ -1,6 +1,7 @@ -/* $Id: docbook2mdoc.c,v 1.23 2014/03/30 17:46:17 kristaps Exp $ */ +/* $Id: docbook2mdoc.c,v 1.55 2019/03/22 16:35:41 schwarze Exp $ */ /* * Copyright (c) 2014 Kristaps Dzonsons + * Copyright (c) 2019 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above @@ -37,11 +38,14 @@ struct parse { enum nodeid node; /* current (NODE_ROOT if pre-tree) */ const char *fname; /* filename */ int stop; /* should we stop now? */ +#define PARSE_EQN 1 + unsigned int flags; /* document-wide flags */ struct pnode *root; /* root of parse tree */ struct pnode *cur; /* current node in tree */ - char *b; /* nil-terminated buffer for pre-print */ + char *b; /* NUL-terminated buffer for pre-print */ size_t bsz; /* current length of b */ size_t mbsz; /* max bsz allocation */ + int level; /* header level, starting at 1 */ int newln; /* output: are we on a fresh line */ }; @@ -64,6 +68,7 @@ struct pattr { struct pnode { enum nodeid node; /* node type */ char *b; /* binary data buffer */ + char *real; /* store for "b" */ size_t bsz; /* data buffer size */ struct pnode *parent; /* parent (or NULL if top) */ struct pnodeq childq; /* queue of children */ @@ -73,7 +78,9 @@ struct pnode { static const char *attrkeys[ATTRKEY__MAX] = { "choice", + "close", "id", + "open", "rep" }; @@ -86,61 +93,140 @@ static const char *attrvals[ATTRVAL__MAX] = { }; static const struct node nodes[NODE__MAX] = { - { NULL, 0 }, - { "acronym", 0 }, - { "arg", 0 }, - { "citerefentry", NODE_IGNTEXT }, - { "cmdsynopsis", NODE_IGNTEXT }, - { "code", 0 }, - { "command", 0 }, - { "date", 0 }, + { NULL, 0 }, + { "acronym", 0 }, + { "affiliation", NODE_IGNTEXT }, + { "anchor", NODE_IGNTEXT }, + { "application", 0 }, + { "arg", 0 }, + { "author", NODE_IGNTEXT }, + { "authorgroup", NODE_IGNTEXT }, + { "blockquote", NODE_IGNTEXT }, + { "book", NODE_IGNTEXT }, + { "bookinfo", NODE_IGNTEXT }, + { "caution", NODE_IGNTEXT }, + { "chapter", NODE_IGNTEXT }, + { "citerefentry", NODE_IGNTEXT }, + { "citetitle", 0 }, + { "cmdsynopsis", NODE_IGNTEXT }, + { "code", 0 }, + { "colspec", NODE_IGNTEXT }, + { "command", 0 }, + { "constant", 0 }, + { "copyright", NODE_IGNTEXT }, + { "date", 0 }, + { "editor", NODE_IGNTEXT }, { "emphasis", 0 }, + { "entry", 0 }, { "envar", 0 }, + { "fieldsynopsis", NODE_IGNTEXT }, { "filename", 0 }, + { "firstname", 0 }, + { "firstterm", 0 }, + { "footnote", 0 }, { "funcdef", 0 }, - { "funcprototype", NODE_IGNTEXT }, - { "funcsynopsis", NODE_IGNTEXT }, - { "funcsynopsisinfo", 0 }, - { "function", 0 }, - { "itemizedlist", NODE_IGNTEXT }, - { "link", 0 }, - { "listitem", NODE_IGNTEXT }, - { "literal", 0 }, - { "manvolnum", 0 }, - { "option", 0 }, - { "orderedlist", NODE_IGNTEXT }, - { "para", 0 }, + { "funcprototype", NODE_IGNTEXT }, + { "funcsynopsis", NODE_IGNTEXT }, + { "funcsynopsisinfo", 0 }, + { "function", 0 }, + { "glossterm", 0 }, + { "group", NODE_IGNTEXT }, + { "holder", NODE_IGNTEXT }, + { "index", NODE_IGNTEXT }, + { "indexterm", NODE_IGNTEXT }, + { "info", NODE_IGNTEXT }, + { "informalequation", NODE_IGNTEXT }, + { "informaltable", NODE_IGNTEXT }, + { "inlineequation", NODE_IGNTEXT }, + { "itemizedlist", NODE_IGNTEXT }, + { "keysym", 0 }, + { "legalnotice", NODE_IGNTEXT }, + { "link", 0 }, + { "listitem", NODE_IGNTEXT }, + { "literal", 0 }, + { "literallayout", 0 }, + { "manvolnum", 0 }, + { "member", 0 }, + { "mml:math", NODE_IGNTEXT }, + { "mml:mfenced", 0 }, + { "mml:mfrac", 0 }, + { "mml:mi", 0 }, + { "mml:mn", 0 }, + { "mml:mo", 0 }, + { "mml:mrow", 0 }, + { "mml:msub", 