=================================================================== RCS file: /cvs/mandoc/cgi.c,v retrieving revision 1.24 retrieving revision 1.41 diff -u -p -r1.24 -r1.41 --- mandoc/cgi.c 2011/12/10 21:51:07 1.24 +++ mandoc/cgi.c 2012/03/24 00:31:55 1.41 @@ -1,4 +1,4 @@ -/* $Id: cgi.c,v 1.24 2011/12/10 21:51:07 kristaps Exp $ */ +/* $Id: cgi.c,v 1.41 2012/03/24 00:31:55 kristaps Exp $ */ /* * Copyright (c) 2011 Kristaps Dzonsons * @@ -41,6 +41,7 @@ #include "man.h" #include "main.h" #include "manpath.h" +#include "mandocdb.h" #ifdef __linux__ # include @@ -67,7 +68,7 @@ struct query { const char *arch; /* architecture */ const char *sec; /* manual section */ const char *expr; /* unparsed expression string */ - int whatis; /* whether whatis mode */ + int manroot; /* manroot index (or -1)*/ int legacy; /* whether legacy mode */ }; @@ -79,13 +80,17 @@ struct req { }; static int atou(const char *, unsigned *); -static void catman(const char *); +static void catman(const struct req *, const char *); static int cmp(const void *, const void *); -static void format(const char *); +static void format(const struct req *, const char *); static void html_print(const char *); +static void html_printquery(const struct req *); static void html_putchar(char); static int http_decode(char *); -static void http_parse(struct query *, char *); +static void http_parse(struct req *, char *); +static void http_print(const char *); +static void http_putchar(char); +static void http_printquery(const struct req *); static int pathstop(DIR *); static void pathgen(DIR *, char *, struct req *); static void pg_index(const struct req *, char *); @@ -102,9 +107,10 @@ static void resp_index(const struct req *); static void resp_search(struct res *, size_t, void *); static void resp_searchform(const struct req *); -static const char *progname; -static const char *cache; -static const char *host; +static const char *progname; /* cgi script name */ +static const char *cache; /* cache directory */ +static const char *css; /* css directory */ +static const char *host; /* hostname */ static const char * const pages[PAGE__MAX] = { "index", /* PAGE_INDEX */ @@ -128,7 +134,7 @@ atou(const char *buf, unsigned *v) return(0); if ((errno == ERANGE && (lval == LONG_MAX || lval == LONG_MIN)) || - (lval > UINT_MAX || lval < 0)) + (lval > INT_MAX || lval < 0)) return(0); *v = (unsigned int)lval; @@ -161,7 +167,41 @@ html_putchar(char c) break; } } +static void +http_printquery(const struct req *req) +{ + printf("&expr="); + http_print(req->q.expr ? req->q.expr : ""); + printf("&sec="); + http_print(req->q.sec ? req->q.sec : ""); + printf("&arch="); + http_print(req->q.arch ? req->q.arch : ""); +} + + +static void +html_printquery(const struct req *req) +{ + + printf("&expr="); + html_print(req->q.expr ? req->q.expr : ""); + printf("&sec="); + html_print(req->q.sec ? req->q.sec : ""); + printf("&arch="); + html_print(req->q.arch ? req->q.arch : ""); +} + +static void +http_print(const char *p) +{ + + if (NULL == p) + return; + while ('\0' != *p) + http_putchar(*p++); +} + /* * Call through to html_putchar(). * Accepts NULL strings. @@ -182,94 +222,92 @@ html_print(const char *p) * uses only GET for simplicity. */ static void -http_parse(struct query *q, char *p) +http_parse(struct req *req, char *p) { - char *key, *val; - size_t sz; - int legacy; + char *key, *val, *manroot; + int i, legacy; - memset(q, 0, sizeof(struct query)); + memset(&req->q, 0, sizeof(struct query)); - q->whatis = 1; legacy = -1; + manroot = NULL; - while (p && '\0' != *p) { - while (' ' == *p) - p++; - + while ('\0' != *p) { key = p; val = NULL; - if (NULL != (p = strchr(p, '='))) { + p += (int)strcspn(p, ";&"); + if ('\0' != *p) *p++ = '\0'; - val = p; + if (NULL != (val = strchr(key, '='))) + *val++ = '\0'; - sz = strcspn(p, ";&"); - /* LINTED */ - p += sz; - - if ('\0' != *p) - *p++ = '\0'; - } else { - p = key; - sz = strcspn(p, ";&"); - /* LINTED */ - p += sz; - - if ('\0' != *p) - p++; + if ('\0' == *key || NULL == val || '\0' == *val) continue; - } - if ('\0' == *key || '\0' == *val) - continue; - /* Just abort handling. */ if ( ! http_decode(key)) break; - if ( ! http_decode(val)) + if (NULL != val && ! http_decode(val)) break; if (0 == strcmp(key, "expr")) - q->expr = val; + req->q.expr = val; else if (0 == strcmp(key, "query")) - q->expr = val; + req->q.expr = val; else if (0 == strcmp(key, "sec")) - q->sec = val; + req->q.sec = val; else if (0 == strcmp(key, "sektion")) - q->sec = val; + req->q.sec = val; else if (0 == strcmp(key, "arch")) - q->arch = val; + req->q.arch = val; + else if (0 == strcmp(key, "manpath")) + manroot = val; else if (0 == strcmp(key, "apropos")) legacy = 0 == strcmp(val, "0"); - else if (0 == strcmp(key, "op")) - q->whatis = 0 == strcasecmp(val, "whatis"); } /* Test for old man.cgi compatibility mode. */ - if (legacy == 0) { - q->whatis = 0; - q->legacy = 1; - } else if (legacy > 0) { - q->legacy = 1; - q->whatis = 1; - } + req->q.legacy = legacy > 0; /* * Section "0" means no section when in legacy mode. * For some man.cgi scripts, "default" arch is none. */ - if (q->legacy && NULL != q->sec) - if (0 == strcmp(q->sec, "0")) - q->sec = NULL; - if (q->legacy && NULL != q->arch) - if (0 == strcmp(q->arch, "default")) - q->arch = NULL; + if (req->q.legacy && NULL != req->q.sec) + if (0 == strcmp(req->q.sec, "0")) + req->q.sec = NULL; + if (req->q.legacy && NULL != req->q.arch) + if (0 == strcmp(req->q.arch, "default")) + req->q.arch = NULL; + + /* Default to first manroot. */ + + if (NULL != manroot) { + for (i = 0; i < (int)req->psz; i++) + if (0 == strcmp(req->p[i].name, manroot)) + break; + req->q.manroot = i < (int)req->psz ? i : -1; + } } +static void +http_putchar(char c) +{ + + if (isalnum((unsigned char)c)) { + putchar((unsigned char)c); + return; + } else if (' ' == c) { + putchar('+'); + return; + } + printf("%%%.2x", c); +} + /* * HTTP-decode a string. The standard explanation is that this turns * "%4e+foo" into "n foo" in the regular way. This is done in-place @@ -325,19 +363,21 @@ resp_begin_html(int code, const char *msg) resp_begin_http(code, msg); - puts("\n" - "\n" - "\n" - "\n" - "\n" - "System Manpage Reference\n" - "\n" - "\n" - ""); + printf("\n" + "\n" + "\n" + "\n" + "\n" + "\n" + "System Manpage Reference\n" + "\n" + "\n" + "\n", css, css); } static void @@ -351,18 +391,17 @@ resp_end_html(void) static void resp_searchform(const struct req *req) { + int i; puts(""); - printf("
