#11200 closed defect (fixed)
dojo/tests/_base/html_docScroll.html::t::testSpeed failure (IE8)
| Reported by: | bill | Owned by: | Eugene Lazutkin |
|---|---|---|---|
| Priority: | low | Milestone: | 2.0 |
| Component: | HTML | Version: | 1.5.0b2 |
| Keywords: | dohfail | Cc: | dgurney@… |
| Blocked By: | Blocking: |
Description (last modified by )
To reproduce:
- start IE8
Optional steps:
- press F12 to get developer console
- select "Console" tab
Mandatory step:
- load dojo/tests/runTests.html
This gets a failure on:
../../dojo/tests/_base/html_docScroll.html::t::testSpeed
It only happens when the developer console is started, and console is selected, so not sure if there's a real bug or not. Would be nice to avoid the test failure.
Change History (19)
comment:1 Changed 9 years ago by
| Milestone: | tbd → 1.6 |
|---|---|
| Owner: | changed from sjmiles to James Burke |
comment:2 Changed 9 years ago by
| Milestone: | 1.6 → future |
|---|
comment:3 Changed 9 years ago by
| Keywords: | dohfail added |
|---|
comment:4 Changed 8 years ago by
| Description: | modified (diff) |
|---|---|
| Keywords: | dohfail removed |
| Summary: | dojo/tests/_base/html_docScroll.html::t::testSpeed failure → dojo/tests/_base/html_docScroll.html::t::testSpeed failure (IE8) |
This is now failing for me even when the console isn't running, even on a fresh start of IE8. (I'm still testing it through runTests.html.) I'm going to comment out the test until it's fixed, and remove the dohfail keyword since I'm not sure this test ever worked.
comment:5 Changed 8 years ago by
| Milestone: | future → tbd |
|---|---|
| Owner: | changed from James Burke to Douglas Hays |
comment:7 Changed 8 years ago by
| Milestone: | tbd → 1.7 |
|---|---|
| Owner: | changed from Douglas Hays to Eugene Lazutkin |
| Priority: | normal → high |
I'm not sure it's a good practice to hide these failures. Dojo 1.7's docScroll takes twice as long as 1.6's version and the testcase was telling us that. I think that counts as a regression.
comment:8 Changed 8 years ago by
| Cc: | dgurney@… added |
|---|
Good catch, I'll reenable the test. The performance regression started in [25705], not sure why.
comment:10 Changed 8 years ago by
| Keywords: | dohfail added |
|---|
comment:11 Changed 8 years ago by
On IE8, all the lost time seems to be contributable to
var doc = win.doc, node = doc.parentWindow || doc.defaultView;
taking a very long time to execute while dojo.global is very fast. The addition of has('quirks') and win.body() seems to have had minimal effect.
(I added the var doc to try and make it a bit faster but it had no effect on IE8)
comment:12 Changed 8 years ago by
| Priority: | high → normal |
|---|
comment:13 Changed 8 years ago by
This does not affect end users, only devs that happen to have console open during test automation. Reducing priority to normal, non-blocking for rc1
comment:14 Changed 8 years ago by
As I mentioned above, this fails even when the console isn't running, even on a fresh start of IE8. (I'm still testing it through runTests.html.) Therefore it does affect end users, although not sure how much.
comment:15 Changed 8 years ago by
| Priority: | high → low |
|---|
comment:16 Changed 7 years ago by
| Milestone: | 1.8 → 2.0 |
|---|
1.8 has been tagged; moving all outstanding tickets to next major release milestone.
comment:17 Changed 6 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
(sadly) punting seemingly abandoned ticket and meta tickets to future