X-Git-Url: https://git.cameronkatri.com/mandoc.git/blobdiff_plain/e1c43d6711ac3ae9f955128631c8bb0dc69772f8..e0443320f54c8a62a9b68da4d72782b1caca44c6:/cgi.c
diff --git a/cgi.c b/cgi.c
index 00f033ef..af56da27 100644
--- a/cgi.c
+++ b/cgi.c
@@ -1,37 +1,45 @@
-/* $Id: cgi.c,v 1.72 2014/07/13 12:55:45 schwarze Exp $ */
+/* $Id: cgi.c,v 1.148 2017/02/22 16:20:01 schwarze Exp $ */
/*
* Copyright (c) 2011, 2012 Kristaps Dzonsons \n"
- "This web interface is documented in the "
- "man.cgi "
- "manual, and the "
- "apropos "
+ resp_searchform(req, FOCUS_QUERY);
+ printf(" \n"
+ "This web interface is documented in the\n"
+ "man.cgi(8)\n"
+ "manual, and the\n"
+ "apropos(1)\n"
"manual explains the query syntax.\n"
- "
"); + resp_searchform(req, FOCUS_QUERY); + puts("
"); puts(msg); - puts("
"); + puts(""); resp_end_html(); } @@ -489,12 +522,12 @@ pg_error_badrequest(const char *msg) { resp_begin_html(400, "Bad Request"); - puts("\n"); + puts("
\n"); puts(msg); printf("Try again from the\n" - "main page.\n" - "
", scriptname); + "main page.\n" + "", scriptname); resp_end_html(); } @@ -502,73 +535,101 @@ static void pg_error_internal(void) { resp_begin_html(500, "Internal Server Error"); - puts("Internal Server Error
"); + puts("Internal Server Error
"); resp_end_html(); } static void pg_searchres(const struct req *req, struct manpage *r, size_t sz) { - size_t i, iuse, isec; + char *arch, *archend; + const char *sec; + size_t i, iuse; + int archprio, archpriouse; int prio, priouse; - char sec; - if (1 == sz) { + for (i = 0; i < sz; i++) { + if (validate_filename(r[i].file)) + continue; + warnx("invalid filename %s in %s database", + r[i].file, req->q.manpath); + pg_error_internal(); + return; + } + + if (req->isquery && sz == 1) { /* * If we have just one result, then jump there now * without any delay. */ printf("Status: 303 See Other\r\n"); - printf("Location: http://%s%s/%s/%s?", - httphost, scriptname, req->q.manpath, r[0].file); - http_printquery(req); + printf("Location: http://%s/%s%s%s/%s", + HTTP_HOST, scriptname, + *scriptname == '\0' ? "" : "/", + req->q.manpath, r[0].file); printf("\r\n" "Content-Type: text/html; charset=utf-8\r\n" "\r\n"); return; } - qsort(r, sz, sizeof(struct manpage), cmp); - resp_begin_html(200, NULL); - resp_searchform(req); - puts("\n" - "q.manpath, r[i].file); - html_printquery(req); - printf("\">"); - html_print(r[i].names); - printf("\n" - " | \n" - ""); - html_print(r[i].output); - puts(" | \n" - "
" + "", + scriptname, *scriptname == '\0' ? "" : "/", + req->q.manpath, r[i].file); + html_print(r[i].names); + printf(" | \n" + ""); + html_print(r[i].output); + puts(" | \n" + "
You specified an invalid manual file.
"); + if ((f = fopen(file, "r")) == NULL) { + puts("You specified an invalid manual file.
