Opened 9 years ago

Closed 9 years ago

#11931 closed defect (fixed)

IE6: html_rtl is failing

Reported by: haysmark Owned by: Douglas Hays
Priority: high Milestone: 1.6
Component: Core Version: 1.5
Keywords: Cc: James Burke
Blocked By: Blocking:

Attachments (2)

11931.patch (2.3 KB) - added by haysmark 9 years ago.
Refs #11931. Converts html_rtl to a robot test.
11931.2.patch (20.0 KB) - added by Douglas Hays 9 years ago.
patch that does a workaround of the IE6 bug with rtl+strict+iframe and automates the IE-event test in html_rtl w/o using the robot

Download all attachments as: .zip

Change History (8)

Changed 9 years ago by haysmark

Attachment: 11931.patch added

Refs #11931. Converts html_rtl to a robot test.

comment:1 Changed 9 years ago by haysmark

Owner: changed from anonymous to Douglas Hays

The last test is failing in IE6, apparently the iframe is not scrolling. The iframe could probably use dojo scroll code.

comment:2 Changed 9 years ago by bill

Cc: James Burke added

See #9514, this ticket is essentially a duplicate.

FYI, I talked to James about robot tests in core a while ago, we ended up with the plan to make a separate test file called something like runTestsRobot.html, leaving runTests.html for the non-robot tests. I forget why, maybe it was so the non-robot tests could run nightly on a test server although now w/your work the robot tests can run on a test server too.

Strangely I never checked that in, I guess I forgot about it. There are already a few robot tests in core so it makes sense to have a script to run them:

  • eventKeyPressRobot.html
  • robot/test_dnd.html
  • (and now this one)

comment:3 Changed 9 years ago by Douglas Hays

Milestone: tbd1.6
Status: newassigned

One if the tests is failing on IE6 and should be fixed in 1.6 if possible. I can also automate the IE test w/o the robot by positioning the rect100 box wherever the cursor happens to be (possibly offscreen) but that still validates the positioning code.

Changed 9 years ago by Douglas Hays

Attachment: 11931.2.patch added

patch that does a workaround of the IE6 bug with rtl+strict+iframe and automates the IE-event test in html_rtl w/o using the robot

comment:4 Changed 9 years ago by Douglas Hays

James, I'd like to commit the 11931_2.patch if you agree. The big question is if you want to workaround an IE6-specific bug by spending bytes in html.js::_fixIeBiDiScrollLeft. I don't really have an opinion. position() returns the wrong value because the HTML element clientWidth is wrong in IE6 when combined with rtl+iframe+vertical scrollbar+strict doctype.
I also automated the IE-event test in html_rtl.html w/o using the robot.
I also added more iframe tests for position() since that seems to be a point of weakness for IE.

comment:5 Changed 9 years ago by Douglas Hays

Description: modified (diff)
Summary: html_rtl blocks automated runs on IEIE6: html_rtl is failing

#9514 automates the test, but this ticket will be used to fix the IE6/position problem.
FAILED test: testScrolledPosition

comment:6 Changed 9 years ago by Douglas Hays

Resolution: fixed
Status: assignedclosed

(In [23193]) Fixes #11931. Add a special case to _fixIeBiDiScrollLeft to fix an IE6+rtl+strict+iframe+vscroll bug when calling position(n,true). Add more tests to cover a greater variety of scrollbar environments when testing position(). !strict

Note: See TracTickets for help on using tickets.