htmlparser: get rid of some todos

* No namespace is correct.
* Queue a microtask can be implemented in elementPoppedImpl, because
  the stack of open elements is not visible to scripts.
* We can just change <image> tag name and reproces, no problem
* Update todo file

2 files changed, 2 insertions, 12 deletions

diff --git a/chame/htmlparser.nim b/chame/htmlparser.nim
index 7ee70372..779cf1ed 100644
--- a/chame/htmlparser.nim
+++ b/chame/htmlparser.nim
@@ -521,7 +521,6 @@ func createElement[Handle, Atom](parser: HTML5Parser[Handle, Atom],
 func toParsedAttrs[Atom](attrs: seq[TokenAttr[Atom]]): seq[ParsedAttr[Atom]] =
   result = @[]
   for attr in attrs:
-    #TODO is no namespace correct?
     result.add((NO_PREFIX, NO_NAMESPACE, attr.name, attr.value))
 
 func createElement[Handle, Atom](parser: HTML5Parser[Handle, Atom],
@@ -1906,13 +1905,7 @@ proc processInHTMLContent[Handle, Atom](parser: var HTML5Parser[Handle, Atom],
         parser.framesetOk = false
       )
       "<image>" => (block:
-        #TODO ew
-        let token = Token(
-          t: START_TAG,
-          tagname: parser.tagTypeToAtom(TAG_IMG),
-          selfclosing: token.selfclosing,
-          attrs: token.attrs
-        )
+        token.tagname = parser.tagTypeToAtom(TAG_IMG)
         reprocess token
       )
       "<textarea>" => (block:
@@ -2023,7 +2016,6 @@ proc processInHTMLContent[Handle, Atom](parser: var HTML5Parser[Handle, Atom],
         reprocess token
       )
       "</script>" => (block:
-        #TODO microtask (?)
         pop_current_node
         parser.insertionMode = parser.oldInsertionMode
       )
diff --git a/todo b/todo
index 0772e4af..4a113baa 100644
--- a/todo
+++ b/todo
@@ -2,13 +2,11 @@ interface:
 * pass TokenAttr to createElement, then append adjusted attributes with a
   separate callback
 * get rid of getParentNode (replace with an insertBeforeOrAppendTo?)
-* optional callbacks without function pointers (when compiles?)
+* use converter + proc for optional callbacks instead of method?
 optimizations:
 * reduce use of getNamespace (maybe store ns on the stack of open elements?)
 * reduce atomToTagType calls (store tagType before the case statement in match?)
-* go through TODOs in htmltokenizer
 * on-demand creation of attribute name atoms (most documents don't even
   need atomMap); use ref array/Option?
-* htmltokenizer: do something with copyBuf, it's ugly and probably broken
 etc:
 * turn test1 into a serializer module