=================================================================== RCS file: /cvs/mandoc/read.c,v retrieving revision 1.165 retrieving revision 1.196 diff -u -p -r1.165 -r1.196 --- mandoc/read.c 2017/05/16 19:06:30 1.165 +++ mandoc/read.c 2018/07/28 18:34:15 1.196 @@ -1,7 +1,7 @@ -/* $Id: read.c,v 1.165 2017/05/16 19:06:30 schwarze Exp $ */ +/* $Id: read.c,v 1.196 2018/07/28 18:34:15 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons - * Copyright (c) 2010-2017 Ingo Schwarze + * Copyright (c) 2010-2018 Ingo Schwarze * Copyright (c) 2010, 2012 Joerg Sonnenberger * * Permission to use, copy, modify, and distribute this software for any @@ -24,9 +24,6 @@ #include #include -#if HAVE_ERR -#include -#endif #include #include #include @@ -42,7 +39,6 @@ #include "mdoc.h" #include "man.h" #include "libmandoc.h" -#include "roff_int.h" #define REPARSE_LIMIT 1000 @@ -53,10 +49,10 @@ struct mparse { const char *file; /* filename of current input file */ struct buf *primary; /* buffer currently being parsed */ struct buf *secondary; /* preprocessed copy of input */ - const char *defos; /* default operating system */ + const char *os_s; /* default operating system */ mandocmsg mmsg; /* warning/error message handler */ enum mandoclevel file_status; /* status of current parse */ - enum mandoclevel wlevel; /* ignore messages below this */ + enum mandocerr mmin; /* ignore messages below this */ int options; /* parser options */ int gzip; /* current input file is gzipped */ int filenc; /* encoding of the current file */ @@ -75,7 +71,7 @@ static void mparse_parse_buffer(struct mparse *, str static const enum mandocerr mandoclimits[MANDOCLEVEL_MAX] = { MANDOCERR_OK, - MANDOCERR_STYLE, + MANDOCERR_OK, MANDOCERR_WARNING, MANDOCERR_ERROR, MANDOCERR_UNSUPP, @@ -86,21 +82,48 @@ static const enum mandocerr mandoclimits[MANDOCLEVEL_M static const char * const mandocerrs[MANDOCERR_MAX] = { "ok", + "base system convention", + + "Mdocdate found", + "Mdocdate missing", + "unknown architecture", + "operating system explicitly specified", + "RCS id missing", + "referenced manual not found", + "generic style suggestion", + + "legacy man(7) date format", + "normalizing date format to", + "lower case character in document title", + "duplicate RCS id", + "possible typo in section name", + "unterminated quoted argument", + "useless macro", + "consider using OS macro", + "errnos out of order", + "duplicate errno", + "trailing delimiter", + "no blank before trailing delimiter", + "fill mode already enabled, skipping", + "fill mode already disabled, skipping", + "verbatim \"--\", maybe consider using \\(em", + "function name without markup", + "whitespace at end of input line", + "bad comment style", + "generic warning", /* related to the prologue */ "missing manual title, using UNTITLED", "missing manual title, using \"\"", - "lower case character in document title", "missing manual section, using \"\"", "unknown manual section", "missing date, using today's date", "cannot parse date, using it verbatim", + "date in the future, using it anyway", "missing Os macro, using \"\"", - "duplicate prologue macro", "late prologue macro", - "skipping late title macro", "prologue macros out of order", /* related to document structure */ @@ -118,6 +141,7 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "sections out of conventional order", "duplicate section title", "unexpected section", + "cross reference to self", "unusual Xr order", "unusual Xr punctuation", "AUTHORS section without An macro", @@ -131,9 +155,9 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "blocks badly nested", "nested displays are not portable", "moving content out of list", - "fill mode already enabled, skipping", - "fill mode already disabled, skipping", + "first macro on line", "line scope broken", + "skipping blank line in line scope", /* related to missing macro arguments */ "skipping empty request", @@ -148,6 +172,7 