feat(rss,atom): use SERVER_{NAME,PORT} for external URL

6 files changed, 85 insertions, 42 deletions

diff --git a/TODO b/TODO
index a2e1b3c..1470d43 100644
--- a/TODO
+++ b/TODO
@@ -1,6 +1,5 @@
 introduce better configuration mechanism | id:1f8f1bd3a99e6f394da4fbb508f1355759c2eeec
 rethink logging / debug output | id:5734e2717f28d80ac65d3f2123a0f46c01d4a0c1
-if possible determine BLOG_SERVER_URL automatically | id:5ad1fabab42a5776ab1446d09659c7a99a104245
 support markdown markup via lowdown | id:90c72d144f21d78d1b56790a927f6eea52df2a2c
 templating: give all functions context (index or not, SCRIPT_NAME, entry struct if applicable) | id:930c2e4f2c4c9ec3efef87b6ddd50a6ba6dd4903
 add tests for timeutil | id:b6cd0665e5c376f52cf9702698330308aa722031
@@ -9,4 +8,4 @@ improve man3 situation | id:bf918930f6361c30f288c42894cbf9b0541c3340
 use errno instead of returning HTTP status codes in make_index | id:bfa35c48dcedbecd3a951eba243c4840178289ed
 templating: unify single_entry and index_entry | id:e6dcb195d890bb0b5186716808d806101d81e4f6
 sandboxing, at least chroot | id:f209aec893fe310c8276f64f6ff4a3208a2a4f28
-document timeutil and stringutil | id:fd05b97115848a8b22231a477b8252373bd87264
+document cgiutil, timeutil and stringutil | id:fd05b97115848a8b22231a477b8252373bd87264
diff --git a/cgiutil.c b/cgiutil.c
index 135028d..7dc8114 100644
--- a/cgiutil.c
+++ b/cgiutil.c
@@ -115,3 +115,16 @@ int urlencode_realloc(char **input, int size) {
 
     return output_size;
 }
+
+char *server_url(bool https) {
+    char *server_name = getenv("SERVER_NAME");
+    char *server_port = getenv("SERVER_PORT");
+
+    if(server_name == NULL || server_port == NULL) {
+        return NULL;
+    }
+
+    char *proto = https ? "https://" : "http://";
+
+    return catn_alloc(4, proto, server_name, ":", server_port);
+}
diff --git a/cgiutil.h b/cgiutil.h
index 4701e83..b90b895 100644
--- a/cgiutil.h
+++ b/cgiutil.h
@@ -3,6 +3,8 @@
  * @brief Simple CGI/HTTP helper functions used by sternenblog.
  */
 
+#include <stdbool.h>
+
 /*!
  * @brief Print a HTTP header
  *
@@ -89,3 +91,5 @@ int http_errno(int err);
  * @return -1 on error, else size of buffer
  */
 int urlencode_realloc(char **input, int size);
+
+char *server_url(bool https);
diff --git a/config.example.h b/config.example.h
index b089d21..372d266 100644
--- a/config.example.h
+++ b/config.example.h
@@ -18,14 +18,15 @@
 #define BLOG_DIR "/srv/sternenblog/"
 
 /*!
- * @brief Public URL of the webserver
+ * @brief Wether the server uses https for sternenblog
  *
- * (External) URL the server that the script is running on is reachable via.
- * This should be the base URL of your webserver without any trailing slashes
- * or subpaths, as sternenblog will use SCRIPT_NAME and PATH_INFO to figure
- * out the path from the root of the server.
+ * sternenblog will use `BLOG_USE_HTTPS` when building an external
+ * URL to pages: If it is `1`, it will use `https://` as prefix, otherwise
+ * `http://`.
+ *
+ * This is currently only relevant for the RSS and Atom feeds.
  */
-#define BLOG_SERVER_URL "http://localhost"
+#define BLOG_USE_HTTPS 1
 
 /*!
  * @brief Enable / Disable strict access check
diff --git a/doc/man/man1/sternenblog.cgi.1 b/doc/man/man1/sternenblog.cgi.1
index d0af4e3..5cab28e 100644
--- a/doc/man/man1/sternenblog.cgi.1
+++ b/doc/man/man1/sternenblog.cgi.1
@@ -65,14 +65,18 @@ although the latter, i. e. an absolute path, is recommended.
 .Pp
 Default value is
 .Pa /srv/sternenblog/ .
-.It Sy BLOG_SERVER_URL
-Public URL of the root of the server, i. e. public address or hostname and used
-protocol.
-This is used to construct global URLs to the entries for the RSS and Atom
-feeds.
+.It Sy BLOG_USE_HTTPS
+Wether the webserver is configured to serve
+.Nm
+via https or not.
+This information will be used whe building external URLs to entries in the RSS
+and Atom feeds to decide wether to prefix the URL with
+.Ql https://
+or
+.Ql http:// .
 .Pp
 Default value is
-.Ql http://localhost .
+.Ql 1 .
 .It Sy BLOG_STRICT_ACCESS
 If set to
 .Ql 1
@@ -254,6 +258,15 @@ Is expected to be the URL part after
 i. e. the path to a subpage of the script.
 This is interpreted as a path to an entry and must start with a leading
 .Ql / .
+.It Ev SERVER_NAME
+Expected to be the hostname or IP(v6) address of the server the request is
+directed to or
+.Nm
+is running on.
+.It Ev SERVER_PORT
+Expected to be the port of the server the request is directed to or
+.Nm
+is running on.
 .El
 .Pp
 Most likely these will be set correctly by your favorite web server without
diff --git a/main.c b/main.c
index f5460ba..b9b6914 100644
--- a/main.c
+++ b/main.c
@@ -267,9 +267,6 @@ void blog_rss(char script_name[]) {
     struct xml_context ctx;
     new_xml_context(&ctx);
 
