=================================================================== RCS file: /cvs/docbook2mdoc/docbook2mdoc.c,v retrieving revision 1.6 retrieving revision 1.72 diff -u -p -r1.6 -r1.72 --- docbook2mdoc/docbook2mdoc.c 2014/03/28 10:08:24 1.6 +++ docbook2mdoc/docbook2mdoc.c 2019/03/25 17:28:32 1.72 @@ -1,6 +1,7 @@ -/* $Id: docbook2mdoc.c,v 1.6 2014/03/28 10:08:24 kristaps Exp $ */ +/* $Id: docbook2mdoc.c,v 1.72 2019/03/25 17:28:32 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 @@ -24,44 +25,14 @@ #include #include #include +#include -/* - * All recognised node types. - */ -enum nodeid { - NODE_ROOT = 0, /* Must comes first. */ - /* Alpha-ordered hereafter. */ - NODE_ARG, - NODE_CITEREFENTRY, - NODE_CMDSYNOPSIS, - NODE_CODE, - NODE_COMMAND, - NODE_FUNCDEF, - NODE_FUNCPROTOTYPE, - NODE_FUNCSYNOPSIS, - NODE_FUNCSYNOPSISINFO, - NODE_FUNCTION, - NODE_MANVOLNUM, - NODE_OPTION, - NODE_PARA, - NODE_PARAMDEF, - NODE_PARAMETER, - NODE_PROGRAMLISTING, - NODE_REFCLASS, - NODE_REFDESCRIPTOR, - NODE_REFENTRY, - NODE_REFENTRYTITLE, - NODE_REFMETA, - NODE_REFMISCINFO, - NODE_REFNAME, - NODE_REFNAMEDIV, - NODE_REFPURPOSE, - NODE_REFSECT1, - NODE_REFSYNOPSISDIV, - NODE_SYNOPSIS, - NODE_TEXT, - NODE_TITLE, - NODE__MAX +#include "extern.h" + +enum linestate { + LINE_NEW = 0, + LINE_TEXT, + LINE_MACRO }; /* @@ -69,302 +40,209 @@ enum nodeid { * Keep this as simple and small as possible. */ struct parse { + XML_Parser xml; 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 */ - unsigned int flags; -#define PARSE_HAS_META 1 - char *b; - size_t bsz; - size_t mbsz; + int level; /* header level, starting at 1 */ + enum linestate linestate; }; struct node { - const char *name; - unsigned int flags; -#define NODE_IGNTEXT 1 /* ignore all contained text */ + const char *name; /* docbook element name */ + enum nodeid node; /* docbook element to generate */ }; TAILQ_HEAD(pnodeq, pnode); +TAILQ_HEAD(pattrq, pattr); +struct pattr { + enum attrkey key; + enum attrval val; + char *rawval; + TAILQ_ENTRY(pattr) child; +}; + 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 */ + struct pattrq attrq; /* attributes of node */ TAILQ_ENTRY(pnode) child; }; -static const struct node nodes[NODE__MAX] = { - { NULL, 0 }, - { "arg", 0 }, - { "citerefentry", NODE_IGNTEXT }, - { "cmdsynopsis", NODE_IGNTEXT }, - { "code", 0 }, - { "command", 0 }, - { "funcdef", 0 }, - { "funcprototype", NODE_IGNTEXT }, - { "funcsynopsis", NODE_IGNTEXT }, - { "funcsynopsisinfo", 0 }, - { "function", 0 }, - { "manvolnum", 0 }, - { "option", 0 }, - { "para", 0 }, - { "paramdef", 0 }, - { "parameter", 0 }, - { "programlisting", 0 }, - { "refclass", NODE_IGNTEXT }, - { "refdescriptor", NODE_IGNTEXT }, - { "refentry", NODE_IGNTEXT }, - { "refentrytitle", 0 }, - { "refmeta", NODE_IGNTEXT }, - { "refmiscinfo", NODE_IGNTEXT }, - { "refname", 0 }, - { "refnamediv", NODE_IGNTEXT }, - { "refpurpose", 0 }, - { "refsect1", 0 }, - { "refsynopsisdiv", NODE_IGNTEXT }, - { "synopsis", 0 }, - { NULL, 0 }, - { "title", 0 }, +static const char *attrkeys[ATTRKEY__MAX] = { + "choice", + "class", + "close", + "id", + "linkend", + "open", + "rep" }; -/* - * Look up whether "parent" is a valid parent for "node". - */ -static int -isparent(enum nodeid node, enum nodeid parent) -{ +static const char *attrvals[ATTRVAL__MAX] = { + "monospaced", + "norepeat", + "opt", + "plain", + "repeat", + "req" +}; - switch (node) { - case (NODE_ROOT): - return(0); - case (NODE_ARG): - switch (parent) { - case (NODE_ARG): - case (NODE_CMDSYNOPSIS): - return(1); - default: - break; - } - return(0); - case (NODE_CITEREFENTRY): - switch (parent) { - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_CMDSYNOPSIS): - switch (parent) { - case (NODE_PARA): - case (NODE_REFSECT1): - case (NODE_REFSYNOPSISDIV): - return(1); - default: - break; - } - return(0); - case (NODE_CODE): - switch (parent) { - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_COMMAND): - switch (parent) { - case (NODE_CMDSYNOPSIS): - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_FUNCDEF): - return(NODE_FUNCPROTOTYPE == parent); - case (NODE_FUNCPROTOTYPE): - return(NODE_FUNCSYNOPSIS == parent); - case (NODE_FUNCSYNOPSIS): - switch (parent) { - case (NODE_PARA): - case (NODE_REFSECT1): - case (NODE_REFSYNOPSISDIV): - return(1); - default: - break; - } - return(0); - case (NODE_FUNCSYNOPSISINFO): - return(NODE_FUNCSYNOPSIS == parent); - case (NODE_FUNCTION): - switch (parent) { - case (NODE_CODE): - case (NODE_FUNCDEF): - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_MANVOLNUM): - switch (parent) { - case (NODE_CITEREFENTRY): - case (NODE_REFMETA): - return(1); - default: - break; - } - return(0); - case (NODE_OPTION): - switch (parent) { - case (NODE_ARG): - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_PARA): - switch (parent) { - case (NODE_REFSECT1): - case (NODE_REFSYNOPSISDIV): - return(1); - default: - break; - } - return(0); - case (NODE_PARAMDEF): - return(NODE_FUNCPROTOTYPE == parent); - case (NODE_PARAMETER): - switch (parent) { - case (NODE_CODE): - case (NODE_FUNCSYNOPSISINFO): - case (NODE_PARA): - case (NODE_PARAMDEF): - case (NODE_PROGRAMLISTING): - case (NODE_REFDESCRIPTOR): - case (NODE_REFENTRYTITLE): - case (NODE_REFNAME): - case (NODE_REFPURPOSE): - case (NODE_SYNOPSIS): - case (NODE_TITLE): - return(1); - default: - break; - } - return(0); - case (NODE_PROGRAMLISTING): - switch (parent) { - case (NODE_PARA): - case (NODE_REFSECT1): - case (NODE_REFSYNOPSISDIV): - return(1); - default: - break; - } - return(0); - case (NODE_REFCLASS): - return(parent == NODE_REFNAMEDIV); - case (NODE_REFDESCRIPTOR): - return(parent == NODE_REFNAMEDIV); - case (NODE_REFENTRY): - return(parent == NODE_ROOT); - case (NODE_REFENTRYTITLE): - switch (parent) { - case (NODE_CITEREFENTRY): - case (NODE_REFMETA): - return(1); - default: - break; - } - case (NODE_REFMETA): - return(parent == NODE_REFENTRY); - case (NODE_REFMISCINFO): - return(parent == NODE_REFMETA); - case (NODE_REFNAME): - return(parent == NODE_REFNAMEDIV); - case (NODE_REFNAMEDIV): - return(parent == NODE_REFENTRY); - case (NODE_REFPURPOSE): - return(parent == NODE_REFNAMEDIV); - case (NODE_REFSECT1): - return(parent == NODE_REFENTRY); - case (NODE_REFSYNOPSISDIV): - return(parent == NODE_REFENTRY); - case (NODE_SYNOPSIS): - switch (parent) { - case (NODE_REFSYNOPSISDIV): - case (NODE_REFSECT1): - return(1); - default: - break; - } - return(0); - case (NODE_TITLE): - switch (parent) { - case (NODE_REFSECT1): - case (NODE_REFSYNOPSISDIV): - return(1); - default: - break; - } - return(0); - case (NODE_TEXT): - return(1); - case (NODE__MAX): - break; - } +static const struct node nodes[] = { + { "acronym", NODE_ACRONYM }, + { "affiliation", NODE_AFFILIATION }, + { "anchor", NODE_ANCHOR }, + { "application", NODE_APPLICATION }, + { "arg", NODE_ARG }, + { "author", NODE_AUTHOR }, + { "authorgroup", NODE_AUTHORGROUP }, + { "blockquote", NODE_BLOCKQUOTE }, + { "book", NODE_BOOK }, + { "bookinfo", NODE_BOOKINFO }, + { "caution", NODE_CAUTION }, + { "chapter", NODE_SECTION }, + { "citerefentry", NODE_CITEREFENTRY }, + { "citetitle", NODE_CITETITLE }, + { "cmdsynopsis", NODE_CMDSYNOPSIS }, + { "code", NODE_CODE }, + { "colspec", NODE_COLSPEC }, + { "command", NODE_COMMAND }, + { "constant", NODE_CONSTANT }, + { "copyright", NODE_COPYRIGHT }, + { "date", NODE_DATE }, + { "editor", NODE_EDITOR }, + { "email", NODE_EMAIL }, + { "emphasis", NODE_EMPHASIS }, + { "entry", NODE_ENTRY }, + { "envar", NODE_ENVAR }, + { "fieldsynopsis", NODE_FIELDSYNOPSIS }, + { "filename", NODE_FILENAME }, + { "firstname", NODE_FIRSTNAME }, + { "firstterm", NODE_FIRSTTERM }, + { "footnote", NODE_FOOTNOTE }, + { "funcdef", NODE_FUNCDEF }, + { "funcprototype", NODE_FUNCPROTOTYPE }, + { "funcsynopsis", NODE_FUNCSYNOPSIS }, + { "funcsynopsisinfo", NODE_FUNCSYNOPSISINFO }, + { "function", NODE_FUNCTION }, + { "glossterm", NODE_GLOSSTERM }, + { "group", NODE_GROUP }, + { "holder", NODE_HOLDER }, + { "index", NODE_INDEX }, + { "indexterm", NODE_INDEXTERM }, + { "info", NODE_INFO }, + { "informalequation", NODE_INFORMALEQUATION }, + { "informaltable", NODE_INFORMALTABLE }, + { "inlineequation", NODE_INLINEEQUATION }, + { "itemizedlist", NODE_ITEMIZEDLIST }, + { "keysym", NODE_KEYSYM }, + { "legalnotice", NODE_LEGALNOTICE }, + { "link", NODE_LINK }, + { "listitem", NODE_LISTITEM }, + { "literal", NODE_LITERAL }, + { "literallayout", NODE_LITERALLAYOUT }, + { "manvolnum", NODE_MANVOLNUM }, + { "member", NODE_MEMBER }, + { "mml:math", NODE_MML_MATH }, + { "mml:mfenced", NODE_MML_MFENCED }, + { "mml:mfrac", NODE_MML_MFRAC }, + { "mml:mi", NODE_MML_MI }, + { "mml:mn", NODE_MML_MN }, + { "mml:mo", NODE_MML_MO }, + { "mml:mrow", NODE_MML_MROW }, + { "mml:msub", NODE_MML_MSUB }, + { "mml:msup", NODE_MML_MSUP }, + { "modifier", NODE_MODIFIER }, + { "note", NODE_NOTE }, + { "option", NODE_OPTION }, + { "orderedlist", NODE_ORDEREDLIST }, + { "orgname", NODE_ORGNAME }, + { "othername", NODE_OTHERNAME }, + { "para", NODE_PARA }, + { "paramdef", NODE_PARAMDEF }, + { "parameter", NODE_PARAMETER }, + { "part", NODE_SECTION }, + { "personname", NODE_PERSONNAME }, + { "phrase", NODE_PHRASE }, + { "preface", NODE_PREFACE }, + { "primary", NODE_PRIMARY }, + { "programlisting", NODE_PROGRAMLISTING }, + { "prompt", NODE_PROMPT }, + { "quote", NODE_QUOTE }, + { "refclass", NODE_REFCLASS }, + { "refdescriptor", NODE_REFDESCRIPTOR }, + { "refentry", NODE_REFENTRY }, + { "refentryinfo", NODE_REFENTRYINFO }, + { "refentrytitle", NODE_REFENTRYTITLE }, + { "refmeta", NODE_REFMETA }, + { "refmetainfo", NODE_REFMETAINFO }, + { "refmiscinfo", NODE_REFMISCINFO }, + { "refname", NODE_REFNAME }, + { "refnamediv", NODE_REFNAMEDIV }, + { "refpurpose", NODE_REFPURPOSE }, + { "refsect1", NODE_SECTION }, + { "refsect2", NODE_SECTION }, + { "refsect3", NODE_SECTION }, + { "refsection", NODE_SECTION }, + { "refsynopsisdiv", NODE_REFSYNOPSISDIV }, + { "releaseinfo", NODE_RELEASEINFO }, + { "replaceable", NODE_REPLACEABLE }, + { "row", NODE_ROW }, + { "sbr", NODE_SBR }, + { "screen", NODE_SCREEN }, + { "secondary", NODE_SECONDARY }, + { "sect1", NODE_SECTION }, + { "sect2", NODE_SECTION }, + { "section", NODE_SECTION }, + { "sgmltag", NODE_SGMLTAG }, + { "simplelist", NODE_SIMPLELIST }, + { "spanspec", NODE_SPANSPEC }, + { "structname", NODE_STRUCTNAME }, + { "subtitle", NODE_SUBTITLE }, + { "surname", NODE_SURNAME }, + { "synopsis", NODE_SYNOPSIS }, + { "table", NODE_TABLE }, + { "tbody", NODE_TBODY }, + { "term", NODE_TERM }, + { "tfoot", NODE_TFOOT }, + { "tgroup", NODE_TGROUP }, + { "thead", NODE_THEAD }, + { "tip", NODE_TIP }, + { "title", NODE_TITLE }, + { "trademark", NODE_TRADEMARK }, + { "type", NODE_TYPE }, + { "ulink", NODE_ULINK }, + { "userinput", NODE_USERINPUT }, + { "variablelist", NODE_VARIABLELIST }, + { "varlistentry", NODE_VARLISTENTRY }, + { "varname", NODE_VARNAME }, + { "warning", NODE_WARNING }, + { "wordasword", NODE_WORDASWORD }, + { "year", NODE_YEAR }, + { NULL, NODE__MAX } +}; - abort(); - return(0); -} +static int warn = 0; static void +pnode_print(struct parse *p, struct pnode *pn); + +/* + * Process a stream of characters. + * We store text as nodes in and of themselves. + * If a text node is already open, append to it. + * If it's not open, open one under the current context. + */ +static void xml_char(void *arg, const XML_Char *p, int sz) { struct parse *ps = arg; @@ -372,129 +250,172 @@ xml_char(void *arg, const XML_Char *p, int sz) int i; /* Stopped or no tree yet. */ - if (ps->stop || NODE_ROOT == ps->node) + if (ps->stop || ps->node == NODE_ROOT) return; - /* Not supposed to be collecting text. */ - assert(NULL != ps->cur); - if (NODE_IGNTEXT & nodes[ps->node].flags) - return; + assert(ps->cur != NULL); /* * Are we in the midst of processing text? * If we're not processing text right now, then create a text * node for doing so. * However, don't do so unless we have some non-whitespace to - * process! + * process: strip out all leading whitespace to be sure. */ - if (NODE_TEXT != ps->node) { + if (ps->node != NODE_TEXT) { for (i = 0; i < sz; i++) - if ( ! isspace((int)p[i])) + if ( ! isspace((unsigned char)p[i])) break; if (i == sz) return; - dat = calloc(1, sizeof(struct pnode)); - if (NULL == dat) { + p += i; + sz -= i; + dat = calloc(1, sizeof(*dat)); + if (dat == NULL) { perror(NULL); - exit(EXIT_FAILURE); + exit(1); } dat->node = ps->node = NODE_TEXT; dat->parent = ps->cur; TAILQ_INIT(&dat->childq); + TAILQ_INIT(&dat->attrq); TAILQ_INSERT_TAIL(&ps->cur->childq, dat, child); ps->cur = dat; - assert(NULL != ps->root); - + assert(ps->root != NULL); } /* Append to current buffer. */ assert(sz >= 0); - ps->cur->b = realloc(ps->cur->b, - ps->cur->bsz + (size_t)sz); - if (NULL == ps->cur->b) { + ps->cur->b = realloc(ps->cur->b, ps->cur->bsz + sz + 1); + if (ps->cur->b == NULL) { perror(NULL); - exit(EXIT_FAILURE); + exit(1); } memcpy(ps->cur->b + ps->cur->bsz, p, sz); - ps->cur->bsz += (size_t)sz; + ps->cur->bsz += sz; + ps->cur->b[ps->cur->bsz] = '\0'; + ps->cur->real = ps->cur->b; } +static void +pnode_trim(struct pnode *pn) +{ + assert(pn->node == NODE_TEXT); + for (; pn->bsz > 0; pn->b[--pn->bsz] = '\0') + if ( ! isspace((unsigned char)pn->b[pn->bsz - 1])) + break; +} + /* * Begin an element. * First, look for the element. * If we don't find it and we're not parsing, keep going. - * If we don't find it (and we're parsing), puke and exit. + * If we don't find it and we're parsing, puke and exit. * If we find it but we're not parsing yet (i.e., it's not a refentry * and thus out of context), keep going. - * If we're at the root and already have a tree, puke and exit. + * If we find it and we're at the root and already have a tree, puke and + * exit (FIXME: I don't think this is right?). + * If we find it but we're parsing a text node, close out the text node, + * return to its parent, and keep going. * Make sure that the element is in the right context. * Lastly, put the node onto our parse tree and continue. */ static void xml_elem_start(void *arg, const XML_Char *name, const XML_Char **atts) { - struct parse *ps = arg; - enum nodeid node; - struct pnode *dat; + struct parse *ps = arg; + const struct node *node; + enum attrkey key; + enum attrval val; + struct pnode *dat; + struct pattr *pattr; + const XML_Char **att; - if (ps->stop) + /* FIXME: find a better way to ditch other namespaces. */ + if (ps->stop || strcmp(name, "xi:include") == 0) return; /* Close out text node, if applicable... */ - if (NODE_TEXT == ps->node) { - assert(NULL != ps->cur); + if (ps->node == NODE_TEXT) { + pnode_trim(ps->cur); ps->cur = ps->cur->parent; - assert(NULL != ps->cur); ps->node = ps->cur->node; } - for (node = 0; node < NODE__MAX; node++) - if (NULL == nodes[node].name) - continue; - else if (0 == strcmp(nodes[node].name, name)) + for (node = nodes; node->name != NULL; node++) + if (strcmp(node->name, name) == 0) break; - if (NODE__MAX == node && NODE_ROOT == ps->node) { - fprintf(stderr, "%s: ignoring node\n", name); - return; - } else if (NODE__MAX == node) { - fprintf(stderr, "%s: unknown node\n", name); + if (node->name == NULL) { + if (ps->node == NODE_ROOT) + return; + fprintf(stderr, "%s:%zu:%zu: unknown node \"%s\"\n", + ps->fname, XML_GetCurrentLineNumber(ps->xml), + XML_GetCurrentColumnNumber(ps->xml), name); ps->stop = 1; return; - } else if (NODE_ROOT == ps->node && NULL != ps->root) { - fprintf(stderr, "%s: reentering?