831232 - Improve spec conformance for table@border

Status: RESOLVED FIXED

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

Description

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v1

ParseIntWithBounds will set Type() to eInteger even if the value is out of bounds. This breaks the default in MapAttributesIntoRule which does

      int32_t borderThickness = 1;

      if (borderValue->Type() == nsAttrValue::eInteger)
        borderThickness = borderValue->GetIntegerValue();

which causes us to fail the final three tests in the attached test.

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 702774 [details] [diff] [review]
Patch v1

I think dbaron should review this change.

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 702774 [details] [diff] [review]
Patch v1

r=dbaron, though it would be good if you:
 renamed the test to table-border-1{,-ref}.html
 added a table-border-2{,-ref,-notref}.html that checks that border="0" does match lack of a border attribute and does not match one of the references from this test (but remember to include only one table)

Since it looks like this changes our behavior from matching Opera to matching IE, could you let somebody at Opera know that we're changing?

Comment 3

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/mozilla-central/rev/454347e2ef1a