Walk the tree rather than use selectors in canonicalizeShortcuts().

bikeshed/Spec.py

@@ -277,7 +277,7 @@ def processDocument(self) -> Spec:
         boilerplate.removeUnwantedBoilerplate(self)
         shorthands.run(self)
         inlineTags.processTags(self)
-        u.canonicalizeShortcuts(self)
+        u.canonicalizeShortcuts(self.body)
         u.addImplicitAlgorithms(self)
         u.fixManualDefTables(self)
         headings.processHeadings(self)

bikeshed/conditional.py

@@ -28,7 +28,7 @@
 
 def processConditionals(doc: t.SpecT, container: t.ElementT | None = None) -> None:
     for el in h.findAll("[include-if], [exclude-if], if-wrapper", container if container is not None else doc):
-        if el.tag == "if-wrapper" and not h.hasAttr(el, "include-if", "exclude-if"):
+        if el.tag == "if-wrapper" and not h.hasAnyAttr(el, "include-if", "exclude-if"):
             m.die(
                 "<if-wrapper> elements must have an include-if and/or exclude-if attribute.",
                 el=el,

bikeshed/h/__init__.py

@@ -29,6 +29,7 @@
     fixupIDs,
     foldWhitespace,
     hasAncestor,
+    hasAnyAttr,
     hasAttr,
     hasAttrs,
     hasChildElements,
@@ -60,6 +61,7 @@
     removeAttr,
     removeClass,
     removeNode,
+    renameAttr,
     replaceContents,
     replaceMacrosTextly,
     replaceNode,

bikeshed/h/dom.py

@@ -696,11 +696,25 @@ def removeAttr(el: t.ElementT, *attrNames: str) -> t.ElementT:
     return el
 
 
-def hasAttr(el: t.ElementT, *attrNames: str) -> bool:
+def renameAttr(el: t.ElementT, oldAttr: str, newAttr: str) -> t.ElementT:
+    # If oldAttr exists on the element, moves its value to newAttr
+    # and deletes oldAttr. Silently does nothing if oldAttr doesn't exist.
+    val = el.get(oldAttr)
+    if val is not None:
+        el.set(newAttr, val)
+        removeAttr(el, oldAttr)
+    return el
+
+
+def hasAnyAttr(el: t.ElementT, *attrNames: str) -> bool:
     # Returns True if the element has at least one of the named attributes
     return any(attrName in el.attrib for attrName in attrNames)
 
 
+def hasAttr(el: t.ElementT, attrName: str) -> bool:
+    return attrName in el.attrib
+
+
 def hasAttrs(el: t.ElementT) -> bool:
     return bool(el.attrib)
 

bikeshed/unsortedJunk.py

@@ -51,60 +51,68 @@ def fixManualDefTables(doc: t.SpecT) -> None:
         h.replaceContents(cell, newContents)
 