\n"); - printf("
\n" + printf("
\n" + "\n" + "
\n" "Search Parameters\n" - " or \n" - " for manuals satisfying \n" - " for manuals satisfying \n" + "q.expr ? req->q.expr : ""); printf("\">, section " "q.arch ? req->q.arch : ""); - puts("\">.\n" + printf("\">"); + if (req->psz > 1) { + puts(", "); + } + puts(".\n" "\n" "
\n" - ""); + "\n" + "
"); puts(""); } @@ -440,58 +495,64 @@ resp_baddb(void) static void resp_search(struct res *r, size_t sz, void *arg) { - int i; + size_t i, matched; const struct req *req; - if (1 == sz) { + req = (const struct req *)arg; + + if (sz > 0) + assert(req->q.manroot >= 0); + + for (matched = i = 0; i < sz; i++) + if (r[i].matched) + matched++; + + if (1 == matched) { + for (i = 0; i < sz; i++) + if (r[i].matched) + break; /* * If we have just one result, then jump there now * without any delay. */ puts("Status: 303 See Other"); - printf("Location: http://%s%s/show/%u/%u.html\n", - host, progname, - r[0].volume, r[0].rec); - puts("Content-Type: text/html; charset=utf-8\n"); + printf("Location: http://%s%s/show/%d/%u/%u.html?", + host, progname, req->q.manroot, + r[i].volume, r[i].rec); + http_printquery(req); + puts("\n" + "Content-Type: text/html; charset=utf-8\n"); return; } - qsort(r, sz, sizeof(struct res), cmp); - resp_begin_html(200, NULL); - - req = (const struct req *)arg; resp_searchform(req); - if (0 == sz) { - printf("

\n" - "No %s results found.\n", - req->q.whatis ? "whatis" : "apropos"); - if (req->q.whatis) { - printf("(Try q.expr ? req->q.expr : ""); - printf("&sec="); - html_print(req->q.sec ? req->q.sec : ""); - printf("&arch="); - html_print(req->q.arch ? req->q.arch : ""); - puts("\">apropos?)"); - } - puts("

"); + puts("
"); + + if (0 == matched) { + puts("

\n" + "No results found.\n" + "

\n" + "
"); resp_end_html(); return; } - puts("

\n" - ""); + qsort(r, sz, sizeof(struct res), cmp); - for (i = 0; i < (int)sz; i++) { + puts("
"); + + for (i = 0; i < sz; i++) { + if ( ! r[i].matched) + continue; printf("\n" ""); } - puts("
\n" - "", r[i].volume, r[i].rec); + "q.manroot, + r[i].volume, r[i].rec); + html_printquery(req); + printf("\">"); html_print(r[i].title); putchar('('); html_print(r[i].cat); @@ -507,7 +568,8 @@ resp_search(struct res *r, size_t sz, void *arg) "
"); + puts("\n" + ""); resp_end_html(); } @@ -520,7 +582,7 @@ pg_index(const struct req *req, char *path) } static void -catman(const char *file) +catman(const struct req *req, const char *file) { FILE *f; size_t len; @@ -533,20 +595,9 @@ catman(const char *file) return; } - resp_begin_http(200, NULL); - puts("\n" - "\n" - "\n" - "\n" - "\n" - "System Manpage Reference\n" - "\n" - "\n" - "\n" + resp_begin_html(200, NULL); + resp_searchform(req); + puts("
\n" "
");
 
 	while (NULL != (p = fgetln(f, &len))) {
@@ -659,6 +710,7 @@ catman(const char *file)
 	}
 
 	puts("
\n" + "
