Support for table column spans

CHANGELOG.md

@@ -272,6 +272,8 @@ Fixes:
    Fix reference link definition label matching in a case when the label ends
    with a Unicode character with non-trivial case folding mapping.
 
+ * [#29](https://github.com/mity/md4c/issues/29):
+   Implement Wiky-style ('|2> text |') table colspan extension.
 
 ## Version 0.4.6
 

CMakeLists.txt

@@ -68,4 +68,23 @@ include(GNUInstallDirs)
 add_subdirectory(src)
 if (BUILD_MD2HTML_EXECUTABLE)
     add_subdirectory(md2html)
-endif ()
+    set(MINIMUM_PYTHON_VERSION "3.6")
+    set(TEST_DIR "${CMAKE_CURRENT_SOURCE_DIR}/test")
+    find_package(Python3 COMPONENTS Interpreter)
+    if(Python3_FOUND AND NOT Python3_VERSION VERSION_LESS ${MINIMUM_PYTHON_VERSION})
+        enable_testing()
+        file(GLOB TEST_FILES "${CMAKE_CURRENT_SOURCE_DIR}/test/*.txt")
+        # get_target_property(PROGRAM md2html LOCATION)
+        set(PROGRAM "${CMAKE_CURRENT_BINARY_DIR}/md2html/md2html")
+        foreach(TEST_FILE ${TEST_FILES})
+            get_filename_component(TEST_NAME ${TEST_FILE} NAME_WE)
+            add_test(NAME ${TEST_NAME}
+                     COMMAND ${Python3_EXECUTABLE} "${TEST_DIR}/run-testsuite.py" -s ${TEST_FILE} -p ${PROGRAM})
+        endforeach()
+            add_test(NAME Pathological_inputs
+                     COMMAND ${Python3_EXECUTABLE} "${TEST_DIR}/pathological-tests.py" -p ${PROGRAM})
+     else()
+        message(WARNING "Python not found or version is less than ${MINIMUM_PYTHON_VERSION}.")
+    endif()
+endif()
+

README.md

@@ -97,6 +97,9 @@ extensions:
 
 * With the flag `MD_FLAG_TABLES`, GitHub-style tables are supported.
 
+* With the flag `MD_FLAG_TABLES_COLSPAN`, wiky-style ('|2> text |') table
+  colspan are supported.
+
 * With the flag `MD_FLAG_TASKLISTS`, GitHub-style task lists are supported.
 
 * With the flag `MD_FLAG_STRIKETHROUGH`, strike-through spans are enabled

md2html/md2html.c

@@ -255,6 +255,7 @@ static const CMDLINE_OPTION cmdline_options[] = {
     {  0,  "fpermissive-www-autolinks",     '.', 0 },
     {  0,  "fstrikethrough",                'S', 0 },
     {  0,  "ftables",                       'T', 0 },
+    {  0,  "ftable-colspan",                'J', 0 },
     {  0,  "ftasklists",                    'X', 0 },
     {  0,  "funderline",                    '_', 0 },
     {  0,  "fverbatim-entities",            'E', 0 },
@@ -310,6 +311,7 @@ usage(void)
         "                       Force all soft breaks to act as hard breaks\n"
         "      --fstrikethrough Enable strike-through spans\n"
         "      --ftables        Enable tables\n"
+        "      --ftable-colspan Enable table colspan extension ('|2> text |')\n"
         "      --ftasklists     Enable task lists\n"
         "      --funderline     Enable underline spans\n"
         "      --fwiki-links    Enable wiki links\n"
@@ -382,6 +384,7 @@ cmdline_callback(int opt, char const* value, void* data)
         case '@':   parser_flags |= MD_FLAG_PERMISSIVEEMAILAUTOLINKS; break;
         case 'V':   parser_flags |= MD_FLAG_PERMISSIVEAUTOLINKS; break;
         case 'T':   parser_flags |= MD_FLAG_TABLES; break;
+        case 'J':   parser_flags |= MD_FLAG_TABLES_COLSPAN; break;
         case 'S':   parser_flags |= MD_FLAG_STRIKETHROUGH; break;
         case 'L':   parser_flags |= MD_FLAG_LATEXMATHSPANS; break;
         case 'K':   parser_flags |= MD_FLAG_WIKILINKS; break;