 
-def canonicalizeShortcuts(doc: t.SpecT) -> None:
+def canonicalizeShortcuts(el: t.ElementT) -> None:
     # Take all the invalid-HTML shortcuts and fix them.
-
-    attrFixup = {
-        "export": "data-export",
-        "noexport": "data-noexport",
-        "link-spec": "data-link-spec",
-        "spec": "data-link-spec",
-        "link-status": "data-link-status",
-        "status": "data-link-status",
-        "dfn-for": "data-dfn-for",
-        "link-for": "data-link-for",
-        "link-for-hint": "data-link-for-hint",
-        "dfn-type": "data-dfn-type",
-        "link-type": "data-link-type",
-        "dfn-force": "data-dfn-force",
-        "force": "data-dfn-force",
-        "section": "data-section",
-        "attribute-info": "data-attribute-info",
-        "dict-member-info": "data-dict-member-info",
-        "lt": "data-lt",
-        "local-lt": "data-local-lt",
-        "algorithm": "data-algorithm",
-        "ignore": "data-var-ignore",
-    }
-    for el in h.findAll(",".join(f"[{attr}]" for attr in attrFixup), doc):
-        for attr, fixedAttr in attrFixup.items():
-            if el.get(attr) is not None:
-                el.set(fixedAttr, t.cast(str, el.get(attr)))
-                del el.attrib[attr]
-
-    # The next two aren't in the above dict because some of the words conflict with existing attributes on some elements.
+    h.renameAttr(el, "export", "data-export")
+    h.renameAttr(el, "noexport", "data-noexport")
+    h.renameAttr(el, "link-spec", "data-link-spec")
+    h.renameAttr(el, "spec", "data-link-spec")
+    h.renameAttr(el, "link-status", "data-link-status")
+    h.renameAttr(el, "status", "data-link-status")
+    h.renameAttr(el, "dfn-for", "data-dfn-for")
+    h.renameAttr(el, "link-for", "data-link-for")
+    h.renameAttr(el, "link-for-hint", "data-link-for-hint")
+    h.renameAttr(el, "dfn-type", "data-dfn-type")
+    h.renameAttr(el, "link-type", "data-link-type")
+    h.renameAttr(el, "dfn-force", "data-dfn-force")
+    h.renameAttr(el, "force", "data-dfn-force")
+    h.renameAttr(el, "section", "data-section")
+    h.renameAttr(el, "attribute-info", "data-attribute-info")
+    h.renameAttr(el, "dict-member-info", "data-dict-member-info")
+    h.renameAttr(el, "lt", "data-lt")
+    h.renameAttr(el, "local-lt", "data-local-lt")
+    h.renameAttr(el, "algorithm", "data-algorithm")
+    h.renameAttr(el, "ignore", "data-var-ignore")
+
+    # Some of these attrs conflict with existing real attributes on some elements.
     # Instead, limit the search/transforms to the relevant elements.
-    for el in h.findAll("dfn, h2, h3, h4, h5, h6", doc):
+    tag = h.tagName(el)
+    if tag in ("dfn", "h2", "h3", "h4", "h5", "h6"):
         for dfnType in config.dfnTypes:
-            if el.get(dfnType) == "":
-                del el.attrib[dfnType]
-                el.set("data-dfn-type", dfnType)
-                break
-    for el in h.findAll("a", doc):
+            if h.hasAttr(el, dfnType):
+                if el.get(dfnType) == "":
+                    h.removeAttr(el, dfnType)
+                    el.set("data-dfn-type", dfnType)
+                    break
+                elif dfnType == "type" and el.get(dfnType) in config.dfnTypes:
+                    realDfnType = el.get(dfnType)
+                    m.die(
+                        f"You used type={realDfnType}, which isn't valid. Instead use just {realDfnType} as a boolean attribute, or dfn-type={realDfnType}.",
+                        el=el,
+                    )
+                    h.removeAttr(el, "type")
+                    break
+
+    if tag == "a":
         for linkType in config.linkTypes:
-            if el.get(linkType) is not None:
-                del el.attrib[linkType]
+            if h.hasAttr(el, linkType):
+                h.removeAttr(el, linkType)
                 el.set("data-link-type", linkType)
                 break
-    for el in h.findAll(config.dfnElementsSelector + ", a", doc):
-        if el.get("for") is None:
-            continue
-        _for = t.cast(str, el.get("for"))
-        del el.attrib["for"]
-        if el.tag == "a":
-            el.set("data-link-for", _for)
-        else:
-            el.set("data-dfn-for", _for)
+
+    if (
+        tag == "a"
+        or (tag == "dfn" and not h.hasAttr(el, "data-var-ignore"))
+        or (tag in ("h2", "h3", "h4", "h5", "h6") and h.hasAttr(el, "data-dfn-type"))
+    ):
+        if h.hasAttr(el, "for"):
+            if el.tag == "a":
+                h.renameAttr(el, "for", "data-link-for")
+            else:
+                h.renameAttr(el, "for", "data-dfn-for")
+
+    for child in h.childElements(el):
+        canonicalizeShortcuts(child)
 
 
 def addImplicitAlgorithms(doc: t.SpecT) -> None:

tests/github/WICG/sms-one-time-codes/index.console.txt

@@ -1,3 +1,6 @@
+LINE 166:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 186:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 196:5: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
 LINE 71:39: Multiple possible 'same site' dfn refs.
 Arbitrarily chose https://html.spec.whatwg.org/multipage/browsers.html#concept-site-same-site
 To auto-select one of the following refs, insert one of these lines into a <pre class=link-defaults> block:

