content-parser: fixed DESCRIPTION variable.

it is now built as a single line variable, that contains the full
unparsed content of the first paragraph found in the source file,
instead of just the first line of it.

this also fixes a bug that prevented creating of DESCRIPTION variable,
if the source file contained only a single line paragraph.

4 files changed, 98 insertions, 10 deletions

diff --git a/man/blogc-source.7.ronn b/man/blogc-source.7.ronn
index 1945329..9b80eb2 100644
--- a/man/blogc-source.7.ronn
+++ b/man/blogc-source.7.ronn
@@ -51,9 +51,10 @@ for building permalinks in templates. This variable can't be overriden by an
 explicit definition in source file.
 
 Another variable, `DESCRIPTION`, will be automatically created by the source
-parser. Its content is the first line of the first paragraph of the content.
-Paragraphs inside blockquotes are ignored. This variable can be overriden by
-an explicit definition in source file.
+parser. It contains the unparsed content of the first paragraph found in the
+source file. Paragraphs inside blockquotes are ignored. This variable can be
+overriden by an explicit definition in source file. This is useful to add
+"description" meta-tags to HTML templates.
 
 ## SOURCE CONTENT - BLOCK ELEMENTS
 
diff --git a/src/content-parser.c b/src/content-parser.c
index 4840a1d..3cb0ddb 100644
--- a/src/content-parser.c
+++ b/src/content-parser.c
@@ -75,6 +75,47 @@ blogc_htmlentities(const char *str)
 }
 
 
+char*
+blogc_fix_description(const char *paragraph)
+{
+    if (paragraph == NULL)
+        return NULL;
+    sb_string_t *rv = sb_string_new();
+    bool last = false;
+    bool newline = false;
+    char *tmp = NULL;
+    size_t start = 0;
+    size_t current = 0;
+    while (true) {
+        switch (paragraph[current]) {
+            case '\0':
+                last = true;
+            case '\r':
+            case '\n':
+                if (newline)
+                    break;
+                tmp = sb_strndup(paragraph + start, current - start);
+                sb_string_append(rv, sb_str_strip(tmp));
+                free(tmp);
+                tmp = NULL;
+                if (!last)
+                    sb_string_append_c(rv, ' ');
+                start = current + 1;
+                newline = true;
+                break;
+            default:
+                newline = false;
+        }
+        if (last)
+            break;
+        current++;
+    }
+    tmp = sb_strdup(sb_str_strip(rv->str));
+    sb_string_free(rv, true);
+    return tmp;
+}
+
+
 typedef enum {
     CONTENT_START_LINE = 1,
     CONTENT_EXCERPT,
@@ -1018,8 +1059,6 @@ blogc_content_parse(const char *src, size_t *end_excerpt, char **description)
                     state = CONTENT_PARAGRAPH_END;
                     end = is_last && c != '\n' && c != '\r' ? src_len :
                         (real_end != 0 ? real_end : current);
-                    if (description != NULL && *description == NULL)
-                        *description = sb_strndup(src + start, end - start);
                 }
                 if (!is_last)
                     break;
@@ -1027,6 +1066,8 @@ blogc_content_parse(const char *src, size_t *end_excerpt, char **description)
             case CONTENT_PARAGRAPH_END:
                 if (c == '\n' || c == '\r' || is_last) {
                     tmp = sb_strndup(src + start, end - start);
+                    if (description != NULL && *description == NULL)
+                        *description = blogc_fix_description(tmp);
                     parsed = blogc_content_parse_inline(tmp);
                     sb_string_append_printf(rv, "<p>%s</p>%s", parsed,
                         line_ending);
diff --git a/src/content-parser.h b/src/content-parser.h
index 148d5ed..85bd063 100644
--- a/src/content-parser.h
+++ b/src/content-parser.h
@@ -14,6 +14,7 @@
 
