docs: fix/remove invalid URLs #20647

2 files changed, 14 insertions, 11 deletions

diff --git a/scripts/gen_help_html.lua b/scripts/gen_help_html.lua
index 06ea1831b0..39c516ee96 100644
--- a/scripts/gen_help_html.lua
+++ b/scripts/gen_help_html.lua
@@ -319,6 +319,16 @@ local function validate_link(node, bufnr, fname)
   return helppage, tagname, ignored
 end
 
+local function validate_url(text, fname)
+  local ignored = false
+  if vim.fs.basename(fname) == 'pi_netrw.txt' then
+    ignored = true
+  elseif text:find('http%:') then
+    invalid_urls[text] = vim.fs.basename(fname)
+  end
+  return ignored
+end
+
 -- Traverses the tree at `root` and checks that |tag| links point to valid helptags.
 local function visit_validate(root, level, lang_tree, opt, stats)
   level = level or 0
@@ -353,9 +363,7 @@ local function visit_validate(root, level, lang_tree, opt, stats)
       end
     end
   elseif node_name == 'url' then
-    if text:find('http%:') then
-      invalid_urls[text] = vim.fs.basename(opt.fname)
-    end
+    validate_url(text, opt.fname)
   elseif node_name == 'taglink' or node_name == 'optionlink' then
     local _, _, _ = validate_link(root, opt.buf, opt.fname)
   end
diff --git a/scripts/gen_vimdoc.py b/scripts/gen_vimdoc.py
index 3ee9d8b5dd..a044c8c39d 100755
--- a/scripts/gen_vimdoc.py
+++ b/scripts/gen_vimdoc.py
@@ -12,15 +12,10 @@ Flow:
             update_params_map /
               render_node
 
-This would be easier using lxml and XSLT, but:
+TODO: eliminate this script and use Lua+treesitter (requires parsers for C and
+Lua markdown-style docstrings).
 
-  1. This should avoid needing Python dependencies, especially ones that are
-     C modules that have library dependencies (lxml requires libxml and
-     libxslt).
-  2. I wouldn't know how to deal with nested indentation in <para> tags using
-     XSLT.
-
-Each function :help block is formatted as follows:
+The generated :help text for each function is formatted as follows:
 
   - Max width of 78 columns (`text_width`).
   - Indent with spaces (not tabs).