ocda: get image from is internal image markup

correct circumstance where regex was incorrectly able to match image
shaped words as well as identified and marked up image
(internal markup).

take all matches.

(assisted by Claude-Code)

3 files changed, 18 insertions, 6 deletions

diff --git a/org/default_regex.org b/org/default_regex.org
index 430828b..d1fc76f 100644
--- a/org/default_regex.org
+++ b/org/default_regex.org
@@ -343,7 +343,9 @@ static smid_inline_link_endnote_url_helper            = ctRegex!(`\{~\^\s+(?P<co
 
 #+NAME: meta_rgx_images
 #+BEGIN_SRC d
-static image                                    = ctRegex!(`([a-zA-Z0-9._-]+?\.(?:png|gif|jpg))`, "mg");
+/+  image post internal markup ┥☼name.png,w640h447 ┝┤├ +/
+static image_in_object_text                     = ctRegex!(`☼(?P<image>[a-zA-Z0-9._-]+?\.(?:png|gif|jpg)),w\d+h\d+`, "mg");
+/+ smid_image for sisu markup source (.sst) images `{ name.png }image` +/
 static smid_image                               = ctRegex!(`(?P<pre>(?:^|[ ])[{┥](?:~\^\s+|\s*))(?P<image>[a-zA-Z0-9._-]+?\.(?:png|gif|jpg))(?P<post>(?:.*?)\s*[}┝](?:image|┤.*?├|(?:(?:https?|git):\/\/|¤?\.\.\/|¤?\.\/|¤|#)\S+?)(?=[;:!,?.]?([ )\]]|$)))`, "mg");
 static smid_image_generic                       = ctRegex!(`(?:^|[ ])[{┥](?:~\^\s+|\s*)\S+\.(?:png|gif|jpg).*?[}┝](?:image|┤.*?├|(?:(?:https?|git):\/\/|¤?\.\.\/|¤?\.\/|¤|#)\S+?)(?=[;:!,?.]?([ )\]]|$))`, "mg");
 static smid_image_with_dimensions               = ctRegex!(`(?P<pre>(?:^|[ ])[{┥](?:~\^\s+|\s*))(?P<image>[a-zA-Z0-9._-]+?\.(?:png|gif|jpg))\s+(?P<width>\d+)x(?P<height>\d+)\s*(?P<post>(?:.*?)\s*[}┝](?:image|┤.*?├|(?:(?:https?|git):\/\/|¤?\.\.\/|¤?\.\/|¤|#)\S+?)(?=[;:!,?.]?([ )\]]|$)))`, "mg");
diff --git a/org/in_source_files.org b/org/in_source_files.org
index 6c6fd2c..e8b5272 100644
--- a/org/in_source_files.org
+++ b/org/in_source_files.org
@@ -301,11 +301,16 @@ template spineRawMarkupContent() {
   mixin spineRgxFiles;
   static auto rgx_files = RgxFiles();
   string[] _images=[];
+  /+ ↓ the images a markup source shows.
+       matched on the image markup itself ({ name.png }image, and the url and
+       ~^ forms), ensure a filename-shaped word in prose is not mistaken for
+       one, and over every match (so a line showing two images reports two).
+  +/
   string[] _extract_images(S)(S content_block) {
     string[] images_;
     string _content_block = content_block.to!string;
-    if (auto m = _content_block.matchAll(rgx.image)) {
-      images_ ~= m.captures[1].to!string;
+    foreach (m; _content_block.matchAll(rgx.smid_image)) {
+      images_ ~= m["image"].to!string;
     }
     return images_;
   }
diff --git a/org/ocda_functions.org b/org/ocda_functions.org
index 6ab4bbf..ca8a7be 100644
--- a/org/ocda_functions.org
+++ b/org/ocda_functions.org
@@ -1441,12 +1441,17 @@ ST_flow_table_array_munge flow_table_array_munge()(
 #+NAME: ocdaFunc_images
 #+HEADER: :noweb yes
 #+BEGIN_SRC d
-// ↓ - images
+/+ ↓ - images, the images an object shows.
+     matched on the internal markup an image has become by this stage
+     (┥☼name.png,w640h447 ┝┤├), so a filename-shaped word in prose is not
+     taken for one, and over every match rather than the first, so a
+     paragraph showing two images reports two.
++/
 string[] extract_images()(string content_block) {
   static auto rgx = RgxI();
   string[] images_;
-  if (auto m = content_block.matchAll(rgx.image)) {
-    images_ ~= m.captures[1];
+  foreach (m; content_block.matchAll(rgx.image_in_object_text)) {
+    images_ ~= m["image"].to!string;
   }
   return images_;
 }