 char* blogc_slugify(const char *str);
 char* blogc_htmlentities(const char *str);
+char* blogc_fix_description(const char *paragraph);
 char* blogc_content_parse_inline(const char *src);
 bool blogc_is_ordered_list_item(const char *str, size_t prefix_len);
 char* blogc_content_parse(const char *src, size_t *end_excerpt,
diff --git a/tests/check_content_parser.c b/tests/check_content_parser.c
index d3ef2b8..ab6853a 100644
--- a/tests/check_content_parser.c
+++ b/tests/check_content_parser.c
@@ -60,6 +60,50 @@ test_htmlentities(void **state)
 
 
 static void
+test_fix_description(void **state)
+{
+    assert_null(blogc_fix_description(NULL));
+    char *s = blogc_fix_description("bola");
+    assert_string_equal(s, "bola");
+    free(s);
+    s = blogc_fix_description("bola\n");
+    assert_string_equal(s, "bola");
+    free(s);
+    s = blogc_fix_description("bola\r\n");
+    assert_string_equal(s, "bola");
+    free(s);
+    s = blogc_fix_description("bola\nguda");
+    assert_string_equal(s, "bola guda");
+    free(s);
+    s = blogc_fix_description("bola\nguda\n");
+    assert_string_equal(s, "bola guda");
+    free(s);
+    s = blogc_fix_description("bola\r\nguda\r\n");
+    assert_string_equal(s, "bola guda");
+    free(s);
+
+    s = blogc_fix_description("bola\n   guda   lol\n asd");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+    s = blogc_fix_description("bola\n   guda   lol\n asd\n");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+    s = blogc_fix_description("bola\r\n   guda   lol\r\n asd\r\n");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+    s = blogc_fix_description("  bola\n   guda   lol\n asd");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+    s = blogc_fix_description("  bola\n   guda   lol\n asd\n");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+    s = blogc_fix_description("  bola\r\n   guda   lol\r\n asd\r\n");
+    assert_string_equal(s, "bola guda   lol asd");
+    free(s);
+}
+
+
+static void
 test_is_ordered_list_item(void **state)
 {
     assert_true(blogc_is_ordered_list_item("1.bola", 2));
@@ -127,7 +171,7 @@ test_content_parse(void **state)
     assert_non_null(html);
     assert_int_equal(l, 0);
     assert_non_null(d);
-    assert_string_equal(d, "bola");
+    assert_string_equal(d, "bola chunda");
     assert_string_equal(html,
         "<h1 id=\"um\">um</h1>\n"
         "<h2 id=\"dois\">dois</h2>\n"
@@ -219,7 +263,7 @@ test_content_parse_crlf(void **state)
     assert_non_null(html);
     assert_int_equal(l, 0);
     assert_non_null(d);
-    assert_string_equal(d, "bola");
+    assert_string_equal(d, "bola chunda");
     assert_string_equal(html,
         "<h1 id=\"um\">um</h1>\r\n"
         "<h2 id=\"dois\">dois</h2>\r\n"
@@ -1136,7 +1180,7 @@ test_content_parse_description(void **state)
         "<p>qwe\n"
         "bar</p>\n");
     assert_non_null(d);
-    assert_string_equal(d, "qwe");
+    assert_string_equal(d, "qwe bar");
     free(html);
     free(d);
     d = NULL;
@@ -1219,7 +1263,7 @@ test_content_parse_description_crlf(void **state)
         "<p>qwe\r\n"
         "bar</p>\r\n");
     assert_non_null(d);
-    assert_string_equal(d, "qwe");
+    assert_string_equal(d, "qwe bar");
     free(html);
     free(d);
     d = NULL;
@@ -1277,7 +1321,7 @@ test_content_parse_invalid_excerpt(void **state)
     assert_non_null(html);
     assert_int_equal(l, 0);
     assert_non_null(d);
-    assert_string_equal(d, "chunda");
+    assert_string_equal(d, "chunda ..");
     assert_string_equal(html,
         "<h1 id=\"test\">test</h1>\n"
         "<p>chunda\n"
@@ -2019,6 +2063,7 @@ main(void)
     const UnitTest tests[] = {
         unit_test(test_slugify),
         unit_test(test_htmlentities),
+        unit_test(test_fix_description),
         unit_test(test_is_ordered_list_item),
         unit_test(test_content_parse),
         unit_test(test_content_parse_crlf),