src/md4c-html.c

@@ -312,15 +312,24 @@ render_open_code_block(MD_HTML* r, const MD_BLOCK_CODE_DETAIL* det)
 static void
 render_open_td_block(MD_HTML* r, const MD_CHAR* cell_type, const MD_BLOCK_TD_DETAIL* det)
 {
+    MD_CHAR tail[2];
+
     RENDER_VERBATIM(r, "<");
     RENDER_VERBATIM(r, cell_type);
 
     switch(det->align) {
-        case MD_ALIGN_LEFT:     RENDER_VERBATIM(r, " align=\"left\">"); break;
-        case MD_ALIGN_CENTER:   RENDER_VERBATIM(r, " align=\"center\">"); break;
-        case MD_ALIGN_RIGHT:    RENDER_VERBATIM(r, " align=\"right\">"); break;
-        default:                RENDER_VERBATIM(r, ">"); break;
+        case MD_ALIGN_LEFT:     RENDER_VERBATIM(r, " align=\"left\""); break;
+        case MD_ALIGN_CENTER:   RENDER_VERBATIM(r, " align=\"center\""); break;
+        case MD_ALIGN_RIGHT:    RENDER_VERBATIM(r, " align=\"right\""); break;
+        case MD_ALIGN_DEFAULT:  break;
+    }
+    if (det->colspan > 1) {
+        RENDER_VERBATIM(r, " colspan=\"");
+        tail[0] = '0' + det->colspan;
+        tail[1] = '\"';
+        render_verbatim(r, tail, 2);
     }
+    RENDER_VERBATIM(r, ">");
 }
 
 static void

src/md4c.c

@@ -4529,24 +4529,22 @@ md_analyze_table_alignment(MD_CTX* ctx, OFF beg, OFF end, MD_ALIGN* align, int n
 static int md_process_normal_block_contents(MD_CTX* ctx, const MD_LINE* lines, MD_SIZE n_lines);
 
 static int
-md_process_table_cell(MD_CTX* ctx, MD_BLOCKTYPE cell_type, MD_ALIGN align, OFF beg, OFF end)
+md_process_table_cell(MD_CTX* ctx, MD_BLOCKTYPE cell_type, MD_BLOCK_TD_DETAIL *det, OFF beg, OFF end)
 {
     MD_LINE line;
-    MD_BLOCK_TD_DETAIL det;
     int ret = 0;
 
     while(beg < end  &&  ISWHITESPACE(beg))
         beg++;
     while(end > beg  &&  ISWHITESPACE(end-1))
         end--;
 
-    det.align = align;
     line.beg = beg;
     line.end = end;
 
-    MD_ENTER_BLOCK(cell_type, &det);
+    MD_ENTER_BLOCK(cell_type, det);
     MD_CHECK(md_process_normal_block_contents(ctx, &line, 1));
-    MD_LEAVE_BLOCK(cell_type, &det);
+    MD_LEAVE_BLOCK(cell_type, det);
 
 abort:
     return ret;
@@ -4557,6 +4555,10 @@ md_process_table_row(MD_CTX* ctx, MD_BLOCKTYPE cell_type, OFF beg, OFF end,
                      const MD_ALIGN* align, int col_count)
 {
     MD_LINE line;
+    MD_LINE cell;
+    MD_BLOCK_TD_DETAIL det;
+
+    unsigned is_spans;
     OFF* pipe_offs = NULL;
     int i, j, k, n;
     int ret = 0;
@@ -4583,19 +4585,36 @@ md_process_table_row(MD_CTX* ctx, MD_BLOCKTYPE cell_type, OFF beg, OFF end,
         MD_MARK* mark = &ctx->marks[i];
         pipe_offs[j++] = mark->end;
     }
-    pipe_offs[j++] = end+1;
+    pipe_offs[j] = end+1;
 
+    is_spans = ctx->parser.flags & MD_FLAG_TABLES_COLSPAN;
     /* Process cells. */
     MD_ENTER_BLOCK(MD_BLOCK_TR, NULL);
     k = 0;
-    for(i = 0; i < j-1  &&  k < col_count; i++) {
-        if(pipe_offs[i] < pipe_offs[i+1]-1)
-            MD_CHECK(md_process_table_cell(ctx, cell_type, align[k++], pipe_offs[i], pipe_offs[i+1]-1));
+    for(i = 0; i < j  &&  k < col_count; i++) {
+        cell.beg = pipe_offs[i];
+        cell.end = pipe_offs[i+1]-1;
+
+        if(cell.beg < cell.end) {
+            det.colspan = 0;
+            if (is_spans && cell.end - cell.beg >= 2 &&
+                ISDIGIT(cell.beg) && CH(cell.beg+1) == '>') {
+                det.colspan = CH(cell.beg) - '0';
+                cell.beg += 2;
+            }
+            det.align = align[k++];
+            MD_CHECK(md_process_table_cell(ctx, cell_type, &det, cell.beg, cell.end));
+            if (is_spans && det.colspan > 1)
+                k += det.colspan - 1;
+        }
     }
     /* Make sure we call enough table cells even if the current table contains
      * too few of them. */
-    while(k < col_count)
-        MD_CHECK(md_process_table_cell(ctx, cell_type, align[k++], 0, 0));
+    det.colspan = 0;
+    while(k < col_count) {
+        det.align = align[k++];
+        MD_CHECK(md_process_table_cell(ctx, cell_type, &det, 0, 0));
+    }
     MD_LEAVE_BLOCK(MD_BLOCK_TR, NULL);
 