0 }, + { "mml:msup", 0 }, + { "modifier", 0 }, + { "note", NODE_IGNTEXT }, + { "option", 0 }, + { "orderedlist", NODE_IGNTEXT }, + { "orgname", 0 }, + { "othername", 0 }, + { "para", 0 }, { "paramdef", 0 }, - { "parameter", 0 }, - { "programlisting", 0 }, - { "prompt", 0 }, - { "refclass", NODE_IGNTEXT }, - { "refdescriptor", NODE_IGNTEXT }, - { "refentry", NODE_IGNTEXT }, - { "refentryinfo", NODE_IGNTEXT }, - { "refentrytitle", 0 }, - { "refmeta", NODE_IGNTEXT }, - { "refmiscinfo", NODE_IGNTEXT }, - { "refname", 0 }, - { "refnamediv", NODE_IGNTEXT }, - { "refpurpose", 0 }, - { "refsect1", NODE_IGNTEXT }, - { "refsect2", NODE_IGNTEXT }, - { "refsynopsisdiv", NODE_IGNTEXT }, - { "replaceable", 0 }, - { "sbr", NODE_IGNTEXT }, - { "screen", NODE_IGNTEXT }, - { "structname", 0 }, - { "synopsis", 0 }, - { "term", 0 }, - { NULL, 0 }, - { "title", 0 }, - { "ulink", 0 }, - { "userinput", 0 }, - { "variablelist", NODE_IGNTEXT }, - { "varlistentry", NODE_IGNTEXT }, + { "parameter", 0 }, + { "part", NODE_IGNTEXT }, + { "phrase", 0 }, + { "preface", NODE_IGNTEXT }, + { "primary", 0 }, + { "programlisting", 0 }, + { "prompt", 0 }, + { "quote", 0 }, + { "refclass", NODE_IGNTEXT }, + { "refdescriptor", NODE_IGNTEXT }, + { "refentry", NODE_IGNTEXT }, + { "refentryinfo", NODE_IGNTEXT }, + { "refentrytitle", 0 }, + { "refmeta", NODE_IGNTEXT }, + { "refmetainfo", NODE_IGNTEXT }, + { "refmiscinfo", NODE_IGNTEXT }, + { "refname", 0 }, + { "refnamediv", NODE_IGNTEXT }, + { "refpurpose", 0 }, + { "refsect1", NODE_IGNTEXT }, + { "refsect2", NODE_IGNTEXT }, + { "refsect3", NODE_IGNTEXT }, + { "refsection", NODE_IGNTEXT }, + { "refsynopsisdiv", NODE_IGNTEXT }, + { "releaseinfo", 0 }, + { "replaceable", 0 }, + { "row", NODE_IGNTEXT }, + { "sbr", NODE_IGNTEXT }, + { "screen", NODE_IGNTEXT }, + { "secondary", 0 }, + { "sect1", NODE_IGNTEXT }, + { "sect2", NODE_IGNTEXT }, + { "section", NODE_IGNTEXT }, + { "sgmltag", 0 }, + { "simplelist", NODE_IGNTEXT }, + { "spanspec", NODE_IGNTEXT }, + { "structname", 0 }, + { "subtitle", 0 }, + { "surname", 0 }, + { "synopsis", 0 }, + { "table", NODE_IGNTEXT }, + { "tbody", NODE_IGNTEXT }, + { "term", 0 }, + { NULL, 0 }, + { "tfoot", NODE_IGNTEXT }, + { "tgroup", NODE_IGNTEXT }, + { "thead", NODE_IGNTEXT }, + { "tip", NODE_IGNTEXT }, + { "title", 0 }, + { "trademark", 0 }, + { "type", 0 }, + { "ulink", 0 }, + { "userinput", 0 }, + { "variablelist", NODE_IGNTEXT }, + { "varlistentry", NODE_IGNTEXT }, + { "varname", 0 }, + { "warning", NODE_IGNTEXT }, + { "wordasword", 0 }, + { "year", NODE_IGNTEXT }, }; +static int warn = 0; + static void pnode_print(struct parse *p, struct pnode *pn); @@ -175,7 +261,7 @@ xml_char(void *arg, const XML_Char *p, int sz) */ if (NODE_TEXT != ps->node) { for (i = 0; i < sz; i++) - if ( ! isspace((int)p[i])) + if ( ! isspace((unsigned char)p[i])) break; if (i == sz) return; @@ -198,7 +284,7 @@ xml_char(void *arg, const XML_Char *p, int sz) /* Append to current buffer. */ assert(sz >= 0); - ps->cur->b = realloc(ps->cur->b, + ps->cur->b = realloc(ps->cur->b, ps->cur->bsz + (size_t)sz); if (NULL == ps->cur->b) { perror(NULL); @@ -206,6 +292,7 @@ xml_char(void *arg, const XML_Char *p, int sz) } memcpy(ps->cur->b + ps->cur->bsz, p, sz); ps->cur->bsz += (size_t)sz; + ps->cur->real = ps->cur->b; } static void @@ -214,7 +301,7 @@ pnode_trim(struct pnode *pn) assert(NODE_TEXT == pn->node); for ( ; pn->bsz > 0; pn->bsz--) - if ( ! isspace((int)pn->b[pn->bsz - 1])) + if ( ! isspace((unsigned char)pn->b[pn->bsz - 1])) break; } @@ -243,7 +330,8 @@ xml_elem_start(void *arg, const XML_Char *name, const struct pattr *pattr; const XML_Char **att; - if (ps->stop) + /* FIXME: find a better way to ditch other namespaces. */ + if (ps->stop || 0 == strcmp(name, "xi:include")) return; /* Close out text node, if applicable... */ @@ -264,7 +352,7 @@ xml_elem_start(void *arg, const XML_Char *name, const if (NODE__MAX == node && NODE_ROOT == ps->node) { return; } else if (NODE__MAX == node) { - fprintf(stderr, "%s:%zu:%zu: unknown node \"%s\"\n", + fprintf(stderr, "%s:%zu:%zu: unknown node \"%s\"\n", ps->fname, XML_GetCurrentLineNumber(ps->xml), XML_GetCurrentColumnNumber(ps->xml), name); ps->stop = 1; @@ -275,21 +363,14 @@ xml_elem_start(void *arg, const XML_Char *name, const XML_GetCurrentColumnNumber(ps->xml)); ps->stop = 1; return; - } else if (NODE_ROOT == ps->node && NODE_REFENTRY != node) { + } else if (NODE_ROOT == ps->node && NODE_REFENTRY != node && + NODE_PART != node && NODE_BOOK != node) { return; - } else if ( ! isparent(node, ps->node)) { - fprintf(stderr, "%s:%zu:%zu: bad parent \"%s\" " - "of node \"%s\"\n", - ps->fname, XML_GetCurrentLineNumber(ps->xml), - XML_GetCurrentColumnNumber(ps->xml), - NULL == nodes[ps->node].name ? - "(none)" : nodes[ps->node].name, - NULL == nodes[node].name ? - "(none)" : nodes[node].name); - ps->stop = 1; - return; } + if (NODE_INLINEEQUATION == node) + ps->flags |= PARSE_EQN; + if (NULL == (dat = calloc(1, sizeof(struct pnode)))) { perror(NULL); exit(EXIT_FAILURE); @@ -315,31 +396,18 @@ xml_elem_start(void *arg, const XML_Char *name, const if (0 == strcmp(*att, attrkeys[key])) break; if (ATTRKEY__MAX == key) { - fprintf(stderr, "%s:%zu:%zu: unknown " - "attribute \"%s\"\n", ps->fname, - XML_GetCurrentLineNumber(ps->xml), - XML_GetCurrentColumnNumber(ps->xml), - *att); + if (warn) + fprintf(stderr, "%s:%zu:%zu: warning: " + "unknown attribute \"%s\"\n", + ps->fname, + XML_GetCurrentLineNumber(ps->xml), + XML_GetCurrentColumnNumber(ps->xml), + *att); continue; - } else if ( ! isattrkey(node, key)) { - fprintf(stderr, "%s:%zu:%zu: bad " - "attribute \"%s\"\n", ps->fname, - XML_GetCurrentLineNumber(ps->xml), - XML_GetCurrentColumnNumber(ps->xml), - *att); - continue; } for (val = 0; val < ATTRVAL__MAX; val++) if (0 == strcmp(*(att + 1), attrvals[val])) break; - if (ATTRVAL__MAX != val && ! isattrval(key, val)) { - fprintf(stderr, "%s:%zu:%zu: bad " - "value \"%s\"\n", ps->fname, - XML_GetCurrentLineNumber(ps->xml), - XML_GetCurrentColumnNumber(ps->xml), - *(att + 1)); - continue; - } pattr = calloc(1, sizeof(struct pattr)); pattr->key = key; pattr->val = val; @@ -360,8 +428,11 @@ xml_elem_end(void *arg, const XML_Char *name) { struct parse *ps = arg; + /* FIXME: find a better way to ditch other namespaces. */ if (ps->stop || NODE_ROOT == ps->node) return; + else if (0 == strcmp(name, "xi:include")) + return; /* Close out text node, if applicable... */ if (NODE_TEXT == ps->node) { @@ -401,7 +472,7 @@ pnode_free(struct pnode *pn) free(ap); } - free(pn->b); + free(pn->real); free(pn); } @@ -441,7 +512,7 @@ bufclear(struct parse *p) /* * Append NODE_TEXT contents to the current buffer, reallocating its * size if necessary. - * The buffer is ALWAYS nil-terminated. + * The buffer is ALWAYS NUL-terminated. */ static void bufappend(struct parse *p, struct pnode *pn) @@ -477,8 +548,28 @@ bufappend_r(struct parse *p, struct pnode *pn) bufappend_r(p, pp); } +/* + * Recursively search and return the first instance of "node". + */ +static struct pnode * +pnode_findfirst(struct pnode *pn, enum nodeid node) +{ + struct pnode *pp, *res; + + res = NULL; + TAILQ_FOREACH(pp, &pn->childq, child) { + res = pp->node == node ? pp : + pnode_findfirst(pp, node); + if (NULL != res) + break; + } + + return(res); +} + #define MACROLINE_NORM 0 #define MACROLINE_UPPER 1 +#define MACROLINE_NOWS 2 /* * Recursively print text presumably on a macro line. * Convert all whitespace to regular spaces. @@ -488,7 +579,7 @@ pnode_printmacrolinetext(struct parse *p, struct pnode { char *cp; - if (0 == p->newln) + if (0 == p->newln && ! (MACROLINE_NOWS & fl)) putchar(' '); bufclear(p); @@ -496,27 +587,27 @@ pnode_printmacrolinetext(struct parse *p, struct pnode /* Convert all space to spaces. */ for (cp = p->b; '\0' != *cp; cp++) - if (isspace((int)*cp)) + if (isspace((unsigned char)*cp)) *cp = ' '; - for (cp = p->b; isspace((int)*cp); cp++) + for (cp = p->b; isspace((unsigned char)*cp); cp++) /* Spin past whitespace (XXX: necessary?) */ ; for ( ; '\0' != *cp; cp++) { /* Escape us if we look like a macro. */ if ((cp == p->b || ' ' == *(cp - 1)) && - isupper((int)*cp) && - '\0' != *(cp + 1) && - islower((int)*(cp + 1)) && - ('\0' == *(cp + 2) || + isupper((unsigned char)*cp) && + '\0' != *(cp + 1) && + islower((unsigned char)*(cp + 1)) && + ('\0' == *(cp + 2) || ' ' == *(cp + 2) || - (islower((int)*(cp + 2)) && - ('\0' == *(cp + 3) || + (islower((unsigned char)*(cp + 2)) && + ('\0' == *(cp + 3) || ' ' == *(cp + 3))))) fputs("\\&", stdout); if (MACROLINE_UPPER & fl) - putchar(toupper((int)*cp)); + putchar(toupper((unsigned char)*cp)); else - putchar((int)*cp); + putchar(*cp); /* If we're a character escape, escape us. */ if ('\\' == *cp) putchar('e'); @@ -565,6 +656,72 @@ pnode_printmclose(struct parse *p, int sv) } /* + * Like pnode_printmclose() except we look to the next node, and, if + * found, see if it starts with punctuation. + * If it does, then we print that punctuation before the newline. + */ +static void +pnode_printmclosepunct(struct parse *p, struct pnode *pn, int sv) +{ + /* We wouldn't have done anything anyway. */ + if ( ! (sv && ! p->newln)) + return; + + /* No next node or it's not text. */ + if (NULL == (pn = TAILQ_NEXT(pn, child))) { + pnode_printmclose(p, sv); + return; + } else if (NODE_TEXT != pn->node) { + pnode_printmclose(p, sv); + return; + } + + /* Only do this for the comma/period. */ + if (pn->bsz > 0 && + (',' == pn->b[0] || '.' == pn->b[0]) && + (1 == pn->bsz || isspace((unsigned char)pn->b[1]))) { + putchar(' '); + putchar(pn->b[0]); + pn->b++; + pn->bsz--; + } + + putchar('\n'); + p->newln = 1; +} + +static void +pnode_printpara(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + assert(p->newln); + if (NULL == pn->parent || NODE_LISTITEM == pn->parent->node) + return; + + pp = TAILQ_PREV(pn, pnodeq, child); + if (NULL == pp) + pp = pn->parent; + switch (pp->node) { + case (NODE_CHAPTER): + case (NODE_PREFACE): + case (NODE_REFSECT1): + case (NODE_REFSECT2): + case (NODE_REFSECT3): + case (NODE_REFSECTION): + case (NODE_SECT1): + case (NODE_SECT2): + case (NODE_SECTION): + if (2 > p->level) + return; + break; + default: + break; + } + puts(".Pp"); +} + +/* * If the SYNOPSIS macro has a superfluous title, kill it. */ static void @@ -572,7 +729,7 @@ pnode_printrefsynopsisdiv(struct parse *p, struct pnod { struct pnode *pp; - TAILQ_FOREACH(pp, &pn->childq, child) + TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_TITLE == pp->node) { pnode_unlink(pp); return; @@ -586,28 +743,71 @@ static