=================================================================== RCS file: /cvs/mandoc/read.c,v retrieving revision 1.57 retrieving revision 1.82 diff -u -p -r1.57 -r1.82 --- mandoc/read.c 2014/07/02 13:10:45 1.57 +++ mandoc/read.c 2014/09/03 23:21:47 1.82 @@ -1,4 +1,4 @@ -/* $Id: read.c,v 1.57 2014/07/02 13:10:45 schwarze Exp $ */ +/* $Id: read.c,v 1.82 2014/09/03 23:21:47 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2010-2014 Ingo Schwarze @@ -16,14 +16,14 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ -#ifdef HAVE_CONFIG_H #include "config.h" -#endif -#ifdef HAVE_MMAP -# include -# include +#include +#if HAVE_MMAP +#include +#include #endif +#include #include #include @@ -65,7 +65,7 @@ struct mparse { mandocmsg mmsg; /* warning/error message handler */ const char *file; struct buf *secondary; - char *defos; /* default operating system */ + const char *defos; /* default operating system */ }; static void resize_buf(struct buf *, size_t); @@ -93,16 +93,19 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "generic warning", /* related to the prologue */ - "missing .TH macro, using \"unknown 1\"", + "missing manual title, using UNTITLED", + "missing manual title, using \"\"", "lower case character in document title", + "missing manual section, using \"\"", "unknown manual section", "unknown manual volume or arch", "missing date, using today's date", "cannot parse date, using it verbatim", - "prologue macros out of order", + "missing Os macro, using \"\"", "duplicate prologue macro", - "incomplete prologue, terminated by", - "skipping prologue macro in body", + "late prologue macro", + "skipping late title macro", + "prologue macros out of order", /* related to document structure */ ".so is fragile, better use ln(1)", @@ -123,36 +126,45 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "nested displays are not portable", "moving content out of list", ".Vt block has child macro", - "fill mode already enabled, skipping .fi", - "fill mode already disabled, skipping .nf", + "fill mode already enabled, skipping", + "fill mode already disabled, skipping", "line scope broken", /* related to missing macro arguments */ + "skipping empty request", + "conditional request controls empty scope", "skipping empty macro", + "empty argument, using 0n", "argument count wrong", - "missing display type", - "list type must come first", - "tag lists require a width argument", - "missing font type", + "missing display type, using -ragged", + "list type is not the first argument", + "missing -width in -tag list, using 8n", + "missing utility name, using \"\"", + "empty head in list item", + "empty list item", + "missing font type, using \\fR", + "unknown font type, using \\fR", + "missing -std argument, adding it", /* related to bad macro arguments */ - "skipping argument", + "unterminated quoted argument", "duplicate argument", - "duplicate display type", - "duplicate list type", + "skipping duplicate argument", + "skipping duplicate display type", + "skipping duplicate list type", + "skipping -width argument", "unknown AT&T UNIX version", - "bad Boolean value", - "unknown font", - "unknown standard specifier", - "bad width argument", + "invalid content in Rs block", + "invalid Boolean argument", + "unknown font, skipping request", /* related to plain text */ - "blank line in non-literal context", - "tab in non-literal context", - "end of line whitespace", + "blank line in fill mode, using .sp", + "tab in filled text", + "whitespace at end of input line", "bad comment style", - "bad escape sequence", - "unterminated quoted string", + "invalid escape sequence", + "undefined string, using \"\"", "generic error", @@ -173,45 +185,45 @@ static const char * const mandocerrs[MANDOCERR_MAX] = "data block still open", "ignoring extra data cells", + /* related to document structure and macros */ "input stack limit exceeded, infinite loop?", "skipping bad character", - "escaped character not allowed in a name", - "manual name not yet set", - "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 item outside list", "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", + "inserting missing end of block", + "appending missing end of block", + /* related to request and macro arguments */ + "escaped character not allowed in a name", + "argument count wrong", + "missing list type, using -item", + "missing manual name, using \"\"", + "uname(3) system call failed, using UNKNOWN", + "unknown standard specifier", + "skipping request without numeric argument", + "skipping all arguments", + "skipping excess arguments", + "generic fatal error", "input too large", - "not a manual", - "column syntax is inconsistent", - "NOT IMPLEMENTED: .Bd -file", - "argument count wrong, violates syntax", - "child violates parent syntax", - "argument count wrong, violates syntax", + "NOT IMPLEMENTED: Bd -file", "NOT IMPLEMENTED: .so with absolute path or \"..