"); return; } - puts(""); + puts("\n" + "\n" + ""); + puts(""); + + p = NULL; + sz = 0; - while (NULL != (p = fgetln(f, &len))) { + while ((len = getline(&p, &sz, f)) != -1) { bold = italic = 0; - for (i = 0; i < (int)len - 1; i++) { - /* + for (i = 0; i < len - 1; i++) { + /* * This means that the catpage is out of state. * Ignore it and keep going (although the * catpage is bogus). @@ -618,22 +683,22 @@ catman(const struct req *req, const char *file) if ('\b' != p[i + 1]) { if (italic) - printf(""); + printf(""); if (bold) - printf(""); + printf(""); italic = bold = 0; html_putchar(p[i]); continue; - } else if (i + 2 >= (int)len) + } else if (i + 2 >= len) continue; /* Italic mode. */ if ('_' == p[i]) { if (bold) - printf(""); + printf(""); if ( ! italic) - printf(""); + printf(""); bold = 0; italic = 1; i += 2; @@ -641,7 +706,7 @@ catman(const struct req *req, const char *file) continue; } - /* + /* * Handle funny behaviour troff-isms. * These grok'd from the original man2html.c. */ @@ -655,9 +720,9 @@ catman(const struct req *req, const char *file) ('*' == p[i] && '|' == p[i + 2]) || ('|' == p[i] && '*' == p[i + 2])) { if (italic) - printf(""); + printf(""); if (bold) - printf(""); + printf(""); italic = bold = 0; putchar('*'); i += 2; @@ -669,9 +734,9 @@ catman(const struct req *req, const char *file) ('+' == p[i] && '|' == p[i + 1]) || ('|' == p[i] && '+' == p[i + 1])) { if (italic) - printf(""); + printf(""); if (bold) - printf(""); + printf(""); italic = bold = 0; putchar('+'); i += 2; @@ -679,114 +744,127 @@ catman(const struct req *req, const char *file) } /* Bold mode. */ - + if (italic) - printf(""); + printf(""); if ( ! bold) - printf(""); + printf(""); bold = 1; italic = 0; i += 2; html_putchar(p[i]); } - /* + /* * Clean up the last character. - * We can get to a newline; don't print that. + * We can get to a newline; don't print that. */ if (italic) - printf(""); + printf(""); if (bold) - printf(""); + printf(""); - if (i == (int)len - 1 && '\n' != p[i]) + if (i == len - 1 && p[i] != '\n') html_putchar(p[i]); putchar('\n'); } + free(p); - puts("\n" - "
You specified an invalid manual file.
"); + puts("You specified an invalid manual file.
"); return; } - mp = mparse_alloc(MPARSE_SO, MANDOCLEVEL_FATAL, NULL, - req->q.manpath); - rc = mparse_readfd(mp, fd, file); + mchars_alloc(); + mp = mparse_alloc(MPARSE_SO | MPARSE_UTF8 | MPARSE_LATIN1, + MANDOCLEVEL_BADARG, NULL, req->q.manpath); + mparse_readfd(mp, fd, file); close(fd); - if (rc >= MANDOCLEVEL_FATAL) { - fprintf(stderr, "fatal mandoc error: %s/%s\n", - req->q.manpath, file); - pg_error_internal(); - return; - } - - snprintf(opts, sizeof(opts), - "fragment,man=%s?query=%%N&sec=%%S", - scriptname); + memset(&conf, 0, sizeof(conf)); + conf.fragment = 1; + conf.style = mandoc_strdup(CSS_DIR "/mandoc.css"); + usepath = strcmp(req->q.manpath, req->p[0]); + mandoc_asprintf(&conf.man, "/%s%s%%N.%%S", + usepath ? req->q.manpath : "", usepath ? "/" : ""); - mparse_result(mp, &mdoc, &man, NULL); - if (NULL == man && NULL == mdoc) { - fprintf(stderr, "fatal mandoc error: %s/%s\n", - req->q.manpath, file); + mparse_result(mp, &man, NULL); + if (man == NULL) { + warnx("fatal mandoc error: %s/%s", req->q.manpath, file); pg_error_internal(); mparse_free(mp); + mchars_free(); return; } - vp = html_alloc(opts); + vp = html_alloc(&conf); - if (NULL != mdoc) - html_mdoc(vp, mdoc); - else + if (man->macroset == MACROSET_MDOC) { + mdoc_validate(man); + html_mdoc(vp, man); + } else { + man_validate(man); html_man(vp, man); + } html_free(vp); mparse_free(mp); + mchars_free(); + free(conf.man); + free(conf.style); } static void resp_show(const struct req *req, const char *file) { - if ('.' == file[0] || '/' == file[1]) + + if ('.' == file[0] && '/' == file[1]) file += 2; if ('c' == *file) - catman(req, file); + resp_catman(req, file); else - format(req, file); + resp_format(req, file); } static void -pg_show(const struct req *req, const char *path) +pg_show(struct req *req, const char *fullpath) { - char *sub; + char *manpath; + const char *file; - if (NULL == path || NULL == (sub = strchr(path, '/'))) { + if ((file = strchr(fullpath, '/')) == NULL) { pg_error_badrequest( "You did not specify a page to show."); return; - } - *sub++ = '\0'; + } + manpath = mandoc_strndup(fullpath, file - fullpath); + file++; + + if ( ! validate_manpath(req, manpath)) { + pg_error_badrequest( + "You specified an invalid manpath."); + free(manpath); + return; + } /* * Begin by chdir()ing into the manpath. @@ -794,15 +872,23 @@ pg_show(const struct req *req, const char *path) * relative to the manpath root. */ - if (-1 == chdir(path)) { + if (chdir(manpath) == -1) { + warn("chdir %s", manpath); + pg_error_internal(); + free(manpath); + return; + } + free(manpath); + + if ( ! validate_filename(file)) { pg_error_badrequest( - "You specified an invalid manpath."); + "You specified an invalid manual file."); return; } resp_begin_html(200, NULL); - resp_searchform(req); - resp_show(req, sub); + resp_searchform(req, FOCUS_NONE); + resp_show(req, file); resp_end_html(); } @@ -812,10 +898,10 @@ pg_search(const struct req *req) struct mansearch search; struct manpaths paths; struct manpage *res; - char **cp; - const char *ep, *start; + char **argv; + char *query, *rp, *wp; size_t ressz; - int i, sz; + int argc; /* * Begin by chdir()ing into the root of the manpath. @@ -823,62 +909,62 @@ pg_search(const struct req *req) * relative to the manpath root. */ - if (-1 == (chdir(req->q.manpath))) { - pg_error_badrequest( - "You specified an invalid manpath."); + if (chdir(req->q.manpath) == -1) { + warn("chdir %s", req->q.manpath); + pg_error_internal(); return; } search.arch = req->q.arch; search.sec = req->q.sec; - search.deftype = req->q.equal ? TYPE_Nm : (TYPE_Nm | TYPE_Nd); - search.flags = req->q.equal ? MANSEARCH_MAN : 0; + search.outkey = "Nd"; + search.argmode = req->q.equal ? ARG_NAME : ARG_EXPR; + search.firstmatch = 1; paths.sz = 1; paths.paths = mandoc_malloc(sizeof(char *)); paths.paths[0] = mandoc_strdup("."); /* - * Poor man's tokenisation: just break apart by spaces. - * Yes, this is half-ass. But it works for now. + * Break apart at spaces with backslash-escaping. */ - ep = req->q.expr; - while (ep && isspace((unsigned char)*ep)) - ep++; + argc = 0; + argv = NULL; + rp = query = mandoc_strdup(req->q.query); + for (;;) { + while (isspace((unsigned char)*rp)) + rp++; + if (*rp == '\0') + break; + argv = mandoc_reallocarray(argv, argc + 1, sizeof(char *)); + argv[argc++] = wp = rp; + for (;;) { + if (isspace((unsigned char)*rp)) { + *wp = '\0'; + rp++; + break; + } + if (rp[0] == '\\' && rp[1] != '\0') + rp++; + if (wp != rp) + *wp = *rp; + if (*rp == '\0') + break; + wp++; + rp++; + } + } - sz = 0; - cp = NULL; - while (ep && '\0' != *ep) { - cp = mandoc_reallocarray(cp, sz + 1, sizeof(char *)); - start = ep; - while ('\0' != *ep && ! isspace((unsigned char)*ep)) - ep++; - cp[sz] = mandoc_malloc((ep - start) + 1); - memcpy(cp[sz], start, ep - start); - cp[sz++][ep - start] = '\0'; - while (isspace((unsigned char)*ep)) - ep++; - } - - if (0 == mansearch(&search, &paths, sz, cp, "Nd", &res, &ressz)) + if (0 == mansearch(&search, &paths, argc, argv, &res, &ressz)) pg_noresult(req, "You entered an invalid query."); else if (0 == ressz) pg_noresult(req, "No results found."); else pg_searchres(req, res, ressz); - for (i = 0; i < sz; i++) - free(cp[i]); - free(cp); - - for (i = 0; i < (int)ressz; i++) { - free(res[i].file); - free(res[i].names); - free(res[i].output); - } - free(res); - + free(query); + mansearch_free(res, ressz); free(paths.paths[0]); free(paths.paths); } @@ -887,17 +973,38 @@ int main(void) { struct req req; + struct itimerval itimer; const char *path; - char *querystring; + const char *querystring; int i; - /* Scan our run-time environment. */ +#if HAVE_PLEDGE + /* + * The "rpath" pledge could be revoked after mparse_readfd() + * if the file desciptor to "/footer.html" would be opened + * up front, but it's probably not worth the complication + * of the code it would cause: it would require scattering + * pledge() calls in multiple low-level resp_*() functions. + */ - if (NULL == (scriptname = getenv("SCRIPT_NAME"))) - scriptname = ""; + if (pledge("stdio rpath", NULL) == -1) { + warn("pledge"); + pg_error_internal(); + return EXIT_FAILURE; + } +#endif + + /* Poor man's ReDoS mitigation. */ - if (NULL == (httphost = getenv("HTTP_HOST"))) - httphost = "localhost"; + itimer.it_value.tv_sec = 2; + itimer.it_value.tv_usec = 0; + itimer.it_interval.tv_sec = 2; + itimer.it_interval.tv_usec = 0; + if (setitimer(ITIMER_VIRTUAL, &itimer, NULL) == -1) { + warn("setitimer"); + pg_error_internal(); + return EXIT_FAILURE; + } /* * First we change directory into the MAN_DIR so that @@ -905,68 +1012,188 @@ main(void) * relative to the same position. */ - if (-1 == chdir(MAN_DIR)) { - fprintf(stderr, "MAN_DIR: %s: %s\n", - MAN_DIR, strerror(errno)); + if (chdir(MAN_DIR) == -1) { + warn("MAN_DIR: %s", MAN_DIR); pg_error_internal(); - return(EXIT_FAILURE); - } + return EXIT_FAILURE; + } memset(&req, 0, sizeof(struct req)); - pathgen(&req); - - /* Next parse out the query string. */ + req.q.equal = 1; + parse_manpath_conf(&req); - if (NULL != (querystring = getenv("QUERY_STRING"))) - http_parse(&req, querystring); + /* Parse the path info and the query string. */ - /* Dispatch to the three different pages. */ - - path = getenv("PATH_INFO"); - if (NULL == path) + if ((path = getenv("PATH_INFO")) == NULL) path = ""; - else if ('/' == *path) + else if (*path == '/') path++; + if (*path != '\0') { + parse_path_info(&req, path); + if (req.q.manpath == NULL || access(path, F_OK) == -1) + path = ""; + } else if ((querystring = getenv("QUERY_STRING")) != NULL) + parse_query_string(&req, querystring); + + /* Validate parsed data and add defaults. */ + + if (req.q.manpath == NULL) + req.q.manpath = mandoc_strdup(req.p[0]); + else if ( ! validate_manpath(&req, req.q.manpath)) { + pg_error_badrequest( + "You specified an invalid manpath."); + return EXIT_FAILURE; + } + + if ( ! (NULL == req.q.arch || validate_urifrag(req.q.arch))) { + pg_error_badrequest( + "You specified an invalid architecture."); + return EXIT_FAILURE; + } + + /* Dispatch to the three different pages. */ + if ('\0' != *path) pg_show(&req, path); - else if (NULL != req.q.expr) + else if (NULL != req.q.query) pg_search(&req); else pg_index(&req); + free(req.q.manpath); + free(req.q.arch); + free(req.q.sec); + free(req.q.query); for (i = 0; i < (int)req.psz; i++) free(req.p[i]); free(req.p); - return(EXIT_SUCCESS); + return EXIT_SUCCESS; } -static int -cmp(const void *p1, const void *p2) +/* + * If PATH_INFO is not a file name, translate it to a query. + */ +static void +parse_path_info(struct req *req, const char *path) { + char *dir[4]; + int i; + + req->isquery = 0; + req->q.equal = 1; + req->q.manpath = mandoc_strdup(path); + req->q.arch = NULL; + + /* Mandatory manual page name. */ + if ((req->q.query = strrchr(req->q.manpath, '/')) == NULL) { + req->q.query = req->q.manpath; + req->q.manpath = NULL; + } else + *req->q.query++ = '\0'; + + /* Optional trailing section. */ + if ((req->q.sec = strrchr(req->q.query, '.')) != NULL) { + if(isdigit((unsigned char)req->q.sec[1])) { + *req->q.sec++ = '\0'; + req->q.sec = mandoc_strdup(req->q.sec); + } else + req->q.sec = NULL; + } + + /* Handle the case of name[.section] only. */ + if (req->q.manpath == NULL) + return; + req->q.query = mandoc_strdup(req->q.query); + + /* Split directory components. */ + dir[i = 0] = req->q.manpath; + while ((dir[i + 1] = strchr(dir[i], '/')) != NULL) { + if (++i == 3) { + pg_error_badrequest( + "You specified too many directory components."); + exit(EXIT_FAILURE); + } + *dir[i]++ = '\0'; + } + + /* Optional manpath. */ + if ((i = validate_manpath(req, req->q.manpath)) == 0) + req->q.manpath = NULL; + else if (dir[1] == NULL) + return; + + /* Optional section. */ + if (strncmp(dir[i], "man", 3) == 0) { + free(req->q.sec); + req->q.sec = mandoc_strdup(dir[i++] + 3); + } + if (dir[i] == NULL) { + if (req->q.manpath == NULL) + free(dir[0]); + return; + } + if (dir[i + 1] != NULL) { + pg_error_badrequest( + "You specified an invalid directory component."); + exit(EXIT_FAILURE); + } - return(strcasecmp(((const struct manpage *)p1)->names, - ((const struct manpage *)p2)->names)); + /* Optional architecture. */ + if (i) { + req->q.arch = mandoc_strdup(dir[i]); + if (req->q.manpath == NULL) + free(dir[0]); + } else + req->q.arch = dir[0]; } /* * Scan for indexable paths. */ static void -pathgen(struct req *req) +parse_manpath_conf(struct req *req) { FILE *fp; char *dp; size_t dpsz; + ssize_t len; - if (NULL == (fp = fopen("manpath.conf", "r"))) - return; + if ((fp = fopen("manpath.conf", "r")) == NULL) { + warn("%s/manpath.conf", MAN_DIR); + pg_error_internal(); + exit(EXIT_FAILURE); + } - while (NULL != (dp = fgetln(fp, &dpsz))) { - if ('\n' == dp[dpsz - 1]) - dpsz--; + dp = NULL; + dpsz = 0; + + while ((len = getline(&dp, &dpsz, fp)) != -1) { + if (dp[len - 1] == '\n') + dp[--len] = '\0'; req->p = mandoc_realloc(req->p, (req->psz + 1) * sizeof(char *)); - req->p[req->psz++] = mandoc_strndup(dp, dpsz); + if ( ! validate_urifrag(dp)) { + warnx("%s/manpath.conf contains " + "unsafe path \"%s\"", MAN_DIR, dp); + pg_error_internal(); + exit(EXIT_FAILURE); + } + if (strchr(dp, '/') != NULL) { + warnx("%s/manpath.conf contains " + "path with slash \"%s\"", MAN_DIR, dp); + pg_error_internal(); + exit(EXIT_FAILURE); + } + req->p[req->psz++] = dp; + dp = NULL; + dpsz = 0; + } + free(dp); + + if (req->p == NULL) { + warnx("%s/manpath.conf is empty", MAN_DIR); + pg_error_internal(); + exit(EXIT_FAILURE); } }