void pnode_printrefsect(struct parse *p, struct pnode *pn) { struct pnode *pp; + const char *title; + int flags, level; + level = ++p->level; + flags = 1 == level ? MACROLINE_UPPER : 0; + if (3 > level) { + switch (pn->node) { + case (NODE_CAUTION): + case (NODE_NOTE): + case (NODE_TIP): + case (NODE_WARNING): + level = 3; + break; + default: + break; + } + } + TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_TITLE == pp->node) break; - if (NODE_REFSECT1 == pn->node) + if (NULL == pp) { + switch (pn->node) { + case (NODE_PREFACE): + title = "Preface"; + break; + case (NODE_CAUTION): + title = "Caution"; + break; + case (NODE_NOTE): + title = "Note"; + break; + case (NODE_TIP): + title = "Tip"; + break; + case (NODE_WARNING): + title = "Warning"; + break; + default: + title = "Unknown"; + break; + } + } + + switch (level) { + case (1): fputs(".Sh", stdout); - else + break; + case (2): fputs(".Ss", stdout); + break; + default: + pnode_printpara(p, pn); + fputs(".Sy", stdout); + break; + } - p->newln = 0; - if (NULL != pp) { - pnode_printmacrolinetext(p, pp, - NODE_REFSECT1 == pn->node ? - MACROLINE_UPPER : 0); + p->newln = 0; + pnode_printmacrolinetext(p, pp, flags); pnode_printmclose(p, 1); pnode_unlink(pp); - } else { - puts("UNKNOWN"); - p->newln = 1; - } + } else + printf(" %s\n", title); } /* @@ -619,16 +819,12 @@ pnode_printciterefentry(struct parse *p, struct pnode struct pnode *pp, *title, *manvol; title = manvol = NULL; - assert(p->newln); TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_MANVOLNUM == pp->node) manvol = pp; else if (NODE_REFENTRYTITLE == pp->node) title = pp; - fputs(".Xr", stdout); - p->newln = 0; - if (NULL != title) { pnode_printmacrolinepart(p, title); } else @@ -638,7 +834,7 @@ pnode_printciterefentry(struct parse *p, struct pnode puts(" 1"); p->newln = 1; } else - pnode_printmacroline(p, manvol); + pnode_printmacrolinepart(p, manvol); } static void @@ -654,7 +850,6 @@ pnode_printrefmeta(struct parse *p, struct pnode *pn) else if (NODE_REFENTRYTITLE == pp->node) title = pp; - puts(".Dd $Mdocdate" "$"); fputs(".Dt", stdout); p->newln = 0; @@ -668,8 +863,6 @@ pnode_printrefmeta(struct parse *p, struct pnode *pn) p->newln = 1; } else pnode_printmacroline(p, manvol); - - puts(".Os"); } static void @@ -717,7 +910,7 @@ pnode_printparamdef(struct parse *p, struct pnode *pn) fputs(".Fa \"", stdout); p->newln = 0; if (NULL != ptype) { - pnode_printmacrolinepart(p, ptype); + pnode_printmacrolinetext(p, ptype, MACROLINE_NOWS); putchar(' '); } @@ -728,14 +921,83 @@ pnode_printparamdef(struct parse *p, struct pnode *pn) p->newln = 1; } +/* + * The node is a little peculiar. + * First, it can have arbitrary open and closing tokens, which default + * to parentheses. + * Second, >1 arguments are separated by commas. + */ static void +pnode_printmathfenced(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + struct pattr *ap; + + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ATTRKEY_OPEN == ap->key) { + printf("left %s ", ap->rawval); + break; + } + if (NULL == ap) + printf("left ( "); + + pp = TAILQ_FIRST(&pn->childq); + pnode_print(p, pp); + + while (NULL != (pp = TAILQ_NEXT(pp, child))) { + putchar(','); + pnode_print(p, pp); + } + + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ATTRKEY_CLOSE == ap->key) { + printf("right %s ", ap->rawval); + break; + } + if (NULL == ap) + printf("right ) "); +} + +/* + * These math nodes require special handling because they have infix + * syntax, instead of the usual prefix or