\n" "\n" ""); @@ -666,7 +718,7 @@ catman(const char *file) } static void -format(const char *file) +format(const struct req *req, const char *file) { struct mparse *mp; int fd; @@ -690,23 +742,31 @@ format(const char *file) return; } - snprintf(opts, sizeof(opts), "style=/man.css," + snprintf(opts, sizeof(opts), "fragment," "man=%s/search.html?sec=%%S&expr=%%N," /*"includes=/cgi-bin/man.cgi/usr/include/%%I"*/, progname); mparse_result(mp, &mdoc, &man); + if (NULL == man && NULL == mdoc) { + resp_baddb(); + mparse_free(mp); + return; + } + + resp_begin_html(200, NULL); + resp_searchform(req); + vp = html_alloc(opts); - if (NULL != mdoc) { - resp_begin_http(200, NULL); + if (NULL != mdoc) html_mdoc(vp, mdoc); - } else if (NULL != man) { - resp_begin_http(200, NULL); + else html_man(vp, man); - } else - resp_baddb(); + puts("\n" + ""); + html_free(vp); mparse_free(mp); } @@ -715,50 +775,66 @@ static void pg_show(const struct req *req, char *path) { struct manpaths ps; + size_t sz; char *sub; char file[MAXPATHLEN]; - const char *fn, *cp; - int rc; - unsigned int vol, rec; + const char *cp; + int rc, catm; + unsigned int vol, rec, mr; DB *idx; DBT key, val; idx = NULL; - if (0 == req->psz || NULL == path) { + /* Parse out mroot, volume, and record from the path. */ + + if (NULL == path || NULL == (sub = strchr(path, '/'))) { resp_error400(); return; - } else if (NULL == (sub = strrchr(path, '/'))) { + } + *sub++ = '\0'; + if ( ! atou(path, &mr)) { resp_error400(); return; - } else - *sub++ = '\0'; + } + path = sub; + if (NULL == (sub = strchr(path, '/'))) { + resp_error400(); + return; + } + *sub++ = '\0'; + if ( ! atou(path, &vol) || ! atou(sub, &rec)) { + resp_error400(); + return; + } else if (mr >= (unsigned int)req->psz) { + resp_error400(); + return; + } /* - * Begin by chdir()ing into the root of the manpath. + * Begin by chdir()ing into the manroot. * This way we can pick up the database files, which are * relative to the manpath root. */ - if (-1 == chdir(req->p[0].path)) { - perror(req->p[0].path); - resp_error400(); + if (-1 == chdir(req->p[(int)mr].path)) { + perror(req->p[(int)mr].path); + resp_baddb(); return; } memset(&ps, 0, sizeof(struct manpaths)); - manpath_manconf("etc/catman.conf", &ps); + manpath_manconf(&ps, "etc/catman.conf"); - if ( ! (atou(path, &vol) && atou(sub, &rec))) { + if (vol >= (unsigned int)ps.sz) { resp_error400(); goto out; - } else if (vol >= (unsigned int)ps.sz) { - resp_error400(); - goto out; } - strlcpy(file, ps.paths[vol], MAXPATHLEN); - strlcat(file, "/mandoc.index", MAXPATHLEN); + sz = strlcpy(file, ps.paths[vol], MAXPATHLEN); + assert(sz < MAXPATHLEN); + strlcat(file, "/", MAXPATHLEN); + strlcat(file, MANDOC_IDX, MAXPATHLEN); /* Open the index recno(3) database. */ @@ -775,21 +851,24 @@ pg_show(const struct req *req, char *path) if (0 != (rc = (*idx->get)(idx, &key, &val, 0))) { rc < 0 ? resp_baddb() : resp_error400(); goto out; - } + } else if (0 == val.size) { + resp_baddb(); + goto out; + } cp = (char *)val.data; + catm = 'c' == *cp++; - if (NULL == (fn = memchr(cp, '\0', val.size))) + if (NULL == memchr(cp, '\0', val.size - 1)) resp_baddb(); - else if (++fn - cp >= (int)val.size) - resp_baddb(); - else if (NULL == memchr(fn, '\0', val.size - (fn - cp))) - resp_baddb(); else { - if (0 == strcmp(cp, "cat")) - catman(fn + 1); + file[(int)sz] = '\0'; + strlcat(file, "/", MAXPATHLEN); + strlcat(file, cp, MAXPATHLEN); + if (catm) + catman(req, file); else - format(fn + 1); + format(req, file); } out: if (idx) @@ -800,15 +879,16 @@ out: static void pg_search(const struct req *req, char *path) { - size_t tt; + size_t tt, ressz; struct manpaths ps; int i, sz, rc; const char *ep, *start; + struct res *res; char **cp; struct opts opt; struct expr *expr; - if (0 == req->psz) { + if (req->q.manroot < 0 || 0 == req->psz) { resp_search(NULL, 0, (void *)req); return; } @@ -821,6 +901,8 @@ pg_search(const struct req *req, char *path) rc = -1; sz = 0; cp = NULL; + ressz = 0; + res = NULL; /* * Begin by chdir()ing into the root of the manpath. @@ -828,14 +910,15 @@ pg_search(const struct req *req, char *path) * relative to the manpath root. */ - if (-1 == (chdir(req->p[0].path))) { - perror(req->p[0].path); + assert(req->q.manroot < (int)req->psz); + if (-1 == (chdir(req->p[req->q.manroot].path))) { + perror(req->p[req->q.manroot].path); resp_search(NULL, 0, (void *)req); return; } memset(&ps, 0, sizeof(struct manpaths)); - manpath_manconf("etc/catman.conf", &ps); + manpath_manconf(&ps, "etc/catman.conf"); /* * Poor man's tokenisation: just break apart by spaces. @@ -862,25 +945,26 @@ pg_search(const struct req *req, char *path) * The resp_search() function is called with the results. */ - expr = req->q.whatis ? + expr = req->q.legacy ? termcomp(sz, cp, &tt) : exprcomp(sz, cp, &tt); if (NULL != expr) rc = apropos_search - (ps.sz, ps.paths, &opt, - expr, tt, (void *)req, resp_search); + (ps.sz, ps.paths, &opt, expr, tt, + (void *)req, &ressz, &res, resp_search); /* ...unless errors occured. */ if (0 == rc) resp_baddb(); else if (-1 == rc) - resp_search(NULL, 0, (void *)req); + resp_search(NULL, 0, NULL); for (i = 0; i < sz; i++) free(cp[i]); free(cp); + resfree(res, ressz); exprfree(expr); manpath_free(&ps); } @@ -896,16 +980,16 @@ main(void) /* Scan our run-time environment. */ - progname = getenv("SCRIPT_NAME"); - if (NULL == progname) + if (NULL == (cache = getenv("CACHE_DIR"))) + cache = "/cache/man.cgi"; + + if (NULL == (progname = getenv("SCRIPT_NAME"))) progname = ""; - cache = getenv("CACHE_DIR"); - if (NULL == cache) - cache = "/cache/man.cgi"; + if (NULL == (css = getenv("CSS_DIR"))) + css = ""; - host = getenv("HTTP_HOST"); - if (NULL == host) + if (NULL == (host = getenv("HTTP_HOST"))) host = "localhost"; /* @@ -933,7 +1017,7 @@ main(void) /* Next parse out the query string. */ if (NULL != (p = getenv("QUERY_STRING"))) - http_parse(&req.q, p); + http_parse(&req, p); /* * Now juggle paths to extract information. @@ -1070,14 +1154,17 @@ pathgen(DIR *dir, char *path, struct req *req) if (rc > 0) { /* This also strips the trailing slash. */ - path[(int)sz - 1] = '\0'; + path[(int)--sz] = '\0'; req->p = mandoc_realloc (req->p, (req->psz + 1) * sizeof(struct paths)); + /* + * Strip out the leading "./" unless we're just a ".", + * in which case use an empty string as our name. + */ req->p[(int)req->psz].path = mandoc_strdup(path); - /* And this strips out the leading "./". */ req->p[(int)req->psz].name = - cp = mandoc_strdup(path + 2); + cp = mandoc_strdup(path + (1 == sz ? 1 : 2)); req->psz++; /* * The name is just the path with all the slashes taken