 abort:

src/md4c.h

@@ -278,6 +278,7 @@ typedef struct MD_BLOCK_TABLE_DETAIL {
 /* Detailed info for MD_BLOCK_TH and MD_BLOCK_TD. */
 typedef struct MD_BLOCK_TD_DETAIL {
     MD_ALIGN align;
+    int colspan;
 } MD_BLOCK_TD_DETAIL;
 
 /* Detailed info for MD_SPAN_A. */
@@ -318,6 +319,7 @@ typedef struct MD_SPAN_WIKILINK {
 #define MD_FLAG_WIKILINKS                   0x2000  /* Enable wiki links extension. */
 #define MD_FLAG_UNDERLINE                   0x4000  /* Enable underline extension (and disables '_' for normal emphasis). */
 #define MD_FLAG_HARD_SOFT_BREAKS            0x8000  /* Force all soft breaks to act as hard breaks. */
+#define MD_FLAG_TABLES_COLSPAN              0x10000  /* Enable table colspan extension '|2> 2-colspan cell |'. */
 
 #define MD_FLAG_PERMISSIVEAUTOLINKS         (MD_FLAG_PERMISSIVEEMAILAUTOLINKS | MD_FLAG_PERMISSIVEURLAUTOLINKS | MD_FLAG_PERMISSIVEWWWAUTOLINKS)
 #define MD_FLAG_NOHTML                      (MD_FLAG_NOHTMLBLOCKS | MD_FLAG_NOHTMLSPANS)

test/spec-tables-colspan.txt

@@ -0,0 +1,74 @@
+
+# Table colspan extension
+
+With the flag `MD_FLAG_TABLES_COLSPAN`, MD4C enables extension for recognition
+of table colspan's.
+
+Basic table example of a table with colspan:
+
+```````````````````````````````` example
+| Column 1 | Column 2 |
+|----------|----------|
+|   foo    | bar      |
+|2> baz               |
+|   quux   | quuz     |
+.
+<table>
+<thead>
+<tr><th>Column 1</th><th>Column 2</th></tr>
+</thead>
+<tbody>
+<tr><td>foo</td><td>bar</td></tr>
+<tr><td colspan="2">baz</td></tr>
+<tr><td>quux</td><td>quuz</td></tr>
+</tbody>
+</table>
+.
+--ftables --ftable-colspan
+````````````````````````````````
+
+Only one digit are allowed:
+
+```````````````````````````````` example
+| Column 1 | Column 2 |
+|----------|----------|
+|   foo    | bar      |
+|22> baz              |
+|   quux   | quuz     |
+.
+<table>
+<thead>
+<tr><th>Column 1</th><th>Column 2</th></tr>
+</thead>
+<tbody>
+<tr><td>foo</td><td>bar</td></tr>
+<tr><td>22&gt; baz</td><td></td></tr>
+<tr><td>quux</td><td>quuz</td></tr>
+</tbody>
+</table>
+.
+--ftables --ftable-colspan
+````````````````````````````````
+
+The arrow character must follow without spaces:
+
+```````````````````````````````` example
+| Column 1 | Column 2 |
+|----------|----------|
+|   foo    | bar      |
+| 2 > baz             |
+|   quux   | quuz     |
+.
+<table>
+<thead>
+<tr><th>Column 1</th><th>Column 2</th></tr>
+</thead>
+<tbody>
+<tr><td>foo</td><td>bar</td></tr>
+<tr><td>2 &gt; baz</td><td></td></tr>
+<tr><td>quux</td><td>quuz</td></tr>
+</tbody>
+</table>
+.
+--ftables --ftable-colspan
+````````````````````````````````