\n", name); + } else if (ps->node == NODE_ROOT && ps->root != NULL) { + fprintf(stderr, "%s:%zu:%zu: multiple refentries\n", + ps->fname, XML_GetCurrentLineNumber(ps->xml), + XML_GetCurrentColumnNumber(ps->xml)); ps->stop = 1; return; - } else if (NODE_ROOT == ps->node && NODE_REFENTRY != node) { - fprintf(stderr, "%s: known node w/o context\n", name); - return; - } else if ( ! isparent(node, ps->node)) { - fprintf(stderr, "%s: bad parent\n", name); - ps->stop = 1; - return; } - if (NULL == (dat = calloc(1, sizeof(struct pnode)))) { + if (node->node == NODE_INLINEEQUATION) + ps->flags |= PARSE_EQN; + + if ((dat = calloc(1, sizeof(*dat))) == NULL) { perror(NULL); - exit(EXIT_FAILURE); + exit(1); } - dat->node = ps->node = node; + dat->node = ps->node = node->node; dat->parent = ps->cur; TAILQ_INIT(&dat->childq); + TAILQ_INIT(&dat->attrq); - if (NULL != ps->cur) + if (ps->cur != NULL) TAILQ_INSERT_TAIL(&ps->cur->childq, dat, child); ps->cur = dat; - if (NULL == ps->root) + if (ps->root == NULL) ps->root = dat; + + /* + * Process attributes. + */ + for (att = atts; *att != NULL; att += 2) { + for (key = 0; key < ATTRKEY__MAX; key++) + if (strcmp(*att, attrkeys[key]) == 0) + break; + if (key == ATTRKEY__MAX) { + if (warn) + fprintf(stderr, "%s:%zu:%zu: warning: " + "unknown attribute \"%s\"\n", + ps->fname, + XML_GetCurrentLineNumber(ps->xml), + XML_GetCurrentColumnNumber(ps->xml), + *att); + continue; + } + for (val = 0; val < ATTRVAL__MAX; val++) + if (strcmp(att[1], attrvals[val]) == 0) + break; + pattr = calloc(1, sizeof(*pattr)); + pattr->key = key; + pattr->val = val; + if (val == ATTRVAL__MAX) + pattr->rawval = strdup(att[1]); + TAILQ_INSERT_TAIL(&dat->attrq, pattr, child); + } + } /* * Roll up the parse tree. - * Does nothing else special. + * If we're at a text node, roll that one up first. * If we hit the root, then assign ourselves as the NODE_ROOT. */ static void @@ -502,49 +423,66 @@ xml_elem_end(void *arg, const XML_Char *name) { struct parse *ps = arg; - if (ps->stop || NODE_ROOT == ps->node) + /* FIXME: find a better way to ditch other namespaces. */ + if (ps->stop || ps->node == NODE_ROOT) return; + else if (strcmp(name, "xi:include") == 0) + return; /* Close out text node, if applicable... */ - if (NODE_TEXT == ps->node) { - assert(NULL != ps->cur); + if (ps->node == NODE_TEXT) { + pnode_trim(ps->cur); ps->cur = ps->cur->parent; - assert(NULL != ps->cur); ps->node = ps->cur->node; } - if (NULL == (ps->cur = ps->cur->parent)) + if ((ps->cur = ps->cur->parent) == NULL) ps->node = NODE_ROOT; else ps->node = ps->cur->node; } +/* + * Recursively free a node (NULL is ok). + */ static void pnode_free(struct pnode *pn) { struct pnode *pp; + struct pattr *ap; - if (NULL == pn) + if (pn == NULL) return; - while (NULL != (pp = TAILQ_FIRST(&pn->childq))) { + while ((pp = TAILQ_FIRST(&pn->childq)) != NULL) { TAILQ_REMOVE(&pn->childq, pp, child); pnode_free(pp); } - free(pn->b); + while ((ap = TAILQ_FIRST(&pn->attrq)) != NULL) { + TAILQ_REMOVE(&pn->attrq, ap, child); + free(ap->rawval); + free(ap); + } + + free(pn->real); free(pn); } +/* + * Unlink a node from its parent and pnode_free() it. + */ static void pnode_unlink(struct pnode *pn) { - - if (NULL != pn->parent) + if (pn->parent != NULL) TAILQ_REMOVE(&pn->parent->childq, pn, child); pnode_free(pn); } +/* + * Unlink all children of a node and pnode_free() them. + */ static void pnode_unlinksub(struct pnode *pn) { @@ -553,148 +491,406 @@ pnode_unlinksub(struct pnode *pn) pnode_unlink(TAILQ_FIRST(&pn->childq)); } -static void -bufclear(struct parse *p) +/* + * Retrieve an enumeration attribute from a node. + * Return ATTRVAL__MAX if the node has no such attribute. + */ +enum attrval +pnode_getattr(struct pnode *pn, enum attrkey key) { + struct pattr *ap; - p->b[p->bsz = 0] = '\0'; + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ap->key == key) + return ap->val; + return ATTRVAL__MAX; } -static void -bufappend(struct parse *p, struct pnode *pn) +/* + * Retrieve an attribute string from a node. + * Return defval if the node has no such attribute. + */ +const char * +pnode_getattr_raw(struct pnode *pn, enum attrkey key, const char *defval) { + struct pattr *ap; - assert(NODE_TEXT == pn->node); - if (p->bsz + pn->bsz + 1 > p->mbsz) { - p->mbsz = p->bsz + pn->bsz + 1; - if (NULL == (p->b = realloc(p->b, p->mbsz))) { - perror(NULL); - exit(EXIT_FAILURE); - } + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ap->key == key) + return ap->val == ATTRVAL__MAX ? ap->rawval : + attrvals[ap->val]; + return defval; +} + +/* + * 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 (res != NULL) + break; } - memcpy(p->b + p->bsz, pn->b, pn->bsz); - p->bsz += pn->bsz; - p->b[p->bsz] = '\0'; + + return res; } static void -bufappend_r(struct parse *p, struct pnode *pn) +macro_open(struct parse *p, const char *name) { - struct pnode *pp; + switch (p->linestate) { + case LINE_TEXT: + putchar('\n'); + /* FALLTHROUGH */ + case LINE_NEW: + putchar('.'); + p->linestate = LINE_MACRO; + break; + case LINE_MACRO: + putchar(' '); + break; + } + fputs(name, stdout); +} - if (NODE_TEXT == pn->node) - bufappend(p, pn); - TAILQ_FOREACH(pp, &pn->childq, child) - bufappend_r(p, pp); +static void +macro_close(struct parse *p) +{ + assert(p->linestate == LINE_MACRO); + putchar('\n'); + p->linestate = LINE_NEW; } +static void +macro_line(struct parse *p, const char *name) +{ + macro_open(p, name); + macro_close(p); +} + +#define ARG_SPACE 1 /* Insert whitespace before this argument. */ +#define ARG_SINGLE 2 /* Quote argument if it contains whitespace. */ +#define ARG_QUOTED 4 /* We are already in a quoted argument. */ +#define ARG_UPPER 8 /* Covert argument to upper case. */ /* - * Print text presumably on a macro line. - * Ignore any child macros. - * Convert all whitespace to regular spaces. + * Print an argument string on a macro line, collapsing whitespace. */ static void -pnode_printmacrolinepart(struct parse *p, struct pnode *pn) +macro_addarg(struct parse *p, const char *arg, int flags) { - char *cp; + const char *cp; - bufclear(p); - bufappend_r(p, pn); + assert(p->linestate == LINE_MACRO); - /* Convert all space to spaces. */ - for (cp = p->b; '\0' != *cp; cp++) - if (isspace((int)*cp)) - *cp = ' '; + /* Quote if requested and necessary. */ - for (cp = p->b; isspace((int)*cp); cp++) - /* Spin past whitespace (XXX: necessary?) */ ; - for ( ; '\0' != *cp; cp++) { + if ((flags & (ARG_SINGLE | ARG_QUOTED)) == ARG_SINGLE) { + for (cp = arg; *cp != '\0'; cp++) + if (isspace((unsigned char)*cp)) + break; + if (*cp != '\0') { + if (flags & ARG_SPACE) { + putchar(' '); + flags &= ~ ARG_SPACE; + } + putchar('"'); + flags = ARG_QUOTED; + } + } + + for (cp = arg; *cp != '\0'; cp++) { + + /* Collapse whitespace. */ + + if (isspace((unsigned char)*cp)) { + flags |= ARG_SPACE; + continue; + } else if (flags & ARG_SPACE) { + putchar(' '); + flags &= ~ ARG_SPACE; + } + /* 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) || - ' ' == *(cp + 2) || - (islower((int)*(cp + 2)) && - ('\0' == *(cp + 3) || - ' ' == *(cp + 3))))) + + if ((flags & ARG_QUOTED) == 0 && + (cp == arg || isspace((unsigned char)cp[-1])) && + isupper((unsigned char)cp[0]) && + islower((unsigned char)cp[1]) && + (cp[2] == '\0' || cp[2] == ' ' || + (islower((unsigned char)cp[2]) && + (cp[3] == '\0' || cp[3] == ' ')))) fputs("\\&", stdout); - putchar(*cp); - /* If we're a character escape, escape us. */ - if ('\\' == *cp) + + if (*cp == '"') + fputs("\\(dq", stdout); + else if (flags & ARG_UPPER) + putchar(toupper((unsigned char)*cp)); + else + putchar(*cp); + if (*cp == '\\') putchar('e'); } } +static void +macro_argline(struct parse *p, const char *name, const char *arg) +{ + macro_open(p, name); + macro_addarg(p, arg, ARG_SPACE); + macro_close(p); +} + /* - * Just pnode_printmacrolinepart() but with a newline. - * If no text, just the newline. + * Recursively append text from the children of a node to a macro line. */ static void -pnode_printmacroline(struct parse *p, struct pnode *pn) +macro_addnode(struct parse *p, struct pnode *pn, int flags) { + int quote_now; - pnode_printmacrolinepart(p, pn); - putchar('\n'); + assert(p->linestate == LINE_MACRO); + + /* + * If the only child is a text node, just add that text, + * letting macro_addarg() decide about quoting. + */ + + pn = TAILQ_FIRST(&pn->childq); + if (pn != NULL && pn->node == NODE_TEXT && + TAILQ_NEXT(pn, child) == NULL) { + macro_addarg(p, pn->b, flags); + return; + } + + /* + * If we want the argument quoted and are not already + * in a quoted context, quote now. + */ + + quote_now = 0; + if (flags & ARG_SINGLE) { + if ((flags & ARG_QUOTED) == 0) { + if (flags & ARG_SPACE) { + putchar(' '); + flags &= ~ARG_SPACE; + } + putchar('"'); + flags |= ARG_QUOTED; + quote_now = 1; + } + flags &= ~ARG_SINGLE; + } + + /* + * Iterate to child and sibling nodes, + * inserting whitespace between nodes. + */ + + while (pn != NULL) { + if (pn->node == NODE_TEXT) + macro_addarg(p, pn->b, flags); + else + macro_addnode(p, pn, flags); + pn = TAILQ_NEXT(pn, child); + flags |= ARG_SPACE; + } + if (quote_now) + putchar('"'); } static void -pnode_printrefsynopsisdiv(struct parse *p, struct pnode *pn) +macro_nodeline(struct parse *p, const char *name, struct pnode *pn, int flags) { + macro_open(p, name); + macro_addnode(p, pn, ARG_SPACE | flags); + macro_close(p); +} + +/* + * If the next node is a text node starting with closing punctuation, + * emit the closing punctuation as a trailing macro argument. + */ +static void +macro_closepunct(struct parse *p, struct pnode *pn) +{ + if ((pn = TAILQ_NEXT(pn, child)) != NULL && + pn->node == NODE_TEXT && pn->bsz > 0 && + (pn->b[0] == ',' || pn->b[0] == '.') && + (pn->bsz == 1 || isspace((unsigned char)pn->b[1]))) { + putchar(' '); + putchar(pn->b[0]); + pn->b++; + pn->bsz--; + } + macro_close(p); +} + +static void +print_text(struct parse *p, const char *word) +{ + switch (p->linestate) { + case LINE_NEW: + break; + case LINE_TEXT: + putchar(' '); + break; + case LINE_MACRO: + macro_close(p); + break; + } + fputs(word, stdout); + p->linestate = LINE_TEXT; +} + +static void +pnode_printpara(struct parse *p, struct pnode *pn) +{ struct pnode *pp; - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_TITLE == pp->node) { + if ((pp = TAILQ_PREV(pn, pnodeq, child)) == NULL && + (pp = pn->parent) == NULL) + return; + + switch (pp->node) { + case NODE_ENTRY: + case NODE_LISTITEM: + return; + case NODE_PREFACE: + case NODE_SECTION: + if (p->level < 3) + return; + break; + default: + break; + } + macro_line(p, "Pp"); +} + +/* + * If the SYNOPSIS macro has a superfluous title, kill it. + */ +static void +pnode_printrefsynopsisdiv(struct parse *p, struct pnode *pn) +{ + struct pnode *pp, *pq; + + TAILQ_FOREACH_SAFE(pp, &pn->childq, child, pq) + if (pp->node == NODE_TITLE) pnode_unlink(pp); - break; - } - puts(".Sh SYNOPSIS"); + macro_line(p, "Sh SYNOPSIS"); } +/* + * Start a hopefully-named `Sh' section. + */ static void pnode_printrefsect(struct parse *p, struct pnode *pn) { struct pnode *pp; + const char *title; + int flags, level; + if (pn->parent == NULL) + return; + + level = ++p->level; + flags = ARG_SPACE; + if (level == 1) + flags |= ARG_UPPER; + if (level < 3) { + 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) + if (pp->node == NODE_TITLE) break; - fputs(".Sh ", stdout); + if (pp == NULL) { + 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; + } + } - if (NULL != pp) { - pnode_printmacroline(p, pp); + switch (level) { + case 1: + macro_open(p, "Sh"); + break; + case 2: + macro_open(p, "Ss"); + break; + default: + pnode_printpara(p, pn); + macro_open(p, "Sy"); + break; + } + + if (pp != NULL) { + macro_addnode(p, pp, flags); pnode_unlink(pp); } else - puts("UNKNOWN"); + macro_addarg(p, title, ARG_SPACE | ARG_QUOTED); + macro_close(p); } +/* + * Start a reference, extracting the title and volume. + */ static void pnode_printciterefentry(struct parse *p, struct pnode *pn) { struct pnode *pp, *title, *manvol; title = manvol = NULL; - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_MANVOLNUM == pp->node) + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node == NODE_MANVOLNUM) manvol = pp; - else if (NODE_REFENTRYTITLE == pp->node) + else if (pp->node == NODE_REFENTRYTITLE) title = pp; - - fputs(".Xr ", stdout); - - if (NULL != title) { - pnode_printmacrolinepart(p, title); - putchar(' '); - } else - fputs("unknown ", stdout); - - if (NULL != manvol) - pnode_printmacroline(p, manvol); + } + macro_open(p, "Xr"); + if (title == NULL) + macro_addarg(p, "unknown", ARG_SPACE); else - puts("1"); + macro_addnode(p, title, ARG_SPACE | ARG_SINGLE); + if (manvol == NULL) + macro_addarg(p, "1", ARG_SPACE); + else + macro_addnode(p, manvol, ARG_SPACE | ARG_SINGLE); + macro_close(p); + pnode_unlinksub(pn); } static void @@ -703,27 +899,23 @@ pnode_printrefmeta(struct parse *p, struct pnode *pn) struct pnode *pp, *title, *manvol; title = manvol = NULL; - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_MANVOLNUM == pp->node) + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node == NODE_MANVOLNUM) manvol = pp; - else if (NODE_REFENTRYTITLE == pp->node) + else if (pp->node == NODE_REFENTRYTITLE) title = pp; - - puts(".Dd $Mdocdate" "$"); - fputs(".Dt ", stdout); - - if (NULL != title) { - pnode_printmacrolinepart(p, title); - putchar(' '); - } else - fputs("UNKNOWN ", stdout); - - if (NULL != manvol) - pnode_printmacroline(p, manvol); + } + macro_open(p, "Dt"); + if (title == NULL) + macro_addarg(p, "UNKNOWN", ARG_SPACE); else - puts("1"); - - puts(".Os"); + macro_addnode(p, title, ARG_SPACE | ARG_SINGLE | ARG_UPPER); + if (manvol == NULL) + macro_addarg(p, "1", ARG_SPACE); + else + macro_addnode(p, manvol, ARG_SPACE | ARG_SINGLE); + macro_close(p); + pnode_unlink(pn); } static void @@ -732,48 +924,77 @@ pnode_printfuncdef(struct parse *p, struct pnode *pn) struct pnode *pp, *ftype, *func; ftype = func = NULL; - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_TEXT == pp->node) + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node == NODE_TEXT) ftype = pp; - else if (NODE_FUNCTION == pp->node) + else if (pp->node == NODE_FUNCTION) func = pp; - - if (NULL != ftype) { - fputs(".Ft ", stdout); - pnode_printmacroline(p, ftype); } - - if (NULL != func) { - fputs(".Fo ", stdout); - pnode_printmacroline(p, func); - } else - puts(".Fo UNKNOWN"); + if (ftype != NULL) + macro_argline(p, "Ft", ftype->b); + macro_open(p, "Fo"); + if (func == NULL) + macro_addarg(p, "UNKNOWN", ARG_SPACE); + else + macro_addnode(p, func, ARG_SPACE | ARG_SINGLE); + macro_close(p); } +/* + * 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_printparamdef(struct parse *p, struct pnode *pn) +pnode_printmathfenced(struct parse *p, struct pnode *pn) { - struct pnode *pp, *ptype, *param; + struct pnode *pp; - ptype = param = NULL; - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_TEXT == pp->node) - ptype = pp; - else if (NODE_PARAMETER == pp->node) - param = pp; + printf("left %s ", pnode_getattr_raw(pn, ATTRKEY_OPEN, "(")); - fputs(".Fa \"", stdout); - if (NULL != ptype) { - pnode_printmacrolinepart(p, ptype); - putchar(' '); + pp = TAILQ_FIRST(&pn->childq); + pnode_print(p, pp); + + while ((pp = TAILQ_NEXT(pp, child)) != NULL) { + putchar(','); + pnode_print(p, pp); } + printf("right %s ", pnode_getattr_raw(pn, ATTRKEY_CLOSE, ")")); + pnode_unlinksub(pn); +} - if (NULL != param) - pnode_printmacrolinepart(p, param); - else - fputs("UNKNOWN", stdout); +/* + * 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; - puts("\""); + 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); + pnode_unlinksub(pn); } static void @@ -782,165 +1003,498 @@ pnode_printfuncprototype(struct parse *p, struct pnode struct pnode *pp, *fdef; TAILQ_FOREACH(fdef, &pn->childq, child) - if (NODE_FUNCDEF == fdef->node) + if (fdef->node == NODE_FUNCDEF) break; - if (NULL != fdef) + if (fdef != NULL) pnode_printfuncdef(p, fdef); else - puts(".Fo UNKNOWN"); + macro_line(p, "Fo