tests/github/WICG/sms-one-time-codes/index.html

@@ -752,7 +752,7 @@ <h3 class="heading settled" data-level="3.1" id="authoring"><span class="secno">
     </div>
    </div>
    <h3 class="heading settled" data-level="3.2" id="parsing"><span class="secno">3.2. </span><span class="content">Parsing</span><a class="self-link" href="#parsing"></a></h3>
-   <p>To <dfn class="dfn-paneled" data-dfn-type="dfn" data-export id="parse-an-origin-bound-one-time-code-message" type="abstract-op">parse an origin-bound one-time code message</dfn> from <var>message</var>, run these steps:</p>
+   <p>To <dfn class="dfn-paneled" data-dfn-type="dfn" data-export id="parse-an-origin-bound-one-time-code-message">parse an origin-bound one-time code message</dfn> from <var>message</var>, run these steps:</p>
    <ol>
     <li data-md>
      <p>Let <var>line</var> be the <a data-link-type="dfn" href="#last-line" id="ref-for-last-line">last line</a> of <var>message</var>, and <var>position</var> be 0.</p>
@@ -791,7 +791,7 @@ <h3 class="heading settled" data-level="3.2" id="parsing"><span class="secno">3.
     <li data-md>
      <p>Return the <a data-link-type="dfn" href="#origin-bound-one-time-code" id="ref-for-origin-bound-one-time-code①⓪">origin-bound one-time code</a> (<var>top-level origin</var>, <var>embedded origin</var>, <var>code</var>).</p>
    </ol>
-   <p>To <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-marked-token" type="abstract-op">extract a marked token</dfn> from <var>string</var> at <var>position</var> with <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#code-point" id="ref-for-code-point②">code point</a> <var>marker</var>, run the following steps:</p>
+   <p>To <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-marked-token">extract a marked token</dfn> from <var>string</var> at <var>position</var> with <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#code-point" id="ref-for-code-point②">code point</a> <var>marker</var>, run the following steps:</p>
    <ol>
     <li data-md>
      <p>If <var>position</var> points past the end of <var>string</var>, return failure.</p>
@@ -808,7 +808,7 @@ <h3 class="heading settled" data-level="3.2" id="parsing"><span class="secno">3.
     <li data-md>
      <p>Return <var>token</var>.</p>
    </ol>
-   <p>The <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="last-line" type="abstract-op">last line</dfn> of <var>string</var> is the result of running these steps:</p>
+   <p>The <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="last-line">last line</dfn> of <var>string</var> is the result of running these steps:</p>
    <ol>
     <li data-md>
      <p><a data-link-type="dfn" href="https://infra.spec.whatwg.org/#normalize-newlines" id="ref-for-normalize-newlines">Normalize newlines</a> in <var>string</var>.</p>

tests/github/privacycg/private-click-measurement/private-click-measurement.console.txt

@@ -1 +1,3 @@
 LINE 147:1: Saw an end tag </div>, but there's no open element corresponding to it.
+LINE 191:9: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 210:9: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.

tests/github/privacycg/private-click-measurement/private-click-measurement.html

@@ -970,7 +970,7 @@ <h2 class="heading settled" data-level="8" id="n-bit-decimal-values"><span class
     <p>The <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string②">strings</a> <code>"00"</code> and <code>"15"</code> are both <a data-link-type="dfn" href="#four-bit-decimal-value" id="ref-for-four-bit-decimal-value①">four-bit decimal values</a>, whereas <code>"7"</code>, <code>"20"</code>, and <code>"!!11one"</code> are not.</p>
    </div>
    <div class="algorithm" data-algorithm="extract a four-bit decimal value">
-    <p>Clients <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-four-bit-decimal-value" type="abstract-op">extract a four-bit decimal value</dfn> from a <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string③">string</a> <var>string</var> by running these steps:</p>
+    <p>Clients <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-four-bit-decimal-value">extract a four-bit decimal value</dfn> from a <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string③">string</a> <var>string</var> by running these steps:</p>
     <ol>
      <li data-md>
       <p>If <var>string</var>’s <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string-length" id="ref-for-string-length">length</a> is not 2, return failure.</p>
@@ -991,7 +991,7 @@ <h2 class="heading settled" data-level="8" id="n-bit-decimal-values"><span class
     <p>The <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string④">strings</a> <code>"00"</code> and <code>"63"</code> are both <a data-link-type="dfn" href="#six-bit-decimal-value" id="ref-for-six-bit-decimal-value①">six-bit decimal values</a>, whereas <code>"7"</code>, <code>"98"</code>, and <code>"!!11one"</code> are not.</p>
    </div>
    <div class="algorithm" data-algorithm="extract a six-bit decimal value">
-    <p>Clients <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-six-bit-decimal-value" type="abstract-op">extract a six-bit decimal value</dfn> from a <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string⑤">string</a> <var>string</var> by running these steps:</p>
+    <p>Clients <dfn class="dfn-paneled" data-dfn-type="dfn" data-noexport id="extract-a-six-bit-decimal-value">extract a six-bit decimal value</dfn> from a <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string" id="ref-for-string⑤">string</a> <var>string</var> by running these steps:</p>
     <ol>
      <li data-md>
       <p>If <var>string</var>’s <a data-link-type="dfn" href="https://infra.spec.whatwg.org/#string-length" id="ref-for-string-length①">length</a> is not 2, return failure.</p>

tests/github/privacycg/storage-access/storage-access.console.txt

@@ -1,3 +1,9 @@
+LINE 121:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 133:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 148:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 155:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 225:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
+LINE 235:4: You used type=abstract-op, which isn't valid. Instead use just abstract-op as a boolean attribute, or dfn-type=abstract-op.
 LINE 301:1: Image doesn't exist, so I couldn't determine its width and height: 'images/storage-access-prompt.png'
 LINE 101:325: Multiple possible 'same site' dfn refs.
 Arbitrarily chose https://html.spec.whatwg.org/multipage/browsers.html#concept-site-same-site