prefix. + * So we need to break up the first and second child node with a + * particular eqn(7) word. + */ +static void +pnode_printmath(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + pp = TAILQ_FIRST(&pn->childq); + pnode_print(p, pp); + + switch (pn->node) { + case (NODE_MML_MSUP): + fputs(" sup ", stdout); + break; + case (NODE_MML_MFRAC): + fputs(" over ", stdout); + break; + case (NODE_MML_MSUB): + fputs(" sub ", stdout); + break; + default: + break; + } + + pp = TAILQ_NEXT(pp, child); + pnode_print(p, pp); +} + +static void pnode_printfuncprototype(struct parse *p, struct pnode *pn) { struct pnode *pp, *fdef; assert(p->newln); TAILQ_FOREACH(fdef, &pn->childq, child) - if (NODE_FUNCDEF == fdef->node) + if (NODE_FUNCDEF == fdef->node) break; if (NULL != fdef) @@ -743,7 +1005,7 @@ pnode_printfuncprototype(struct parse *p, struct pnode else puts(".Fo UNKNOWN"); - TAILQ_FOREACH(pp, &pn->childq, child) + TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_PARAMDEF == pp->node) pnode_printparamdef(p, pp); @@ -751,7 +1013,7 @@ pnode_printfuncprototype(struct parse *p, struct pnode p->newln = 1; } -/* +/* * The element is more complicated than it should be because text * nodes are treated like ".Ar foo", but non-text nodes need to be * re-sent into the printer (i.e., without the preceding ".Ar"). @@ -767,10 +1029,10 @@ pnode_printarg(struct parse *p, struct pnode *pn) isop = 1; isrep = 0; - TAILQ_FOREACH(ap, &pn->attrq, child) + TAILQ_FOREACH(ap, &pn->attrq, child) if (ATTRKEY_CHOICE == ap->key && (ATTRVAL_PLAIN == ap->val || - ATTRVAL_REQ == ap->val)) + ATTRVAL_REQ == ap->val)) isop = 0; else if (ATTRKEY_REP == ap->key && (ATTRVAL_REPEAT == ap->val)) @@ -785,90 +1047,185 @@ pnode_printarg(struct parse *p, struct pnode *pn) if (NODE_TEXT == pp->node) { pnode_printmopen(p); fputs("Ar", stdout); - } + } pnode_print(p, pp); - if (NODE_TEXT == pp->node && isrep) + if (NODE_TEXT == pp->node && isrep) fputs("...", stdout); } } -/* - * Recursively search and return the first instance of "node". - */ -static struct pnode * -pnode_findfirst(struct pnode *pn, enum nodeid node) +static void +pnode_printgroup(struct parse *p, struct pnode *pn) { - struct pnode *pp, *res; + struct pnode *pp, *np; + struct pattr *ap; + int isop, sv; - res = NULL; - TAILQ_FOREACH(pp, &pn->childq, child) { - res = pp->node == node ? pp : - pnode_findfirst(pp, node); - if (NULL != res) + isop = 1; + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ATTRKEY_CHOICE == ap->key && + (ATTRVAL_PLAIN == ap->val || + ATTRVAL_REQ == ap->val)) { + isop = 0; break; + } + + /* + * Make sure we're on a macro line. + * This will prevent pnode_print() for putting us on a + * subsequent line. + */ + sv = p->newln; + pnode_printmopen(p); + if (isop) + fputs("Op", stdout); + else if (sv) + fputs("No", stdout); + + /* + * Keep on printing text separated by the vertical bar as long + * as we're within the same origin node as the group. + * This is kind of a nightmare. + * Eh, DocBook... + * FIXME: if there's a "Fl", we don't cut off the leading "-" + * like we do in pnode_print(). + */ + TAILQ_FOREACH(pp, &pn->childq, child) { + pnode_print(p, pp); + np = TAILQ_NEXT(pp, child); + while (NULL != np) { + if (pp->node != np->node) + break; + fputs(" |", stdout); + pnode_printmacrolinepart(p, np); + pp = np; + np = TAILQ_NEXT(np, child); + } } - return(res); + pnode_printmclose(p, sv); } static void pnode_printprologue(struct parse *p, struct pnode *pn) { struct pnode *pp; + struct pattr *ap; + const char *name; pp = NULL == p->root ? NULL : pnode_findfirst(p->root, NODE_REFMETA); + puts(".Dd $Mdocdate" "$"); if (NULL != pp) { pnode_printrefmeta(p, pp); pnode_unlink(pp); } else { - puts(".