UNKNOWN"); - TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_PARAMDEF == pp->node) - pnode_printparamdef(p, pp); + TAILQ_FOREACH(pp, &pn->childq, child) + if (pp->node == NODE_PARAMDEF) + macro_nodeline(p, "Fa", pp, ARG_SINGLE); - puts(".Fc"); + macro_line(p, "Fc"); + pnode_unlinksub(pn); } -/* TODO: handle "optional" values. */ +/* + * 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"). + * This also handles the case of "repetition" (or in other words, the + * ellipsis following an argument) and optionality. + */ static void -pnode_printarg(struct parse *p, struct pnode *pn, int nested) +pnode_printarg(struct parse *p, struct pnode *pn) { struct pnode *pp; - int sv = nested; + struct pattr *ap; + int isop, isrep; - if ( ! nested) - fputs(".", stdout); - nested = 1; + isop = 1; + isrep = 0; + TAILQ_FOREACH(ap, &pn->attrq, child) { + if (ap->key == ATTRKEY_CHOICE && + (ap->val == ATTRVAL_PLAIN || ap->val == ATTRVAL_REQ)) + isop = 0; + else if (ap->key == ATTRKEY_REP && ap->val == ATTRVAL_REPEAT) + isrep = 1; + } + if (isop) + macro_open(p, "Op"); + + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node == NODE_TEXT) + macro_open(p, "Ar"); + pnode_print(p, pp); + if (isrep && pp->node == NODE_TEXT) + macro_addarg(p, "...", ARG_SPACE); + } + pnode_unlinksub(pn); +} + +static void +pnode_printgroup(struct parse *p, struct pnode *pn) +{ + struct pnode *pp, *np; + struct pattr *ap; + int isop, sv; + + isop = 1; + TAILQ_FOREACH(ap, &pn->attrq, child) + if (ap->key == ATTRKEY_CHOICE && + (ap->val == ATTRVAL_PLAIN || ap->val == ATTRVAL_REQ)) { + 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->linestate == LINE_NEW; + if (isop) + macro_open(p, "Op"); + else if (sv) + macro_open(p, "No"); + + /* + * 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 (np != NULL) { + if (pp->node != np->node) + break; + macro_addarg(p, "|", ARG_SPACE); + macro_addnode(p, np, ARG_SPACE); + pp = np; + np = TAILQ_NEXT(np, child); + } + } + if (sv) + macro_close(p); + pnode_unlinksub(pn); +} + +static void +pnode_printprologue(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + pp = p->root == NULL ? NULL : + pnode_findfirst(p->root, NODE_REFMETA); + + macro_line(p, "Dd $Mdocdate" "$"); + if (pp != NULL) + pnode_printrefmeta(p, pp); + else { + macro_open(p, "Dt"); + macro_addarg(p, + pnode_getattr_raw(p->root, ATTRKEY_ID, "UNKNOWN"), + ARG_SPACE | ARG_SINGLE | ARG_UPPER); + macro_addarg(p, "1", ARG_SPACE); + macro_close(p); + } + macro_line(p, "Os"); + + if (p->flags & PARSE_EQN) { + macro_line(p, "EQ"); + print_text(p, "delim $$"); + macro_line(p, "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; + + macro_open(p, "It"); + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node != NODE_TERM) + continue; + if ( ! first) + macro_addarg(p, ",", 0); + pnode_print(p, pp); + first = 0; + } + macro_close(p); TAILQ_FOREACH(pp, &pn->childq, child) - if (NODE_OPTION == pp->node) { - fputs("Fl ", stdout); - pnode_printmacrolinepart(p, pp); - } else if (NODE_TEXT == pp->node) { - fputs("Ar ", stdout); - pnode_printmacrolinepart(p, pp); - } else if (NODE_ARG == pp->node) - pnode_printarg(p, pp, nested); + if (pp->node != NODE_TERM) + pnode_print(p, pp); + pnode_unlinksub(pn); +} - if ( ! sv) - puts(""); +static void +pnode_printtitle(struct parse *p, struct pnode *pn) +{ + struct pnode *pp, *pq; + + TAILQ_FOREACH_SAFE(pp, &pn->childq, child, pq) { + if (pp->node == NODE_TITLE) { + pnode_printpara(p, pp); + pnode_print(p, pp); + pnode_unlink(pp); + } + } } +static void +pnode_printrow(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + macro_line(p, "Bl -dash -compact"); + TAILQ_FOREACH(pp, &pn->childq, child) { + macro_line(p, "It"); + pnode_print(p, pp); + } + macro_line(p, "El"); + pnode_unlink(pn); +} + +static void +pnode_printtable(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + pnode_printtitle(p, pn); + macro_line(p, "Bl -ohang"); + while ((pp = pnode_findfirst(pn, NODE_ROW)) != NULL) { + macro_line(p, "It Table Row"); + pnode_printrow(p, pp); + } + macro_line(p, "El"); + pnode_unlinksub(pn); +} + +static void +pnode_printlist(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + pnode_printtitle(p, pn); + macro_argline(p, "Bl", + pn->node == NODE_ORDEREDLIST ? "-enum" : "-bullet"); + TAILQ_FOREACH(pp, &pn->childq, child) { + macro_line(p, "It"); + pnode_print(p, pp); + } + macro_line(p, "El"); + pnode_unlinksub(pn); +} + +static void +pnode_printvariablelist(struct parse *p, struct pnode *pn) +{ + struct pnode *pp; + + pnode_printtitle(p, pn); + macro_line(p, "Bl -tag -width Ds"); + TAILQ_FOREACH(pp, &pn->childq, child) { + if (pp->node == NODE_VARLISTENTRY) + pnode_print(p, pp); + else + macro_nodeline(p, "It", pp, 0); + } + macro_line(p, "El"); + pnode_unlinksub(pn); +} + /* * Print a parsed node (or ignore it--whatever). * This is a recursive function. - * FIXME: macro line continuation? + * FIXME: if we're in a literal context ( or or + * whatever), don't print inline macros. */ static void pnode_print(struct parse *p, struct pnode *pn) { struct pnode *pp; + const char *ccp; char *cp; int last; + enum linestate sv; - if (NULL == pn) + if (pn == NULL) return; - if (NODE_TEXT != pn->node && NODE_ROOT != pn->node) - printf(".