1257208 - Fix some confusing naming and casts in DirectionalityUtils.cpp

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[(no longer active)]

This node argument may point to any node type, and is not necessarily
an Element.  Its value is only useful when it's pointing to a text node,
but calling it a text node before we have checked its type is confusing.

Comment 1

[(no longer active)]

Attachment: Fix some confusing naming and casts in DirectionalityUtils.cpp

Comment 2

[Peter Van der Beken [:peterv]]

I don't really understand this patch, since afaict we only ever pass a text node to aChangedNode for ResetTextNodeDirection. We always check HasTextNodeDirectionalityMap before calling it. HasTextNodeDirectionalityMap only returns true if a nsTextNodeDirectionalityMap was constructed for the node, which only ever happens from AddEntryToMap, which is only ever called with the result of WalkDescendantsSetDirectionFromText (which only returns a text node). Could we make HasTextNodeDirectionalityMap assert that the node type is nsIDOMNode::TEXT_NODE? Maybe we can also make WalkDescendantsSetDirectionFromText return a nsTextNode* and nsTextNodeDirectionalityMap and AddEntryToMap take one?

At that point we could make TextNodeChangedDirection take a nsTextNode* (nsGenericDOMDataNode::SetTextInternal checks the nodetype before calling it) and then all the callers of ResetTextNodeDirection pass in a nsTextNode* for aChangedNode afaict.

Comment 3

[(no longer active)]

(In reply to Peter Van der Beken [:peterv] from comment #2)
> I don't really understand this patch, since afaict we only ever pass a text
> node to aChangedNode for ResetTextNodeDirection. We always check
> HasTextNodeDirectionalityMap before calling it. HasTextNodeDirectionalityMap
> only returns true if a nsTextNodeDirectionalityMap was constructed for the
> node, which only ever happens from AddEntryToMap, which is only ever called
> with the result of WalkDescendantsSetDirectionFromText (which only returns a
> text node).

D'oh, I was missing this part!  Sorry...

> Could we make HasTextNodeDirectionalityMap assert that the node
> type is nsIDOMNode::TEXT_NODE? Maybe we can also make
> WalkDescendantsSetDirectionFromText return a nsTextNode* and
> nsTextNodeDirectionalityMap and AddEntryToMap take one?
> 
> At that point we could make TextNodeChangedDirection take a nsTextNode*
> (nsGenericDOMDataNode::SetTextInternal checks the nodetype before calling
> it) and then all the callers of ResetTextNodeDirection pass in a nsTextNode*
> for aChangedNode afaict.

Will do!

Comment 4

[(no longer active)]

Attachment: Use the nsTextNode concrete type in several places in DirectionalityUtils.cpp instead of nsINode and nsIContent

Comment 5

[Peter Van der Beken [:peterv]]

Comment on attachment 8731276 [details] [diff] [review]
Fix some confusing naming and casts in DirectionalityUtils.cpp

Review of attachment 8731276 [details] [diff] [review]:
-----------------------------------------------------------------

I don't really understand this patch, since afaict we only ever pass a text node to aChangedNode for ResetTextNodeDirection. We always check HasTextNodeDirectionalityMap before calling it. HasTextNodeDirectionalityMap only returns true if a nsTextNodeDirectionalityMap was constructed for the node, which only ever happens from AddEntryToMap, which is only ever called with the result of WalkDescendantsSetDirectionFromText (which only returns a text node). Could we make HasTextNodeDirectionalityMap assert that the node type is nsIDOMNode::TEXT_NODE, and maybe we can also make WalkDescendantsSetDirectionFromText return a nsTextNode* and nsTextNodeDirectionalityMap and AddEntryToMap take one?

At that point we could make TextNodeChangedDirection take a nsTextNode* (nsGenericDOMDataNode::SetTextInternal checks the nodetype before calling it) and then all the callers of ResetTextNodeDirection pass in a nsTextNode* for aChangedNode afaict.

Comment 6

[Peter Van der Beken [:peterv]]

Comment on attachment 8732380 [details] [diff] [review]
Use the nsTextNode concrete type in several places in DirectionalityUtils.cpp instead of nsINode and nsIContent

Review of attachment 8732380 [details] [diff] [review]:
-----------------------------------------------------------------

Thanks for doing this.

Comment 7

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/01eff93538c8

Comment 8

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Backed out for Reftest failures at least on Linux x64 debug, e.g. dir_auto-set-contained-dir-L.html

Backout: https://hg.mozilla.org/integration/mozilla-inbound/rev/f112e7dad6f7

First non-busted run after push (which also shows the failures): https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=52195522dbac
Failure log: https://treeherder.mozilla.org/logviewer.html#?job_id=24440903&repo=mozilla-inbound
17:10:46     INFO -  REFTEST TEST-LOAD | file:///home/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/dirAuto/dir_auto-contained-dir-L-ref.html | 531 / 1712 (31%)
17:10:46     INFO -  ++DOMWINDOW == 54 (0x7f94d3023c00) [pid = 1128] [serial = 1144] [outer = 0x7f94d9455400]
17:10:46     INFO -  REFTEST TEST-UNEXPECTED-FAIL | file:///home/worker/workspace/build/tests/reftest/tests/layout/reftests/bidi/dirAuto/dir_auto-set-contained-dir-L.html | image comparison (==), max difference: 255, number of differing pixels: 3284

Comment 9

[(no longer active)]

D'oh, I made such a rookie mistake: used IsNodeOfType() instead of NodeType().  Embarrassing!

Comment 10

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/337789a5f9c6

Comment 11

[Peter Van der Beken [:peterv]]

(In reply to :Ehsan Akhgari from comment #9)
> D'oh, I made such a rookie mistake: used IsNodeOfType() instead of
> NodeType().  Embarrassing!

Hmpf, sorry for missing that.

Comment 12

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/337789a5f9c6