REGRESSION (r90699): 12 svg tests failing on Windows XP Debug (Tests) and 13 tests failing on Windows 7 Release (WebKit2 Tests)

[Adam Roben (:aroben)]

The following tests started failing on Windows XP Debug (Tests) in r90699 <http://trac.webkit.org/changeset/90699>:

    svg/W3C-I18N/g-dirLTR-ubNone.svg
    svg/W3C-I18N/g-dirLTR-ubOverride.svg
    svg/W3C-I18N/g-dirRTL-ubNone.svg
    svg/W3C-I18N/g-dirRTL-ubOverride.svg
    svg/W3C-I18N/text-dirLTR-ubNone.svg
    svg/W3C-I18N/text-dirRTL-ubNone.svg
    svg/W3C-I18N/text-dirRTL-ubOverride.svg
    svg/W3C-I18N/tspan-dirLTR-ubEmbed-in-rtl-context.svg
    svg/W3C-I18N/tspan-dirLTR-ubNone-in-rtl-context.svg
    svg/W3C-I18N/tspan-dirRTL-ubOverride-in-rtl-context.svg
    svg/W3C-I18N/tspan-direction-rtl.svg
    svg/text/bidi-tspans.svg

http://build.webkit.org/results/Windows%20XP%20Debug%20(Tests)/r90698%20(30527)/results.html passed
http://build.webkit.org/results/Windows%20XP%20Debug%20(Tests)/r90699%20(30528)/results.html failed

[Adam Roben (:aroben)]

Looks like the win-xp results for these tests were removed in r90699. I wonder what it was that made Adam think they were "redundant"?

[Adam Roben (:aroben)]

Well, I'll just put them back I guess.

[Adam Roben (:aroben)]

*** Bug 64278 has been marked as a duplicate of this bug. ***

[Adam Roben (:aroben)]

This also caused 13 tests to fail on WebKit2:

    editing/selection/caret-ltr-right.html
    editing/selection/drag-text-delay.html
    fast/css/pseudo-any.html
    fast/forms/textfield-overflow.html
    fast/ruby/after-block-doesnt-crash.html
    fast/ruby/after-table-doesnt-crash.html
    fast/ruby/before-block-doesnt-crash.html
    fast/ruby/before-table-doesnt-crash.html
    fast/ruby/generated-after-counter-doesnt-crash.html
    fast/ruby/generated-before-and-after-counter-doesnt-crash.html
    fast/ruby/generated-before-counter-doesnt-crash.html
    plugins/resize-from-plugin.html
    plugins/npruntime/embed-property-equality.html

[Adam Roben (:aroben)]

Committed r90744: <http://trac.webkit.org/changeset/90744>

[Adam Barth]

The deduplicate result script said they were.  We should figure out why the script was wrong.

[Adam Barth]

Apparently the script did some wrong things for WebKit2 as well.  Sounds like the script needs some more work to be reliable.

[Tony Chang]

(In reply to comment #7)
> Apparently the script did some wrong things for WebKit2 as well.  Sounds like the script needs some more work to be reliable.

Are the win-xp and the webkit2 bots using NRWT?  It gets the result fallback order from port.py.

[Adam Roben (:aroben)]

(In reply to comment #8)
> (In reply to comment #7)
> > Apparently the script did some wrong things for WebKit2 as well.  Sounds like the script needs some more work to be reliable.
> 
> Are the win-xp and the webkit2 bots using NRWT?

No.

[Adam Barth]

The wk2 fallback paths are supposed to be correct for NRWT.  Maybe there's some problem with how Eric added them?

[Eric Seidel (no email)]

Sorry.  Yes, the WK2 fallback paths are supposed to be correct, but I'm not surprised if they're not since https://bugs.webkit.org/show_bug.cgi?id=56729 is not solved yet.

NRWT also doesn't work for win https://bugs.webkit.org/show_bug.cgi?id=38756, so I'm not surprised the fallback paths were wrong there too.

We should fix deduplicate tests to not operate on ports that we know NRWT doesn't support yet. :(

[Eric Seidel (no email)]

(We should also fix the fallback paths on those port objects in NRWT, but if we're not using them, they're likely to go wrong again.)