=================================================================== RCS file: /cvs/mandoc/mandoc.h,v retrieving revision 1.60 retrieving revision 1.87 diff -u -p -r1.60 -r1.87 --- mandoc/mandoc.h 2011/03/17 09:16:38 1.60 +++ mandoc/mandoc.h 2011/07/21 14:13:00 1.87 @@ -1,4 +1,4 @@ -/* $Id: mandoc.h,v 1.60 2011/03/17 09:16:38 kristaps Exp $ */ +/* $Id: mandoc.h,v 1.87 2011/07/21 14:13:00 kristaps Exp $ */ /* * Copyright (c) 2010, 2011 Kristaps Dzonsons * @@ -74,6 +74,7 @@ enum mandocerr { MANDOCERR_CHILD, /* child violates parent syntax */ MANDOCERR_NESTEDDISP, /* nested displays are not portable */ MANDOCERR_SCOPEREP, /* already in literal mode */ + MANDOCERR_LINESCOPE, /* line scope broken */ /* related to missing macro arguments */ MANDOCERR_MACROEMPTY, /* skipping empty macro */ @@ -103,8 +104,17 @@ enum mandocerr { MANDOCERR_BADESCAPE, /* unknown escape sequence */ MANDOCERR_BADQUOTE, /* unterminated quoted string */ + /* related to equations */ + MANDOCERR_EQNQUOTE, /* unexpected literal in equation */ + MANDOCERR_ERROR, /* ===== start of errors ===== */ + /* related to equations */ + MANDOCERR_EQNARGS, /* bad equation macro arguments */ + MANDOCERR_EQNNEST, /* too many nested equation defines */ + MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/ + MANDOCERR_EQNSCOPE, /* equation scope open on exit */ + /* related to tables */ MANDOCERR_TBL, /* bad table syntax */ MANDOCERR_TBLOPT, /* bad table option */ @@ -121,7 +131,6 @@ enum mandocerr { MANDOCERR_NOTEXT, /* skipping text before the first section header */ MANDOCERR_MACRO, /* skipping unknown macro */ MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */ - MANDOCERR_LINESCOPE, /* line scope broken */ MANDOCERR_ARGCOUNT, /* argument count wrong */ MANDOCERR_NOSCOPE, /* skipping end of block that is not open */ MANDOCERR_SCOPEBROKEN, /* missing end of block */ @@ -137,6 +146,7 @@ enum mandocerr { MANDOCERR_FATAL, /* ===== start of fatal errors ===== */ + MANDOCERR_NOTMANUAL, /* manual isn't really a manual */ MANDOCERR_COLUMNS, /* column syntax is inconsistent */ MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */ MANDOCERR_SYNTLINESCOPE, /* line scope broken, syntax violated */ @@ -269,75 +279,126 @@ struct tbl_span { struct tbl_span *next; }; -struct eqn { - size_t sz; - char *data; - int line; /* invocation line */ - int pos; /* invocation position */ +enum eqn_boxt { + EQN_ROOT, /* root of parse tree */ + EQN_TEXT, /* text (number, variable, whatever) */ + EQN_SUBEXPR /* nested subexpression */ }; -/* - * Available registers (set in libroff, accessed elsewhere). - */ -enum regs { - REG_nS = 0, - REG__MAX +enum eqn_markt { + EQNMARK_NONE = 0, + EQNMARK_DOT, + EQNMARK_DOTDOT, + EQNMARK_HAT, + EQNMARK_TILDE, + EQNMARK_VEC, + EQNMARK_DYAD, + EQNMARK_BAR, + EQNMARK_UNDER, + EQNMARK__MAX }; -/* - * A register (struct reg) can consist of many types: this consists of - * normalised types from the original string form. - */ -union regval { - unsigned u; /* unsigned integer */ +enum eqn_fontt { + EQNFONT_NONE = 0, + EQNFONT_ROMAN, + EQNFONT_BOLD, + EQNFONT_ITALIC, + EQNFONT__MAX }; -/* - * A single register entity. If "set" is zero, the value of the - * register should be the default one, which is per-register. It's - * assumed that callers know which type in "v" corresponds to which - * register value. +enum eqn_post { + EQNPOS_NONE = 0, + EQNPOS_OVER, + EQNPOS_SUP, + EQNPOS_SUB, + EQNPOS_TO, + EQNPOS_FROM, + EQNPOS_ABOVE, + EQNPOS__MAX +}; + + /* + * A "box" is a parsed mathematical expression as defined by the eqn.7 + * grammar. */ -struct reg { - int set; /* whether set or not */ - union regval v; /* parsed data */ +struct eqn_box { + enum eqn_boxt type; /* type of node */ + struct eqn_box *child; /* child node */ + struct eqn_box *next; /* next in tree */ + enum eqn_post pos; /* position of next box */ + char *text; /* text (or NULL) */ + enum eqn_markt mark; /* a mark about the box */ + enum eqn_fontt font; /* font of box */ }; /* - * The primary interface to setting register values is in libroff, - * although libmdoc and libman from time to time will manipulate - * registers (such as `.Sh SYNOPSIS' enabling REG_nS). + * An equation consists of a tree of expressions starting at a given + * line and position. */ -struct regset { - struct reg regs[REG__MAX]; +struct eqn { + struct eqn_box *root; /* root mathematical expression */ + int ln; /* invocation line */ + int pos; /* invocation position */ }; /* - * A punctuation delimiter, used only in mdoc(7) documents, is opening, - * closing, or "middle mark" punctuation. These govern spacing. - * Opening punctuation (e.g., the opening parenthesis) suppresses the - * following space; closing punctuation (e.g., the closing parenthesis) - * suppresses the leading space; middle punctuation (e.g., the vertical - * bar) can do either. The middle punctuation delimiter bends the rules - * depending on usage. + * The type of parse sequence. This value is usually passed via the + * mandoc(1) command line of -man and -mdoc. It's almost exclusively + * -mandoc but the others have been retained for compatibility. */ -enum mdelim { - DELIM_NONE = 0, - DELIM_OPEN, - DELIM_MIDDLE, - DELIM_CLOSE +enum mparset { + MPARSE_AUTO, /* magically determine the document type */ + MPARSE_MDOC, /* assume -mdoc */ + MPARSE_MAN /* assume -man */ }; -typedef int (*mandocmsg)(enum mandocerr, void *, - int, int, const char *); +enum mandoc_esc { + ESCAPE_ERROR = 0, /* bail! unparsable escape */ + ESCAPE_IGNORE, /* escape to be ignored */ + ESCAPE_SPECIAL, /* a regular special character */ + ESCAPE_FONT, /* a generic font mode */ + ESCAPE_FONTBOLD, /* bold font mode */ + ESCAPE_FONTITALIC, /* italic font mode */ + ESCAPE_FONTROMAN, /* roman font mode */ + ESCAPE_FONTPREV, /* previous font mode */ + ESCAPE_NUMBERED, /* a numbered glyph */ + ESCAPE_UNICODE, /* a unicode codepoint */ + ESCAPE_NOSPACE /* suppress space if the last on a line */ +}; +typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel, + const char *, int, int, const char *); + +struct mparse; +struct mchars; +struct mdoc; +struct man; + __BEGIN_DECLS +void mparse_free(struct mparse *); +void mparse_reset(struct mparse *); +struct mparse *mparse_alloc(enum mparset, + enum mandoclevel, mandocmsg, void *); +enum mandoclevel mparse_readfd(struct mparse *, int, const char *); +void mparse_result(struct mparse *, struct mdoc **, struct man **); +const char *mparse_strerror(enum mandocerr); +const char *mparse_strlevel(enum mandoclevel); + void *mandoc_calloc(size_t, size_t); void *mandoc_malloc(size_t); void *mandoc_realloc(void *, size_t); -#define DELIMSZ 6 /* hint: max possible size of a delimiter */ -enum mdelim mandoc_isdelim(const char *); +char *mandoc_strdup(const char *); + +enum mandoc_esc mandoc_escape(const char **, const char **, int *); + +struct mchars *mchars_alloc(void); +char mchars_num2char(const char *, size_t); +int mchars_num2uc(const char *, size_t); +const char *mchars_spec2str(struct mchars *, const char *, size_t, size_t *); +int mchars_spec2cp(struct mchars *, const char *, size_t); +void mchars_free(struct mchars *); + __END_DECLS