\\\" Supplying bogus prologue..."); - puts(".Dd $Mdocdate" "$"); - puts(".Dt UNKNOWN 1"); - puts(".Os"); + name = "UNKNOWN"; + TAILQ_FOREACH(ap, &p->root->attrq, child) { + if (ATTRKEY_ID == ap->key) { + name = ap->rawval; + break; + } + } + printf(".Dt %s 1\n", name); } + puts(".Os"); + + if (PARSE_EQN & p->flags) { + puts(".EQ"); + puts("delim $$"); + puts(".EN"); + } } +/* + * We can have multiple elements within a , which + * we should comma-separate as list headers. + */ static void pnode_printvarlistentry(struct parse *p, struct pnode *pn) { struct pnode *pp; + int first = 1; assert(p->newln); + fputs(".It", stdout); + p->newln = 0; + TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_TERM == pp->node) { - fputs(".It", stdout); - p->newln = 0; + if ( ! first) + putchar(','); pnode_print(p, pp); pnode_unlink(pp); - pnode_printmclose(p, 1); - return; - } + first = 0; + } else + break; - puts(".It"); + putchar('\n'); p->newln = 1; } static void -pnode_printitemizedlist(struct parse *p, struct pnode *pn) +pnode_printrow(struct parse *p, struct pnode *pn) { struct pnode *pp; + puts(".Bl -dash -compact"); + + TAILQ_FOREACH(pp, &pn->childq, child) { + assert(p->newln); + puts(".It"); + pnode_print(p, pp); + pnode_printmclose(p, 1); + } assert(p->newln); + puts(".El"); +} + +static void +pnode_printtable(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + assert(p->newln); TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_TITLE == pp->node) { - puts(".Pp"); + pnode_printpara(p, pp); pnode_print(p, pp); pnode_unlink(pp); } + assert(p->newln); + puts(".Bl -ohang"); + while (NULL != (pp = pnode_findfirst(pn, NODE_ROW))) { + puts(".It Table Row"); + pnode_printrow(p, pp); + pnode_printmclose(p, 1); + pnode_unlink(pp); + } + assert(p->newln); + puts(".El"); +} +static void +pnode_printlist(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + assert(p->newln); + TAILQ_FOREACH(pp, &pn->childq, child) + if (NODE_TITLE == pp->node) { + pnode_printpara(p, pp); + pnode_print(p, pp); + pnode_unlink(pp); + } + assert(p->newln); if (NODE_ORDEREDLIST == pn->node) puts(".Bl -enum"); else - puts(".Bl -item"); + puts(".Bl -bullet"); TAILQ_FOREACH(pp, &pn->childq, child) { assert(p->newln); @@ -888,7 +1245,7 @@ pnode_printvariablelist(struct parse *p, struct pnode assert(p->newln); TAILQ_FOREACH(pp, &pn->childq, child) if (NODE_TITLE == pp->node) { - puts(".Pp"); + pnode_printpara(p, pp); pnode_print(p, pp); pnode_unlink(pp); } @@ -926,13 +1283,34 @@ pnode_print(struct parse *p, struct pnode *pn) sv = p->newln; + /* XXX fprintf(stderr, "NODE %s\n", nodes[pn->node].name); */ switch (pn->node) { + case (NODE_APPLICATION): + pnode_printmopen(p); + fputs("Nm", stdout); + break; + case (NODE_ANCHOR): + /* Don't print anything! */ + return; case (NODE_ARG): pnode_printarg(p, pn); pnode_unlinksub(pn); break; - case (NODE_CITEREFENTRY): + case (NODE_AUTHOR): + pnode_printmopen(p); + fputs("An", stdout); + break; + case (NODE_AUTHORGROUP): assert(p->newln); + puts(".An -split"); + break; + case (NODE_BOOKINFO): + assert(p->newln); + puts(".Sh NAME"); + break; + case (NODE_CITEREFENTRY): + pnode_printmopen(p); + fputs("Xr", stdout); pnode_printciterefentry(p, pn); pnode_unlinksub(pn); break; @@ -944,6 +1322,15 @@ pnode_print(struct parse *p, struct pnode *pn) pnode_printmopen(p); fputs("Nm", stdout); break; + case (NODE_CONSTANT): + pnode_printmopen(p); + fputs("Dv", stdout); + break; + case (NODE_EDITOR): + puts("editor: "); + pnode_printmopen(p); + fputs("An", stdout); + break; case (NODE_EMPHASIS): pnode_printmopen(p); fputs("Em", stdout); @@ -969,35 +1356,79 @@ pnode_print(struct parse *p, struct pnode *pn) pnode_printmopen(p); fputs("Fd", stdout); break; + case (NODE_INDEXTERM): + return; + case (NODE_INFORMALEQUATION): + if ( ! p->newln) + putchar('\n'); + puts(".EQ"); + p->newln = 0; + break; + case (NODE_INLINEEQUATION): + fputc('$', stdout); + p->newln = 0; + break; case (NODE_ITEMIZEDLIST): - /* FALLTHROUGH */ - case (NODE_ORDEREDLIST): assert(p->newln); - pnode_printitemizedlist(p, pn); + pnode_printlist(p, pn); + pnode_unlinksub(pn); break; + case (NODE_GROUP): + pnode_printgroup(p, pn); + pnode_unlinksub(pn); + break; + case (NODE_LEGALNOTICE): + assert(p->newln); + puts(".Sh LEGAL NOTICE"); + break; case (NODE_LITERAL): pnode_printmopen(p); fputs("Li", stdout); break; + case (NODE_MML_MFENCED): + pnode_printmathfenced(p, pn); + pnode_unlinksub(pn); + break; + case (NODE_MML_MROW): + case (NODE_MML_MI): + case (NODE_MML_MN): + case (NODE_MML_MO): + if (TAILQ_EMPTY(&pn->childq)) + break; + fputs(" { ", stdout); + break; + case (NODE_MML_MFRAC): + case (NODE_MML_MSUB): + case (NODE_MML_MSUP): + pnode_printmath(p, pn); + pnode_unlinksub(pn); + break; case (NODE_OPTION): pnode_printmopen(p); fputs("Fl", stdout); break; - case (NODE_PARA): + case (NODE_ORDEREDLIST): assert(p->newln); - if (NULL != pn->parent && - NODE_LISTITEM == pn->parent->node) - break; - puts(".Pp"); + pnode_printlist(p, pn); + pnode_unlinksub(pn); break; + case (NODE_PARA): + pnode_printpara(p, pn); + break; case (NODE_PARAMETER): /* Suppress non-text children... */ pnode_printmopen(p); fputs("Fa \"", stdout); - pnode_printmacrolinepart(p, pn); - puts("\""); + pnode_printmacrolinetext(p, pn, MACROLINE_NOWS); + fputs("\"", stdout); pnode_unlinksub(pn); break; + case (NODE_QUOTE): + pnode_printmopen(p); + fputs("Qo", stdout); + break; + case (NODE_LITERALLAYOUT): + /* FALLTHROUGH */ case (NODE_PROGRAMLISTING): /* FALLTHROUGH */ case (NODE_SCREEN): @@ -1033,9 +1464,19 @@ pnode_print(struct parse *p, struct pnode *pn) pnode_printrefsynopsisdiv(p, pn); puts(".Sh SYNOPSIS"); break; + case (NODE_PREFACE): case (NODE_REFSECT1): - /* FALLTHROUGH */ case (NODE_REFSECT2): + case (NODE_REFSECT3): + case (NODE_REFSECTION): + case (NODE_CHAPTER): + case (NODE_SECT1): + case (NODE_SECT2): + case (NODE_SECTION): + case (NODE_NOTE): + case (NODE_TIP): + case (NODE_CAUTION): + case (NODE_WARNING): assert(p->newln); pnode_printrefsect(p, pn); break; @@ -1047,23 +1488,42 @@ pnode_print(struct parse *p, struct pnode *pn) assert(p->newln); puts(".br"); break; + case (NODE_SGMLTAG): + pnode_printmopen(p); + fputs("Li", stdout); + break; case (NODE_STRUCTNAME): pnode_printmopen(p); fputs("Vt", stdout); break; + case (NODE_TABLE): + /* FALLTHROUGH */ + case (NODE_INFORMALTABLE): + assert(p->newln); + pnode_printtable(p, pn); + pnode_unlinksub(pn); + break; case (NODE_TEXT): if (0 == p->newln) putchar(' '); + bufclear(p); bufappend(p, pn); + + if (0 == p->bsz) { + assert(pn->real != pn->b); + break; + } + /* * Output all characters, squeezing out whitespace - * between newlines. + * between newlines. * XXX: all whitespace, including tabs (?). * Remember to escape control characters and escapes. */ assert(p->bsz); cp = p->b; + /* * There's often a superfluous "-" in its