\"", ".so request failed", - "no document prologue", - "static buffer exhausted", /* system errors */ + "cannot dup file descriptor", + "cannot exec", + "gunzip failed with code", + "cannot fork", NULL, - "cannot stat file", + "cannot open pipe", "cannot read file", + "gunzip died from signal", + "cannot stat file", + "wait failed", }; static const char * const mandoclevels[MANDOCLEVEL_MAX] = { @@ -258,18 +270,9 @@ pset(const char *buf, int pos, struct mparse *curp) } if (MPARSE_MDOC & curp->options) { - 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; } else if (MPARSE_MAN & curp->options) { - if (NULL == curp->pman) - curp->pman = man_alloc(curp->roff, curp, - MPARSE_QUICK & curp->options ? 1 : 0); - assert(curp->pman); curp->man = curp->pman; return; } @@ -360,8 +363,8 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int if ( ! (isascii(c) && (isgraph(c) || isblank(c)))) { - mandoc_msg(MANDOCERR_BADCHAR, curp, - curp->line, pos, NULL); + mandoc_vmsg(MANDOCERR_BADCHAR, curp, + curp->line, pos, "0x%x", c); i++; ln.buf[pos++] = '?'; continue; @@ -417,8 +420,8 @@ mparse_buf_r(struct mparse *curp, struct buf blk, int if ( ! (isascii(c) && (isgraph(c) || isblank(c)))) { - mandoc_msg(MANDOCERR_BADCHAR, curp, - curp->line, pos, NULL); + mandoc_vmsg(MANDOCERR_BADCHAR, curp, + curp->line, pos, "0x%x", c); i += 2; ln.buf[pos++] = '?'; continue; @@ -595,7 +598,7 @@ read_whole_file(struct mparse *curp, const char *file, size_t off; ssize_t ssz; -#ifdef HAVE_MMAP +#if HAVE_MMAP struct stat st; if (-1 == fstat(fd, &st)) { curp->file_status = MANDOCLEVEL_SYSERR; @@ -677,6 +680,19 @@ mparse_end(struct mparse *curp) if (MANDOCLEVEL_FATAL <= curp->file_status) return; + if (curp->mdoc == NULL && + curp->man == NULL && + curp->sodest == NULL) { + if (curp->options & MPARSE_MDOC) + curp->mdoc = curp->pmdoc; + else { + if (curp->pman == NULL) + curp->pman = man_alloc(curp->roff, curp, + curp->options & MPARSE_QUICK ? 1 : 0); + curp->man = curp->pman; + } + } + if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) { assert(MANDOCLEVEL_FATAL <= curp->file_status); return; @@ -687,12 +703,6 @@ mparse_end(struct mparse *curp) return; } - if ( ! (curp->mdoc || curp->man || curp->sodest)) { - mandoc_msg(MANDOCERR_NOTMANUAL, curp, 0, 0, NULL); - curp->file_status = MANDOCLEVEL_FATAL; - return; - } - roff_endparse(curp->roff); } @@ -761,7 +771,7 @@ mparse_readfd(struct mparse *curp, int fd, const char mparse_parse_buffer(curp, blk, file); -#ifdef HAVE_MMAP +#if HAVE_MMAP if (with_mmap) munmap(blk.buf, blk.sz); else @@ -774,9 +784,94 @@ out: return(curp->file_status); } +enum mandoclevel +mparse_open(struct mparse *curp, int *fd, const char *file, + pid_t *child_pid) +{ + int pfd[2]; + char *cp; + enum mandocerr err; + + pfd[1] = -1; + curp->file = file; + if ((cp = strrchr(file, '.')) == NULL || + strcmp(cp + 1, "gz")) { + *child_pid = 0; + if ((*fd = open(file, O_RDONLY)) == -1) { + err = MANDOCERR_SYSOPEN; + goto out; + } + return(MANDOCLEVEL_OK); + } + + if (pipe(pfd) == -1) { + err = MANDOCERR_SYSPIPE; + goto out; + } + + switch (*child_pid = fork()) { + case -1: + err = MANDOCERR_SYSFORK; + close(pfd[0]); + close(pfd[1]); + pfd[1] = -1; + break; + case 0: + close(pfd[0]); + if (dup2(pfd[1], STDOUT_FILENO) == -1) { + err = MANDOCERR_SYSDUP; + break; + } + execlp("gunzip", "gunzip", "-c", file, NULL); + err = MANDOCERR_SYSEXEC; + break; + default: + close(pfd[1]); + *fd = pfd[0]; + return(MANDOCLEVEL_OK); + } + +out: + *fd = -1; + *child_pid = 0; + curp->file_status = MANDOCLEVEL_SYSERR; + if (curp->mmsg) + (*curp->mmsg)(err, curp->file_status, file, + 0, 0, strerror(errno)); + if (pfd[1] != -1) + exit(1); + return(curp->file_status); +} + +enum mandoclevel +mparse_wait(struct mparse *curp, pid_t child_pid) +{ + int status; + + if (waitpid(child_pid, &status, 0) == -1) { + mandoc_msg(MANDOCERR_SYSWAIT, curp, 0, 0, + strerror(errno)); + curp->file_status = MANDOCLEVEL_SYSERR; + return(curp->file_status); + } + if (WIFSIGNALED(status)) { + mandoc_vmsg(MANDOCERR_SYSSIG, curp, 0, 0, + "%d", WTERMSIG(status)); + curp->file_status = MANDOCLEVEL_SYSERR; + return(curp->file_status); + } + if (WEXITSTATUS(status)) { + mandoc_vmsg(MANDOCERR_SYSEXIT, curp, 0, 0, + "%d", WEXITSTATUS(status)); + curp->file_status = MANDOCLEVEL_SYSERR; + return(curp->file_status); + } + return(MANDOCLEVEL_OK); +} + struct mparse * mparse_alloc(int options, enum mandoclevel wlevel, - mandocmsg mmsg, char *defos) + mandocmsg mmsg, const char *defos) { struct mparse *curp; @@ -790,6 +885,14 @@ mparse_alloc(int options, enum mandoclevel wlevel, curp->defos = defos; curp->roff = roff_alloc(curp, options); + if (curp->options & MPARSE_MDOC) + curp->pmdoc = mdoc_alloc( + curp->roff, curp, curp->defos, + curp->options & MPARSE_QUICK ? 1 : 0); + if (curp->options & MPARSE_MAN) + curp->pman = man_alloc(curp->roff, curp, + curp->options & MPARSE_QUICK ? 1 : 0); + return(curp); }