377874 - repeat-attrs on table causing complex content binding error

Status: RESOLVED FIXED

(Core Graveyard :: XForms, defect)

Description

[Steve Speicher]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a4pre) Gecko/20070416 Minefield/3.0a4pre
Build Identifier: 

This fails non-normative W3C test suite case 9.3.2.a

Reproducible: Always

Comment 1

[Steve Speicher]

Attachment: test case

Comment 2

[aaronr]

problem here is the initial output from the document doesn't live under a range since the range is introduced via anonymous content.  So when we run it through UpdateRepeatState, it is coming back as eType_NotApplicable.  Will have to change the logic there to look for xforms repeat attrs.  Well, at least the attr(s) that introduce repeats into anonymous content.

Comment 3

[aaronr]

Attachment: patch

made fix using check for repeat attrs, so no more error message popping up.

Comment 4

[aaronr]

Comment on attachment 262051 [details] [diff] [review]
patch

removing request.  I screwed something up.

Comment 5

[aaronr]

(In reply to comment #4)
> (From update of attachment 262051 [details] [diff] [review])
> removing request.  I screwed something up.
> 

Crap.  If we detect repeat attrs on the parent, we need to know if the current element is a repeat or not.  If it is then it is the anonymous repeat that is contained by the element with repeat attrs and we need to give it eType_NotApplicable.  Otherwise it'll be eType_Template.  However, at the stub element level we don't have access to mElement.  So looks like we are going to have to store mElement in stub element instead of in control stub, action, etc.  It is more overhead for every xforms element now, but not as much as having to move the repeat state and functions out of stub element.

Comment 6

[aaronr]

Attachment: patch2

Overrode UpdateRepeatState on repeat to figure out if it is an attr based repeat.  Beats moving mElement into stub element.

Comment 7

[aaronr]

Attachment: patch for 1.8 and 1.8.0 branches

Comment 8

[aaronr]

Attachment: patch2 for 1.8 and 1.8.0 branches

oops, forgot I had trace in the tree from another bug.

Comment 9

[alexander :surkov (:asurkov)]

Comment on attachment 262094 [details] [diff] [review]
patch2


>+  /**
>+   * Override of nsXFormsStubElement's function.  Needed to properly handle
>+   * the case where the repeat is generated by anonymous content.

nit: nice if you would mention here about attr-based repeat binding because it's not clear about what anon content you said.

>+    if (NS_SUCCEEDED(rv))
>+      // if repeat is attribute based, it won't go through the normal
>+      // attribute change processing, so should set the mIsAttributeBased
>+      // variable here
>+      mIsAttributeBased = isAttributeBased;
>+      if (isAttributeBased) {
>+        repeatState = eType_NotApplicable;
>+        SetRepeatState(repeatState);
>+        return repeatState;
>+      }
>+  }

nit: either line up properly if (isAttributeBased) or add brackets for top if statement.

Comment 10

[aaronr]

Attachment: patch for checkin

fixes surkov's nits, ready for checkin once the tree goes green

Comment 11

[aaronr]

checked into trunk

Comment 12

[aaronr]

checked into 1.8 and 1.8.0 branches