=================================================================== RCS file: /cvs/mandoc/read.c,v retrieving revision 1.44 retrieving revision 1.59 diff -u -p -r1.44 -r1.59 --- mandoc/read.c 2014/03/19 21:51:20 1.44 +++ mandoc/read.c 2014/07/03 23:24:56 1.59 @@ -1,4 +1,4 @@ -/* $Id: read.c,v 1.44 2014/03/19 21:51:20 schwarze Exp $ */ +/* $Id: read.c,v 1.59 2014/07/03 23:24:56 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2010-2014 Ingo Schwarze @@ -37,6 +37,7 @@ #include #include "mandoc.h" +#include "mandoc_aux.h" #include "libmandoc.h" #include "mdoc.h" #include "man.h" @@ -45,7 +46,7 @@ #define REPARSE_LIMIT 1000 struct buf { - char *buf; /* binary input buffer */ + char *buf; /* binary input buffer */ size_t sz; /* size of binary buffer */ }; @@ -59,9 +60,10 @@ struct mparse { struct man *man; /* man parser */ struct mdoc *mdoc; /* mdoc parser */ struct roff *roff; /* roff parser (!NULL) */ + char *sodest; /* filename pointed to by .so */ int reparse_count; /* finite interp. stack */ mandocmsg mmsg; /* warning/error message handler */ - const char *file; + const char *file; struct buf *secondary; char *defos; /* default operating system */ }; @@ -91,44 +93,50 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "generic warning", /* related to the prologue */ - "no title in document", - "document title should be all caps", + "missing .TH macro, using \"unknown 1\"", + "lower case character in document title", "unknown manual section", "unknown manual volume or arch", - "date missing, using today's date", + "missing date, using today's date", "cannot parse date, using it verbatim", "prologue macros out of order", "duplicate prologue macro", - "macro not allowed in prologue", - "macro not allowed in body", + "incomplete prologue, terminated by", + "skipping prologue macro in body", /* related to document structure */ ".so is fragile, better use ln(1)", - "NAME section must come first", + "no document body", + "content before first section header", + "first section is not \"NAME\"", "bad NAME section contents", "sections out of conventional order", - "duplicate section name", - "section header suited to sections 2, 3, and 9 only", + "duplicate section title", + "unexpected section", /* related to macros and nesting */ - "skipping obsolete macro", + "obsolete macro", "skipping paragraph macro", "moving paragraph macro out of list", "skipping no-space macro", "blocks badly nested", - "child violates parent syntax", "nested displays are not portable", - "already in literal mode", + "moving content out of list", + ".Vt block has child macro", + "fill mode already enabled, skipping .fi", + "fill mode already disabled, skipping .nf", "line scope broken", /* related to missing macro arguments */ + "skipping empty request", + "conditional request controls empty scope", "skipping empty macro", "argument count wrong", "missing display type", "list type must come first", "tag lists require a width argument", + "missing head in list item", "missing font type", - "skipping end of block that is not open", /* related to bad macro arguments */ "skipping argument", @@ -149,9 +157,6 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "bad escape sequence", "unterminated quoted string", - /* related to equations */ - "unexpected literal in equation", - "generic error", /* related to equations */ @@ -175,22 +180,21 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "skipping bad character", "escaped character not allowed in a name", "manual name not yet set", - "skipping text before the first section header", + "skipping text before first section header", "skipping unknown macro", "NOT IMPLEMENTED, please use groff: skipping request", "argument count wrong", + "skipping invalid content in .Rs block", "skipping column outside column list", "skipping end of block that is not open", "missing end of block", "scope open on exit", "uname(3) system call failed", - "macro requires line argument(s)", "macro requires body argument(s)", "macro requires argument(s)", "request requires a numeric argument", "missing list type", "line argument(s) will be lost", - "body argument(s) will be lost", "generic fatal error", @@ -202,12 +206,12 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "child violates parent syntax", "argument count wrong, violates syntax", "NOT IMPLEMENTED: .so with absolute path or \"..