-    ctx.closing_slash = 1;
-    ctx.warn = stderr; // dev
-
     xml_raw(&ctx, "<?xml version=\"1.0\" encoding=\"UTF-8\" ?>");
     xml_open_tag_attrs(&ctx, "rss", 2, "version", "2.0", "xmlns:atom", "http://www.w3.org/2005/Atom");
     xml_open_tag(&ctx, "channel");
@@ -284,9 +281,14 @@ void blog_rss(char script_name[]) {
     xml_close_cdata(&ctx);
     xml_close_tag(&ctx, "description");
 
-    xml_open_tag(&ctx, "link");
-    xml_escaped(&ctx, BLOG_SERVER_URL);
-    xml_close_tag(&ctx, "link");
+    char *external_url = server_url(BLOG_USE_HTTPS);
+
+    if(external_url != NULL) {
+        xml_open_tag(&ctx, "link");
+        xml_escaped(&ctx, external_url);
+        xml_escaped(&ctx, script_name);
+        xml_close_tag(&ctx, "link");
+    }
 
     if(count > 0) {
         time_t update_time = entries[0].time;
@@ -299,13 +301,15 @@ void blog_rss(char script_name[]) {
         }
     }
 
-    char *rss_link = catn_alloc(3, BLOG_SERVER_URL, script_name, "/rss.xml");
-    if(rss_link != NULL) {
-        xml_empty_tag(&ctx, "atom:link", 3,
-                      "rel", "self",
-                      "href", rss_link,
-                      "type", "application/rss+xml");
-        free(rss_link);
+    if(external_url != NULL) {
+        char *rss_link = catn_alloc(3, external_url, script_name, "/rss.xml");
+        if(rss_link != NULL) {
+            xml_empty_tag(&ctx, "atom:link", 3,
+                          "rel", "self",
+                          "href", rss_link,
+                          "type", "application/rss+xml");
+            free(rss_link);
+        }
     }
 
     for(int i = 0; i < count; i++) {
@@ -315,15 +319,17 @@ void blog_rss(char script_name[]) {
             xml_escaped(&ctx, entries[i].title);
             xml_close_tag(&ctx, "title");
 
-            xml_open_tag(&ctx, "link");
-            xml_escaped(&ctx, BLOG_SERVER_URL);
-            xml_escaped(&ctx, entries[i].link);
-            xml_close_tag(&ctx, "link");
+            if(external_url != NULL) {
+                xml_open_tag(&ctx, "link");
+                xml_escaped(&ctx, external_url);
+                xml_escaped(&ctx, entries[i].link);
+                xml_close_tag(&ctx, "link");
 
-            xml_open_tag(&ctx, "guid");
-            xml_escaped(&ctx, BLOG_SERVER_URL);
-            xml_escaped(&ctx, entries[i].link);
-            xml_close_tag(&ctx, "guid");
+                xml_open_tag(&ctx, "guid");
+                xml_escaped(&ctx, external_url);
+                xml_escaped(&ctx, entries[i].link);
+                xml_close_tag(&ctx, "guid");
+            }
 
             if(entries[i].text_size > 0) {
                 xml_open_tag(&ctx, "description");
@@ -349,6 +355,8 @@ void blog_rss(char script_name[]) {
 
     xml_close_all(&ctx);
 
+    free(external_url);
+
     free_index(&entries, count);
 
     del_xml_context(&ctx);
@@ -371,8 +379,9 @@ void blog_atom(char script_name[]) {
     struct xml_context ctx;
     new_xml_context(&ctx);
 
-    char *self_url = catn_alloc(3, BLOG_SERVER_URL, script_name, "/atom.xml");
-    char *html_url = catn_alloc(2, BLOG_SERVER_URL, script_name);
+    char *external_url = server_url(BLOG_USE_HTTPS);
+    char *self_url = catn_alloc(3, external_url, script_name, "/atom.xml");
+    char *html_url = catn_alloc(2, external_url, script_name);
 
     send_standard_headers(200, "application/atom+xml");
 
@@ -432,10 +441,12 @@ void blog_atom(char script_name[]) {
         if(entry_get_text(&entries[i]) != -1) {
             xml_open_tag(&ctx, "entry");
 
-            xml_open_tag(&ctx, "id");
-            xml_escaped(&ctx, BLOG_SERVER_URL);
-            xml_escaped(&ctx, entries[i].link);
-            xml_close_tag(&ctx, "id");
+            if(external_url != NULL) {
+                xml_open_tag(&ctx, "id");
+                xml_escaped(&ctx, external_url);
+                xml_escaped(&ctx, entries[i].link);
+                xml_close_tag(&ctx, "id");
+            }
 
             xml_open_tag(&ctx, "title");
             xml_escaped(&ctx, entries[i].title);
@@ -448,7 +459,7 @@ void blog_atom(char script_name[]) {
                 xml_close_tag(&ctx, "updated");
             }
 
-            char *entry_url = catn_alloc(2, BLOG_SERVER_URL, entries[i].link);
+            char *entry_url = catn_alloc(2, external_url, entries[i].link);
             if(entry_url != NULL) {
                 xml_empty_tag(&ctx, "link", 3, "rel", "alternate", "type", "text/html", "href", entry_url);
                 free(entry_url);
@@ -468,6 +479,8 @@ void blog_atom(char script_name[]) {
 
     xml_close_tag(&ctx, "feed");
 
+    free(external_url);
+
     free_index(&entries, count);
     del_xml_context(&ctx);
 }