\\\" %s\n", nodes[pn->node].name); + sv = p->linestate; switch (pn->node) { - case (NODE_ARG): - pnode_printarg(p, pn, 0); - pnode_unlinksub(pn); + case NODE_APPLICATION: + macro_open(p, "Nm"); break; - case (NODE_CITEREFENTRY): + case NODE_ANCHOR: + /* Don't print anything! */ + return; + case NODE_ARG: + pnode_printarg(p, pn); + break; + case NODE_AUTHOR: + macro_open(p, "An"); + break; + case NODE_AUTHORGROUP: + macro_line(p, "An -split"); + break; + case NODE_BOOKINFO: + macro_line(p, "Sh NAME"); + break; + case NODE_CITEREFENTRY: pnode_printciterefentry(p, pn); - pnode_unlinksub(pn); break; - case (NODE_CODE): - fputs(".Li ", stdout); - pnode_printmacroline(p, pn); - pnode_unlinksub(pn); + case NODE_CITETITLE: + macro_open(p, "%T"); break; - case (NODE_COMMAND): - fputs(".Nm ", stdout); - pnode_printmacroline(p, pn); - pnode_unlinksub(pn); + case NODE_CODE: + macro_open(p, "Li"); break; - case (NODE_FUNCTION): - fputs(".Fn ", stdout); - pnode_printmacroline(p, pn); - pnode_unlinksub(pn); + case NODE_COMMAND: + macro_open(p, "Nm"); break; - case (NODE_FUNCPROTOTYPE): + case NODE_CONSTANT: + macro_open(p, "Dv"); + break; + case NODE_EDITOR: + print_text(p, "editor:"); + macro_open(p, "An"); + break; + case NODE_EMAIL: + macro_open(p, "Aq Mt"); + break; + case NODE_EMPHASIS: + case NODE_FIRSTTERM: + macro_open(p, "Em"); + break; + case NODE_ENVAR: + macro_open(p, "Ev"); + break; + case NODE_FILENAME: + macro_open(p, "Pa"); + break; + case NODE_FUNCTION: + macro_open(p, "Fn"); + break; + case NODE_FUNCPROTOTYPE: pnode_printfuncprototype(p, pn); - pnode_unlinksub(pn); break; - case (NODE_FUNCSYNOPSISINFO): - fputs(".Fd ", stdout); - pnode_printmacroline(p, pn); - pnode_unlinksub(pn); + case NODE_FUNCSYNOPSISINFO: + macro_open(p, "Fd"); break; - case (NODE_PARA): - /* FIXME: not always. */ - puts(".Pp"); + case NODE_INDEXTERM: + return; + case NODE_INFORMALEQUATION: + macro_line(p, "EQ"); break; - case (NODE_PARAMETER): - fputs(".Fa \"", stdout); - pnode_printmacrolinepart(p, pn); - puts("\""); - pnode_unlinksub(pn); + case NODE_INLINEEQUATION: + if (p->linestate == LINE_NEW) + p->linestate = LINE_TEXT; + putchar('$'); break; - case (NODE_PROGRAMLISTING): - puts(".Bd -literal"); + case NODE_ITEMIZEDLIST: + pnode_printlist(p, pn); break; - case (NODE_REFMETA): - pnode_printrefmeta(p, pn); + case NODE_GROUP: + pnode_printgroup(p, pn); + break; + case NODE_KEYSYM: + macro_open(p, "Sy"); + break; + case NODE_LEGALNOTICE: + macro_line(p, "Sh LEGAL NOTICE"); + break; + case NODE_LINK: + ccp = pnode_getattr_raw(pn, ATTRKEY_LINKEND, NULL); + if (ccp == NULL) + break; + macro_argline(p, "Sx", ccp); + return; + case NODE_LITERAL: + macro_open(p, "Li"); + break; + case NODE_LITERALLAYOUT: + macro_argline(p, "Bd", pnode_getattr(pn, ATTRKEY_CLASS) == + ATTRVAL_MONOSPACED ? "-literal" : "-unfilled"); + break; + case NODE_MML_MFENCED: + pnode_printmathfenced(p, 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); + break; + case NODE_OPTION: + macro_open(p, "Fl"); + break; + case NODE_ORDEREDLIST: + pnode_printlist(p, pn); + break; + case NODE_PARA: + pnode_printpara(p, pn); + break; + case NODE_PARAMETER: + macro_nodeline(p, "Fa", pn, ARG_SINGLE); pnode_unlinksub(pn); - p->flags |= PARSE_HAS_META; break; - case (NODE_REFNAME): - fputs(".Nm ", stdout); - pnode_printmacroline(p, pn); + case NODE_QUOTE: + macro_open(p, "Qo"); + break; + case NODE_PROGRAMLISTING: + case NODE_SCREEN: + macro_line(p, "Bd -literal"); + break; + case NODE_REFENTRYINFO: + /* Suppress. */ pnode_unlinksub(pn); - return; - case (NODE_REFNAMEDIV): - if ( ! (PARSE_HAS_META & p->flags)) { - puts(".Dd $Mdocdate" "$"); - puts(".Dt UNKNOWN 1"); - puts(".Os"); - } - puts(".Sh NAME"); break; - case (NODE_REFPURPOSE): - fputs(".Nd ", stdout); - pnode_printmacroline(p, pn); + case NODE_REFMETA: + abort(); + break; + case NODE_REFNAME: + /* Suppress non-text children... */ + macro_open(p, "Nm"); + macro_addnode(p, pn, ARG_SPACE | ARG_SINGLE); pnode_unlinksub(pn); - return; - case (NODE_REFSYNOPSISDIV): + break; + case NODE_REFNAMEDIV: + macro_line(p, "Sh NAME"); + break; + case NODE_REFPURPOSE: + macro_open(p, "Nd"); + break; + case NODE_REFSYNOPSISDIV: pnode_printrefsynopsisdiv(p, pn); break; - case (NODE_REFSECT1): + case NODE_PREFACE: + case NODE_SECTION: + case NODE_NOTE: + case NODE_TIP: + case NODE_CAUTION: + case NODE_WARNING: pnode_printrefsect(p, pn); break; - case (NODE_TEXT): - bufclear(p); - bufappend(p, pn); + case NODE_REPLACEABLE: + macro_open(p, "Ar"); + break; + case NODE_SBR: + macro_line(p, "br"); + break; + case NODE_SGMLTAG: + macro_open(p, "Li"); + break; + case NODE_STRUCTNAME: + macro_open(p, "Vt"); + break; + case NODE_TABLE: + case NODE_INFORMALTABLE: + pnode_printtable(p, pn); + break; + case NODE_TEXT: + if (pn->bsz == 0) { + assert(pn->real != pn->b); + break; + } + if (p->linestate == LINE_NEW) + p->linestate = LINE_TEXT; + else + putchar(' '); + /* * Output all characters, squeezing out whitespace - * between newlines. + * between newlines. * XXX: all whitespace, including tabs (?). * Remember to escape control characters and escapes. */ - for (last = '\n', cp = p->b; '\0' != *cp; ) { - if ('\n' == last) { + cp = pn->b; + + /* + * There's often a superfluous "-" in its