\"", - "no document body", + ".so request failed", "no document prologue", "static buffer exhausted", /* system errors */ - "cannot open file", + NULL, "cannot stat file", "cannot read file", }; @@ -222,6 +226,7 @@ static const char * const mandoclevels[MANDOCLEVEL_MAX "SYSERR" }; + static void resize_buf(struct buf *buf, size_t initial) { @@ -255,7 +260,7 @@ pset(const char *buf, int pos, struct mparse *curp) } if (MPARSE_MDOC & curp->options) { - if (NULL == curp->pmdoc) + if (NULL == curp->pmdoc) curp->pmdoc = mdoc_alloc( curp->roff, curp, curp->defos, MPARSE_QUICK & curp->options ? 1 : 0); @@ -263,7 +268,7 @@ pset(const char *buf, int pos, struct mparse *curp) curp->mdoc = curp->pmdoc; return; } else if (MPARSE_MAN & curp->options) { - if (NULL == curp->pman) + if (NULL == curp->pman) curp->pman = man_alloc(curp->roff, curp, MPARSE_QUICK & curp->options ? 1 : 0); assert(curp->pman); @@ -272,16 +277,16 @@ pset(const char *buf, int pos, struct mparse *curp) } if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3)) { - if (NULL == curp->pmdoc) + if (NULL == curp->pmdoc) curp->pmdoc = mdoc_alloc( curp->roff, curp, curp->defos, MPARSE_QUICK & curp->options ? 1 : 0); assert(curp->pmdoc); curp->mdoc = curp->pmdoc; return; - } + } - if (NULL == curp->pman) + if (NULL == curp->pman) curp->pman = man_alloc(curp->roff, curp, MPARSE_QUICK & curp->options ? 1 : 0); assert(curp->pman); @@ -306,8 +311,8 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int memset(&ln, 0, sizeof(struct buf)); - lnn = curp->line; - pos = 0; + lnn = curp->line; + pos = 0; for (i = 0; i < (int)blk.sz; ) { if (0 == pos && '\0' == blk.buf[i]) @@ -344,7 +349,7 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int if (pos + 2 >= (int)ln.sz) resize_buf(&ln, 256); - /* + /* * Warn about bogus characters. If you're using * non-ASCII encoding, you're screwing your * readers. Since I'd rather this not happen, @@ -355,10 +360,10 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int c = (unsigned char) blk.buf[i]; - if ( ! (isascii(c) && - (isgraph(c) || isblank(c)))) { + if ( ! (isascii(c) && + (isgraph(c) || isblank(c)))) { mandoc_msg(MANDOCERR_BADCHAR, curp, - curp->line, pos, NULL); + curp->line, pos, NULL); i++; ln.buf[pos++] = '?'; continue; @@ -412,10 +417,10 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int c = (unsigned char) blk.buf[i+1]; - if ( ! (isascii(c) && - (isgraph(c) || isblank(c)))) { + if ( ! (isascii(c) && + (isgraph(c) || isblank(c)))) { mandoc_msg(MANDOCERR_BADCHAR, curp, - curp->line, pos, NULL); + curp->line, pos, NULL); i += 2; ln.buf[pos++] = '?'; continue; @@ -427,7 +432,7 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int ln.buf[pos++] = blk.buf[i++]; } - if (pos >= (int)ln.sz) + if (pos >= (int)ln.sz) resize_buf(&ln, 256); ln.buf[pos] = '\0'; @@ -450,13 +455,12 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int */ if (curp->secondary) { - curp->secondary->buf = - mandoc_realloc - (curp->secondary->buf, - curp->secondary->sz + pos + 2); - memcpy(curp->secondary->buf + - curp->secondary->sz, - ln.buf, pos); + curp->secondary->buf = mandoc_realloc( + curp->secondary->buf, + curp->secondary->sz + pos + 2); + memcpy(curp->secondary->buf + + curp->secondary->sz, + ln.buf, pos); curp->secondary->sz += pos; curp->secondary->buf [curp->secondary->sz] = '\n'; @@ -465,41 +469,50 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int [curp->secondary->sz] = '\0'; } rerun: - rr = roff_parseln - (curp->roff, curp->line, - &ln.buf, &ln.sz, of, &of); + rr = roff_parseln(curp->roff, curp->line, + &ln.buf, &ln.sz, of, &of); switch (rr) { - case (ROFF_REPARSE): + case ROFF_REPARSE: if (REPARSE_LIMIT >= ++curp->reparse_count) mparse_buf_r(curp, ln, 0); else mandoc_msg(MANDOCERR_ROFFLOOP, curp, - curp->line, pos, NULL); + curp->line, pos, NULL); pos = 0; continue; - case (ROFF_APPEND): + case ROFF_APPEND: pos = (int)strlen(ln.buf); continue; - case (ROFF_RERUN): + case ROFF_RERUN: goto rerun; - case (ROFF_IGN): + case ROFF_IGN: pos = 0; continue; - case (ROFF_ERR): + case ROFF_ERR: assert(MANDOCLEVEL_FATAL <= curp->file_status); break; - case (ROFF_SO): + case ROFF_SO: + if (0 == (MPARSE_SO & curp->options) && + (i >= (int)blk.sz || '\0' == blk.buf[i])) { + curp->sodest = mandoc_strdup(ln.buf + of); + free(ln.buf); + return; + } /* * We remove `so' clauses from our lookaside * buffer because we're going to descend into * the file recursively. */ - if (curp->secondary) + if (curp->secondary) curp->secondary->sz -= pos + 1; mparse_readfd(curp, -1, ln.buf + of); - if (MANDOCLEVEL_FATAL <= curp->file_status) + if (MANDOCLEVEL_FATAL <= curp->file_status) { + mandoc_vmsg(MANDOCERR_SO_FAIL, + curp, curp->line, pos, + ".so %s", ln.buf + of); break; + } pos = 0; continue; default: @@ -524,7 +537,7 @@ rerun: if ( ! (curp->man || curp->mdoc)) pset(ln.buf + of, pos - of, curp); - /* + /* * Lastly, push down into the parsers themselves. One * of these will have already been set in the pset() * routine. @@ -540,23 +553,23 @@ rerun: if (ROFF_TBL == rr) while (NULL != (span = roff_span(curp->roff))) { rc = curp->man ? - man_addspan(curp->man, span) : - mdoc_addspan(curp->mdoc, span); + man_addspan(curp->man, span) : + mdoc_addspan(curp->mdoc, span); if (0 == rc) break; } else if (ROFF_EQN == rr) - rc = curp->mdoc ? - mdoc_addeqn(curp->mdoc, - roff_eqn(curp->roff)) : - man_addeqn(curp->man, - roff_eqn(curp->roff)); + rc = curp->mdoc ? + mdoc_addeqn(curp->mdoc, + roff_eqn(curp->roff)) : + man_addeqn(curp->man, + roff_eqn(curp->roff)); else if (curp->man || curp->mdoc) rc = curp->man ? - man_parseln(curp->man, - curp->line, ln.buf, of) : - mdoc_parseln(curp->mdoc, - curp->line, ln.buf, of); + man_parseln(curp->man, + curp->line, ln.buf, of) : + mdoc_parseln(curp->mdoc, + curp->line, ln.buf, of); if (0 == rc) { assert(MANDOCLEVEL_FATAL <= curp->file_status); @@ -676,8 +689,8 @@ mparse_end(struct mparse *curp) return; } - if ( ! (curp->man || curp->mdoc)) { - mandoc_msg(MANDOCERR_NOTMANUAL, curp, 1, 0, NULL); + if ( ! (curp->mdoc || curp->man || curp->sodest)) { + mandoc_msg(MANDOCERR_NOTMANUAL, curp, 0, 0, NULL); curp->file_status = MANDOCLEVEL_FATAL; return; } @@ -798,6 +811,9 @@ mparse_reset(struct mparse *curp) curp->file_status = MANDOCLEVEL_OK; curp->mdoc = NULL; curp->man = NULL; + + free(curp->sodest); + curp->sodest = NULL; } void @@ -814,13 +830,20 @@ mparse_free(struct mparse *curp) free(curp->secondary->buf); free(curp->secondary); + free(curp->sodest); free(curp); } void -mparse_result(struct mparse *curp, struct mdoc **mdoc, struct man **man) +mparse_result(struct mparse *curp, + struct mdoc **mdoc, struct man **man, char **sodest) { + if (sodest && NULL != (*sodest = curp->sodest)) { + *mdoc = NULL; + *man = NULL; + return; + } if (mdoc) *mdoc = curp->mdoc; if (man) @@ -835,14 +858,14 @@ mandoc_vmsg(enum mandocerr t, struct mparse *m, va_list ap; va_start(ap, fmt); - vsnprintf(buf, sizeof(buf) - 1, fmt, ap); + (void)vsnprintf(buf, sizeof(buf), fmt, ap); va_end(ap); mandoc_msg(t, m, ln, pos, buf); } void -mandoc_msg(enum mandocerr er, struct mparse *m, +mandoc_msg(enum mandocerr er, struct mparse *m, int ln, int col, const char *msg) { enum mandoclevel level;