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "missing function name, using \"\"", "empty head in list item", "empty list item", + "missing argument, using next line", "missing font type, using \\fR", "unknown font type, using \\fR", "nothing follows prefix", @@ -159,7 +184,6 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "missing eqn box, using \"\"", /* related to bad macro arguments */ - "unterminated quoted argument", "duplicate argument", "skipping duplicate argument", "skipping duplicate display type", @@ -169,6 +193,7 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "unknown AT&T UNIX version", "comma in function argument", "parenthesis in function name", + "unknown library name", "invalid content in Rs block", "invalid Boolean argument", "unknown font, skipping request", @@ -177,9 +202,7 @@ static const char * const mandocerrs[MANDOCERR_MAX] = /* related to plain text */ "blank line in fill mode, using .sp", "tab in filled text", - "whitespace at end of input line", "new sentence, new line", - "bad comment style", "invalid escape sequence", "undefined string, using \"\"", @@ -205,6 +228,8 @@ static const char * const mandocerrs[MANDOCERR_MAX] = /* related to document structure and macros */ NULL, + "duplicate prologue macro", + "skipping late title macro", "input stack limit exceeded, infinite loop?", "skipping bad character", "skipping unknown macro", @@ -221,6 +246,7 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "NOT IMPLEMENTED: Bd -file", "skipping display without arguments", "missing list type, using -item", + "argument is not numeric, using 1", "missing manual name, using \"\"", "uname(3) system call failed, using UNKNOWN", "unknown standard specifier", @@ -315,12 +341,10 @@ choose_parser(struct mparse *curp) static int mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start) { - const struct tbl_span *span; struct buf ln; const char *save_file; char *cp; size_t pos; /* byte number in the ln buffer */ - size_t j; /* auxiliary byte number in the blk buffer */ enum rofferr rr; int of; int lnn; /* line number in the real file */ @@ -402,79 +426,14 @@ mparse_buf_r(struct mparse *curp, struct buf blk, size continue; } - /* Trailing backslash = a plain char. */ - - if (blk.buf[i] != '\\' || i + 1 == blk.sz) { - ln.buf[pos++] = blk.buf[i++]; - continue; - } - - /* - * Found escape and at least one other character. - * When it's a newline character, skip it. - * When there is a carriage return in between, - * skip that one as well. - */ - - if ('\r' == blk.buf[i + 1] && i + 2 < blk.sz && - '\n' == blk.buf[i + 2]) - ++i; - if ('\n' == blk.buf[i + 1]) { - i += 2; - ++lnn; - continue; - } - - if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) { - j = i; - i += 2; - /* Comment, skip to end of line */ - for (; i < blk.sz; ++i) { - if (blk.buf[i] != '\n') - continue; - if (blk.buf[i - 1] == ' ' || - blk.buf[i - 1] == '\t') - mandoc_msg( - MANDOCERR_SPACE_EOL, - curp, curp->line, - pos + i-1 - j, NULL); - ++i; - ++lnn; - break; - } - - /* Backout trailing whitespaces */ - for (; pos > 0; --pos) { - if (ln.buf[pos - 1] != ' ') - break; - if (pos > 2 && ln.buf[pos - 2] == '\\') - break; - } - break; - } - - /* Catch escaped bogus characters. */ - - c = (unsigned char) blk.buf[i+1]; - - if ( ! (isascii(c) && - (isgraph(c) || isblank(c)))) { - mandoc_vmsg(MANDOCERR_CHAR_BAD, curp, - curp->line, pos, "0x%x", c); - i += 2; - ln.buf[pos++] = '?'; - continue; - } - - /* Some other escape sequence, copy & cont. */ - ln.buf[pos++] = blk.buf[i++]; - ln.buf[pos++] = blk.buf[i++]; } - if (pos >= ln.sz) + if (pos + 1 >= ln.sz) resize_buf(&ln, 256); + if (i == blk.sz || blk.buf[i] == '\0') + ln.buf[pos++] = '\n'; ln.buf[pos] = '\0'; /* @@ -572,21 +531,7 @@ rerun: if (curp->man->macroset == MACROSET_NONE) choose_parser(curp); - /* - * Lastly, push down into the parsers themselves. - * If libroff returns ROFF_TBL, then add it to the - * currently open parse. Since we only get here if - * there does exist data (see tbl_data.c), we're - * guaranteed that something's been allocated. - * Do the same for ROFF_EQN. - */ - - if (rr == ROFF_TBL) - while ((span = roff_span(curp->roff)) != NULL) - roff_addtbl(curp->man, span); - else if (rr == ROFF_EQN) - roff_addeqn(curp->man, roff_eqn(curp->roff)); - else if ((curp->man->macroset == MACROSET_MDOC ? + if ((curp->man->macroset == MACROSET_MDOC ? mdoc_parseln(curp->man, curp->line, ln.buf, of) : man_parseln(curp->man, curp->line, ln.buf, of)) == 2) break; @@ -613,9 +558,13 @@ read_whole_file(struct mparse *curp, const char *file, gzFile gz; size_t off; ssize_t ssz; + int gzerrnum, retval; - if (fstat(fd, &st) == -1) - err((int)MANDOCLEVEL_SYSERR, "%s", file); + if (fstat(fd, &st) == -1) { + mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, + "fstat: %s", strerror(errno)); + return 0; + } /* * If we're a regular file, try just reading in the whole entry @@ -637,8 +586,24 @@ read_whole_file(struct mparse *curp, const char *file, } if (curp->gzip) { - if ((gz = gzdopen(fd, "rb")) == NULL) - err((int)MANDOCLEVEL_SYSERR, "%s", file); + /* + * Duplicating the file descriptor is required + * because we will have to call gzclose(3) + * to free memory used internally by zlib, + * but that will also close the file descriptor, + * which this function must not do. + */ + if ((fd = dup(fd)) == -1) { + mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, + "dup: %s", strerror(errno)); + return 0; + } + if ((gz = gzdopen(fd, "rb")) == NULL) { + mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, + "gzdopen: %s", strerror(errno)); + close(fd); + return 0; + } } else gz = NULL; @@ -649,6 +614,7 @@ read_whole_file(struct mparse *curp, const char *file, *with_mmap = 0; off = 0; + retval = 0; fb->sz = 0; fb->buf = NULL; for (;;) { @@ -665,16 +631,29 @@ read_whole_file(struct mparse *curp, const char *file, read(fd, fb->buf + (int)off, fb->sz - off); if (ssz == 0) { fb->sz = off; - return 1; + retval = 1; + break; } - if (ssz == -1) - err((int)MANDOCLEVEL_SYSERR, "%s", file); + if (ssz == -1) { + if (curp->gzip) + (void)gzerror(gz, &gzerrnum); + mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, "read: %s", + curp->gzip && gzerrnum != Z_ERRNO ? + zError(gzerrnum) : strerror(errno)); + break; + } off += (size_t)ssz; } - free(fb->buf); - fb->buf = NULL; - return 0; + if (curp->gzip && (gzerrnum = gzclose(gz)) != Z_OK) + mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, "gzclose: %s", + gzerrnum == Z_ERRNO ? strerror(errno) : + zError(gzerrnum)); + if (retval == 0) { + free(fb->buf); + fb->buf = NULL; + } + return retval; } static void @@ -804,20 +783,20 @@ mparse_open(struct mparse *curp, const char *file) } struct mparse * -mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg, - const char *defos) +mparse_alloc(int options, enum mandocerr mmin, mandocmsg mmsg, + enum mandoc_os os_e, const char *os_s) { struct mparse *curp; curp = mandoc_calloc(1, sizeof(struct mparse)); curp->options = options; - curp->wlevel = wlevel; + curp->mmin = mmin; curp->mmsg = mmsg; - curp->defos = defos; + curp->os_s = os_s; curp->roff = roff_alloc(curp, options); - curp->man = roff_man_alloc( curp->roff, curp, curp->defos, + curp->man = roff_man_alloc(curp->roff, curp, curp->os_s, curp->options & MPARSE_QUICK ? 1 : 0); if (curp->options & MPARSE_MDOC) { curp->man->macroset = MACROSET_MDOC; @@ -829,6 +808,7 @@ mparse_alloc(int options, enum mandoclevel wlevel, man curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX); } curp->man->first->tok = TOKEN_NONE; + curp->man->meta.os_e = os_e; return curp; } @@ -904,12 +884,12 @@ mandoc_msg(enum mandocerr er, struct mparse *m, { enum mandoclevel level; + if (er < m->mmin && er != MANDOCERR_FILE) + return; + level = MANDOCLEVEL_UNSUPP; while (er < mandoclimits[level]) level--; - - if (level < m->wlevel && er != MANDOCERR_FILE) - return; if (m->mmsg) (*m->mmsg)(er, level, m->file, ln, col, msg);